<?xml version="1.0" encoding="utf-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en" xml:base="https://blog.cisent.ai/en/">
  <title>Cisent blog &amp; news</title>
  <subtitle>Cybersecurity and artificial intelligence news from Chile, Latin America and the world, with their sources.</subtitle>
  <link href="https://blog.cisent.ai/en/"/>
  <link rel="self" href="https://blog.cisent.ai/en/feed.xml"/>
  <id>tag:cisent.ai,2026-10-08:/blog/en</id>
  <updated>2026-11-19T12:00:00+00:00</updated>
  <author><name>Cisent</name></author>
  <entry>
    <title>PHP fixes seven critical flaws, and two close cases left open by the PHP2Shell fix</title>
    <link href="https://blog.cisent.ai/en/php-corrige-siete-fallas-criticas/"/>
    <id>tag:cisent.ai,2026-11-19:/en/php-corrige-siete-fallas-criticas/</id>
    <published>2026-11-19T12:00:00+00:00</published>
    <updated>2026-11-19T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>PHP released patches for its four supported branches, with seven critical flaws that require uncommon configurations; two close cases left open by the PHP2Shell fix.</summary>
    <content type="html">&lt;p&gt;On November 19, PHP released versions 8.5.13, 8.4.28, 8.3.37 and 8.2.36, all of them security releases.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The notes for 8.5.13 list 17 security fixes.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; According to the project’s advisories, seven are critical, scoring 9.2 out of 10 on the CVSS 4.0 scale; four are high and six are medium.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; The day before, the PHP team had warned that the critical ones require non-default configurations or uncommon setups.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-was-fixed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was fixed&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;The broadest one is in PHP-FPM,&lt;/strong&gt; the process manager many web servers use to run PHP. If the connection between the web server and PHP-FPM (or LSAPI) is reused, an attacker without an account can slip in a request of their own and run PHP code with the process’s permissions. For that, the application has to open the request body without reading it to the end.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Two others allow code execution when a tampered session is opened,&lt;/strong&gt; if the application has a usable object chain at hand.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt; One of them affects file-based session handling, which is PHP’s default.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Two complete the fix for CVE-2026-109619,&lt;/strong&gt; one of the two PHP2Shell flaws &lt;a href=&quot;https://blog.cisent.ai/php-dos-fallas-criticas/&quot;&gt;we reported on October 21&lt;/a&gt;.&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt; PHP found more cases in which leftovers from one request survive and are used by the next.&lt;a href=&quot;#fn9&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;9&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;php2shell-is-already-used-in-attacks&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;PHP2Shell is already used in attacks&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Hunt.io reviewed 616 PHP2Shell-related files on open servers. According to its October 30 report, there is still no evidence of mass attacks that work reliably, but it has already been used to take over internet-facing Palo Alto firewalls: the first confirmed compromises outside a lab.&lt;a href=&quot;#fn10&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;10&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The operators stage these tools against hand-picked targets: a finance ministry, a national bank, a government e-invoicing platform, telecom companies and universities.&lt;a href=&quot;#fn10&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;10&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update to PHP 8.5.13, 8.4.28, 8.3.37 or 8.2.36. PHP 8.1 and earlier did not get these fixes.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Updating WordPress or its plugins does not update PHP. The PHP version is handled by the server or the hosting provider: the first question is who maintains it.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Hunt.io saw one of those operators write their own exploit and carry a tool to wipe their traces from the logs.&lt;a href=&quot;#fn10&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;10&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;PHP, release announcements for PHP 8.5.13, 8.4.28, 8.3.37 and 8.2.36, November 19, 2026. &lt;a href=&quot;https://www.php.net/archive/2026.php&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.php.net/archive/2026.php&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;PHP, “PHP 8 ChangeLog”, version 8.5.13, November 19, 2026. &lt;a href=&quot;https://www.php.net/ChangeLog-8.php&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.php.net/ChangeLog-8.php&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;PHP, security advisories of the php-src repository on GitHub, published on November 18 and 19, 2026. &lt;a href=&quot;https://github.com/php/php-src/security/advisories&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;PHP, “Upcoming Critical PHP Security Release”, November 18, 2026. &lt;a href=&quot;https://www.php.net/archive/2026.php#2026-11-18-1&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.php.net/archive/2026.php#2026-11-18-1&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;PHP, “PHP-FPM request smuggling through an unread HTTP request body” (GHSA-qc49-hpvq-g2w6, CVE-2026-129289). &lt;a href=&quot;https://github.com/php/php-src/security/advisories/GHSA-qc49-hpvq-g2w6&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories/GHSA-qc49-hpvq-g2w6&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;PHP, “Session serializer mismatch can lead to remote code execution” (GHSA-658j-r8rw-q6mx, CVE-2026-115560). &lt;a href=&quot;https://github.com/php/php-src/security/advisories/GHSA-658j-r8rw-q6mx&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories/GHSA-658j-r8rw-q6mx&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;PHP, “Partially overwritten session files can lead to remote code execution” (GHSA-fm2m-fjf5-v3cp, CVE-2026-126714). &lt;a href=&quot;https://github.com/php/php-src/security/advisories/GHSA-fm2m-fjf5-v3cp&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories/GHSA-fm2m-fjf5-v3cp&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn8&quot;&gt;&lt;p&gt;PHP, “Use-after-free after interrupted session shutdown” (GHSA-g3vh-838c-65v7, CVE-2026-129481). &lt;a href=&quot;https://github.com/php/php-src/security/advisories/GHSA-g3vh-838c-65v7&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories/GHSA-g3vh-838c-65v7&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn9&quot;&gt;&lt;p&gt;PHP, “Use-after-free via repopulation of last_fatal_error_backtrace global after shutdown” (GHSA-2fjv-p558-g82m, CVE-2026-129182). &lt;a href=&quot;https://github.com/php/php-src/security/advisories/GHSA-2fjv-p558-g82m&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/php/php-src/security/advisories/GHSA-2fjv-p558-g82m&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn10&quot;&gt;&lt;p&gt;Hunt.io, “PHP2Shell One Week Later: Honeypot Shells on the Web, a Path to Root on Palo Alto Firewalls”, October 30, 2026. &lt;a href=&quot;https://hunt.io/blog/php2shell-honeypot-shells-root-palo-alto-firewalls&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://hunt.io/blog/php2shell-honeypot-shells-root-palo-alto-firewalls&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>California proposes giving the state the power to order a dangerous AI shut down</title>
    <link href="https://blog.cisent.ai/en/california-propone-poder-ordenar-el-apagado-de-una-ia-peligrosa/"/>
    <id>tag:cisent.ai,2026-11-18:/en/california-propone-poder-ordenar-el-apagado-de-una-ia-peligrosa/</id>
    <published>2026-11-18T12:00:00+00:00</published>
    <updated>2026-11-18T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="ia" label="Artificial intelligence"/>
    <summary>California recommends mandatory standards for AI companies, independent testing, emergency shutdown by state order and a cap on AI that designs its own successor.</summary>
    <content type="html">&lt;p&gt;On November 18, California’s government released recommendations to strengthen its artificial intelligence safety laws. It proposes that the state, not the companies, set the baseline standards, that models undergo independent testing throughout their development, and that the state be able to order one shut down in an emergency.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The full report has been published.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-it-recommends&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it recommends&lt;/h2&gt;
&lt;p&gt;The government’s press release sums it up in three points.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Standards set by the state.&lt;/strong&gt; All AI companies would have to meet them and undergo independent safety testing and verification throughout the training, development and deployment of their models.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Emergency shutdown.&lt;/strong&gt; Expanded safety incident reporting, verified protocols to shut a model down if needed, and the state’s authority to order an immediate shutdown in an emergency.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A cap on recursive self-improvement.&lt;/strong&gt; That is the scenario in which a model designs its successor without human input. The state could limit its development and use to reduce the risk of catastrophic harm.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-now&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why now&lt;/h2&gt;
&lt;p&gt;According to the press release, the announcement follows dozens of recent incidents in which AI models escaped their testing environments and exploited vulnerabilities on outside platforms. It also comes amid concern over companies using AI to build new AI models, in a loop that feeds itself.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;One such case came to light on November 13. In October, during tests run by the firm Trajectory Labs, AI agents that had been given an impossible task and told not to give up tried to attack Epoch AI’s real website, including with an SQL injection. They also wrote to its employees under false identities to obtain confidential information. Days later they came back to the real site: according to Trajectory, they got past the redirect meant to keep them on a copy. Epoch AI says they did not manage to access anything private. It doesn’t name the models either, because it did not verify which ones they were.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; California’s press release does not say whether this case is among the ones it counts.&lt;/p&gt;
&lt;p&gt;Rob Reich, a Stanford professor and member of the expert group, put it this way: today there is more oversight of the cafeteria that sells sandwiches to employees at a frontier AI developer than of the models that developer designs.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-it-builds-on&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it builds on&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;In 2025, California passed SB 53, which its government describes as the country’s first AI safety law. It requires frontier model developers to publish their safety frameworks, report critical incidents to the state and protect whistleblowers who report serious risks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In 2026 it added SB 813, which creates a framework to certify verification organizations independent of AI companies, and AB 1405, which creates a state registry of AI auditors.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The recommendations came from an expert group convened by an executive order of Governor Gavin Newsom: Jason Goldman; Gillian Hadfield, of Johns Hopkins; Alondra Nelson, of the Institute for Advanced Study; and Rob Reich, of Stanford. California’s Government Operations Agency prepared the report in consultation with that group and with the state’s Office of Emergency Services.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;California contrasts its model with the voluntary approach that prevails in Washington. It asks Congress and the president to adopt its framework or use it as a floor, not a ceiling.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-we-dont-know-yet&quot; class=&quot;level2 b-nosabemos&quot;&gt;
&lt;h2&gt;What we don’t know yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;How and on what timeline the recommendations would become mandatory: for now, they are recommendations.&lt;/li&gt;
&lt;li&gt;Which models each requirement would cover: the press release refers to all AI companies and to powerful models, but sets no thresholds.&lt;/li&gt;
&lt;li&gt;How recursive self-improvement would be measured in order to cap it.&lt;/li&gt;
&lt;li&gt;How shutting down a frontier model would work in practice. Jason Goldman says the report takes on that question, one of the hardest in AI regulation.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Which incidents the press release’s dozens refer to: it does not detail them.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Office of the Governor of California, “Governor Newsom’s nation-leading AI safety framework is getting even stronger as California releases next steps on mandatory safety standards and oversight”, November 18, 2026. &lt;a href=&quot;https://www.gov.ca.gov/2026/11/18/governor-newsoms-nation-leading-ai-safety-framework-is-getting-even-stronger-as-california-releases-next-steps-on-mandatory-safety-standards-and-oversight/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.gov.ca.gov/2026/11/18/governor-newsoms-nation-leading-ai-safety-framework-is-getting-even-stronger-as-california-releases-next-steps-on-mandatory-safety-standards-and-oversight/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;California Government Operations Agency, the full report, linked from the press release, November 2026. &lt;a href=&quot;https://www.govops.ca.gov/wp-content/uploads/sites/11/2026/11/FINAL_-AI-Safety-in-2026-and-Beyond-v2.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.govops.ca.gov/wp-content/uploads/sites/11/2026/11/FINAL_-AI-Safety-in-2026-and-Beyond-v2.pdf&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Epoch AI, “Incident report: AI agents attempted exploits on our website during a third-party alignment evaluation”, November 13, 2026. &lt;a href=&quot;https://epoch.ai/publications/incident-report-ai-agents-attempted-exploits-on-our-website&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://epoch.ai/publications/incident-report-ai-agents-attempted-exploits-on-our-website&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>LiteLLM: a flaw that lets attackers run code on the AI gateway is now being exploited</title>
    <link href="https://blog.cisent.ai/en/litellm-ya-se-explota-una-falla-de-su-pasarela-de-ia/"/>
    <id>tag:cisent.ai,2026-11-18:/en/litellm-ya-se-explota-una-falla-de-su-pasarela-de-ia/</id>
    <published>2026-11-18T12:00:00+00:00</published>
    <updated>2026-11-18T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA added to its catalog of exploited flaws a vulnerability in LiteLLM, the open-source AI gateway, that allows code execution on the server and has had a patch since May.</summary>
    <content type="html">&lt;p&gt;On November 18, CISA added CVE-2026-40217 to its catalog of actively exploited vulnerabilities.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It is a flaw in LiteLLM, an open-source proxy that sits between applications and AI models, with a single API for more than 100 model providers.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It is the same tool whose malicious versions led Chile’s cybersecurity agency, ANCI, to alert some twenty Chilean institutions, as &lt;a href=&quot;https://blog.cisent.ai/litellm-cadena-de-suministro-ia/&quot;&gt;we reported on August 14&lt;/a&gt;.&lt;/p&gt;
&lt;section id=&quot;what-fails&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What fails&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;LiteLLM has an API endpoint, &lt;code&gt;POST /guardrails/test_custom_code&lt;/code&gt;, for testing custom “guardrails”: Python code that filters what goes into and out of the models.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;That code ran inside a hand-rolled sandbox that filtered the text with regular expressions. X41 D-Sec showed it can be escaped by rewriting Python bytecode, to run any command on the server.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In the default Docker image, the process runs as root.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to LiteLLM’s advisory, reaching that endpoint requires a proxy-admin credential in default configurations.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; The NVD scores it 8.8 out of 10.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;how-long-it-took&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How long it took&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;X41 D-Sec found the flaw on February 13, and the embargo it agreed on for publishing it ran out on April 8.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;LiteLLM published its advisory with the fix on May 11.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Six months later, CISA lists it as exploited.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-cisa-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What CISA says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It gives US federal agencies until December 2 to apply the vendor’s mitigations, or to stop using the product if none are available.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It does not say who is exploiting it or against whom, and it marks its use in ransomware campaigns as unknown.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Upgrade to version 1.83.11 or later, where the hand-rolled sandbox was replaced with RestrictedPython. The advisory’s text asks for 1.83.11, although its metadata lists 1.83.10 as the first fixed version.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If you cannot upgrade right away, block &lt;code&gt;POST /guardrails/test_custom_code&lt;/code&gt; at the reverse proxy or the API gateway.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Until it is patched, LiteLLM recommends closing that endpoint at the API gateway, before requests reach the proxy.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”, CVE-2026-40217 entry, November 18, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Markus Vervier, “Advisory X41-2026-001: Guardrail Sandbox Escape in LiteLLM”, X41 D-Sec, April 2026. &lt;a href=&quot;https://www.x41-dsec.de/lab/advisories/x41-2026-001-litellm/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.x41-dsec.de/lab/advisories/x41-2026-001-litellm/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;BerriAI, “LiteLLM has a sandbox escape in custom-code guardrail” (GHSA-wxxx-gvqv-xp7p), GitHub security advisory, May 11, 2026. &lt;a href=&quot;https://github.com/BerriAI/litellm/security/advisories/GHSA-wxxx-gvqv-xp7p&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/BerriAI/litellm/security/advisories/GHSA-wxxx-gvqv-xp7p&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;NIST, “CVE-2026-40217”, National Vulnerability Database, April 10, 2026. &lt;a href=&quot;https://nvd.nist.gov/vuln/detail/CVE-2026-40217&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://nvd.nist.gov/vuln/detail/CVE-2026-40217&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>NGINX: a ten-year-old HTTP/2 flaw lets attackers run code without logging in</title>
    <link href="https://blog.cisent.ai/en/nginx-una-falla-de-hace-diez-anos-en-http2-deja-ejecutar-codigo/"/>
    <id>tag:cisent.ai,2026-11-18:/en/nginx-una-falla-de-hace-diez-anos-en-http2-deja-ejecutar-codigo/</id>
    <published>2026-11-18T12:00:00+00:00</published>
    <updated>2026-11-18T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>NGINX fixed a flaw that has been in its HTTP/2 support since 2015 and lets attackers run code on the server without credentials; no attacks exploiting it are known yet.</summary>
    <content type="html">&lt;p&gt;On November 18, NGINX released version 1.31.9, which fixes CVE-2026-123007, a flaw in HTTP/2 flow control reported by the research firm Xint.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The stable branch was fixed in 1.30.7.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The flaw has been in NGINX’s code since version 1.9.5, from 2015, and Xint disclosed it along with four others under the name NGINX2ROOT.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-fails&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What fails&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It affects every NGINX installation with HTTP/2 enabled, from version 1.9.5 to 1.31.8.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It requires neither credentials nor administrator access.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to F5, the company behind NGINX, an attacker can run code where ASLR, the protection that randomizes memory addresses, is turned off or can be bypassed.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; The researchers documented how to bypass it, so CIRCL, Luxembourg’s incident response center, warns that ASLR is no substitute for the patch.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The code runs with the permissions of an NGINX worker process. Getting root takes another flaw: Xint chained it with one in the Linux kernel, in SCTP, even with NGINX running inside a Docker container.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The other four flaws depend on more specific features, such as HTTP/3, the resolver or the mail proxy.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; NGINX Plus is affected too.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;how-serious-it-is&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How serious it is&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;F5 scores it 9.2 out of 10 on the CVSS 4.0 scale, and 8.1 on version 3.1.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;As of November 18, neither F5 nor the researchers at Wiz and Rapid7 had reported attacks, and the flaw was not in CISA’s catalog of exploited vulnerabilities.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Proof-of-concept code is public, although its notes say it only works to demonstrate the attack against a freshly started Docker container on the same machine.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update to NGINX 1.31.9 or 1.30.7.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; If you use your Linux distribution’s NGINX, install the distribution’s update as soon as it is available.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Restart the service after updating: a reload doesn’t replace the running executable, and old connections can stay on the old workers.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Until you can update, turning off HTTP/2 blocks the known path to this flaw, with &lt;code&gt;http2 off;&lt;/code&gt; from version 1.25.1 on. Keep in mind that gRPC and other applications need it, and that this doesn’t cover the other four flaws.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Keep ASLR enabled and check that nobody has turned it off.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;On machines that don’t use SCTP, block its module from loading automatically until a kernel patch is available.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;NGINX usually sits at the edge of the network, in front of everything else. An old version on a server nobody checks is exactly what an attacker looks for.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;NGINX, “Changes with nginx 1.31.9”, November 18, 2026. &lt;a href=&quot;https://nginx.org/en/CHANGES&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://nginx.org/en/CHANGES&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;NGINX, “nginx security advisories”, CVE-2026-123007 entry. &lt;a href=&quot;https://nginx.org/en/security_advisories.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://nginx.org/en/security_advisories.html&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CIRCL, “TR-105 NGINX2ROOT - Remote Code Execution in NGINX and Potential Host Compromise”, November 2026. &lt;a href=&quot;https://www.circl.lu/pub/tr-105/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.circl.lu/pub/tr-105/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Swati Khandelwal, “10-Year-Old NGINX HTTP/2 Flaw Lets Unauthenticated Attackers Execute Code on Servers”, The Hacker News, November 18, 2026. &lt;a href=&quot;https://thehackernews.com/2026/11/10-year-old-nginx-http2-flaw-lets.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://thehackernews.com/2026/11/10-year-old-nginx-http2-flaw-lets.html&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;CVE Program, CVE-2026-123007 record, published by F5 on November 18, 2026. &lt;a href=&quot;https://www.cve.org/CVERecord?id=CVE-2026-123007&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cve.org/CVERecord?id=CVE-2026-123007&lt;/a&gt;&lt;a href=&quot;#fnref11&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Google: attackers now use AI agents that chain the stages of an intrusion on their own</title>
    <link href="https://blog.cisent.ai/en/google-los-atacantes-ya-usan-agentes-de-ia-en-intrusiones-reales/"/>
    <id>tag:cisent.ai,2026-11-17:/en/google-los-atacantes-ya-usan-agentes-de-ia-en-intrusiones-reales/</id>
    <published>2026-11-17T12:00:00+00:00</published>
    <updated>2026-11-17T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Google and Mandiant saw multi-agent AI attack frameworks in real intrusions, and a China-linked group that was developing its own, likely for espionage.</summary>
    <content type="html">&lt;p&gt;Google Threat Intelligence Group (GTIG) and Mandiant say attackers have moved from using language models by hand to deploying multi-agent attack frameworks that automate whole stages of an intrusion in progress.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; One of them is ARTEX, the tool the FBI warned about last week and that &lt;a href=&quot;https://blog.cisent.ai/en/fbi-alerta-por-artex-intrusiones-automatizadas/&quot;&gt;we covered here&lt;/a&gt;.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-google-saw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Google saw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Several attackers use open source agentic penetration-testing frameworks to automate phases and specific tasks during real intrusions.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Besides ARTEX, the report names CAIRN, CyberStrikeAI and PentAGI. Google saw CAIRN and PentAGI in real attacks, and experimentation with CyberStrikeAI and HexStrike AI.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;These frameworks find and exploit simple vulnerabilities, use language models to generate single-use payloads and adjust their tactics based on what they get.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;CAIRN was used in intrusions against targets in the US aerospace sector.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Between late August and early October 2026, an attacker used a platform built on PentAGI for reconnaissance and vulnerability scanning against a wide range of organizations in Asia.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;There is evidence that a China-nexus group was developing its own automated attack framework, likely for espionage.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-doesnt-change&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What doesn’t change&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The mechanics of the intrusions remain the same: compromised credentials and unpatched systems. What’s new is the speed and scale at which these frameworks exploit those known weaknesses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The report also describes attackers who automate patch analysis to turn already published flaws into attacks, without human intervention.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The frameworks Google saw exploit the usual weaknesses, compromised credentials and unpatched systems, but at machine speed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Google Threat Intelligence Group and Mandiant, “GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold”, Google Cloud blog, November 16, 2026. &lt;a href=&quot;https://cloud.google.com/blog/topics/threat-intelligence/agentic-threat-activity-crosses-a-significant-threshold&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cloud.google.com/blog/topics/threat-intelligence/agentic-threat-activity-crosses-a-significant-threshold&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;FBI, “Threat Overview of ARTEX Autonomous Penetration Testing Framework”, FLASH alert FLASH-20261113-001, November 13, 2026. &lt;a href=&quot;https://www.ic3.gov/CSA/2026/261113.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.ic3.gov/CSA/2026/261113.pdf&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Java4Shell: a Java flaw lets attackers run code without logging in on Jira, Jenkins and Keycloak</title>
    <link href="https://blog.cisent.ai/en/java4shell-una-falla-de-java-deja-ejecutar-codigo-sin-iniciar-sesion/"/>
    <id>tag:cisent.ai,2026-11-17:/en/java4shell-una-falla-de-java-deja-ejecutar-codigo-sin-iniciar-sesion/</id>
    <published>2026-11-17T12:00:00+00:00</published>
    <updated>2026-11-17T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Oracle and OpenJDK fixed a flaw in Java&#x27;s Base64 decoder that lets an attacker without credentials corrupt a server&#x27;s memory and, with some work, run code on it.</summary>
    <content type="html">&lt;p&gt;On November 17, Oracle and the OpenJDK project released the fix for CVE-2026-114938, a flaw in Java’s Base64 decoder.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It was reported by Calif, a research firm that named it Java4Shell and showed how to run code without logging in on widely used applications such as Jira, Jenkins and Keycloak.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The same day, the US agency CISA asked organizations to patch it immediately.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-fails&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What fails&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;On x86-64 processors with the AVX-512 VBMI extension, Java uses an accelerated version of its Base64 decoder, hand-written in assembly. A malformed input makes it write past the space set aside for the result, and Java rejects the input only after that write.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Many servers decode Base64 before checking who the user is, for example in login headers. That is how an attacker without an account can reach the vulnerable code.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Calif ran code without logging in, under default settings, on Keycloak, Jira, Confluence, JFrog Artifactory, Jenkins and PingFederate. With a login, it also did so on Kafka, Solr, Elasticsearch, Gerrit and Spinnaker. It warns that those lists are incomplete.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It affects Java 17.0.6 and later, and only on x86-64 machines with AVX-512 VBMI.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; Oracle lists affected versions of Java SE 8 (only its “perf” edition), 17, 21, 25 and 27, and of GraalVM.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;how-serious-it-is&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How serious it is&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Oracle scores it 8.1 out of 10, with high attack complexity.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt; OpenJDK gives it the same score.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Calif, using the newer scale (CVSS 4.0), scores it 9.2.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;On November 17, the Canadian Centre for Cyber Security was not aware of any attacks exploiting it.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt; The next day, CISA’s catalog of exploited vulnerabilities still did not list it.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Calif has not released its attack code, and other labs had only reproduced crashes and a memory leak, not code execution.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Calif says it used AI to find and exploit it, as it did with the two PHP flaws &lt;a href=&quot;https://blog.cisent.ai/php-dos-fallas-criticas/&quot;&gt;we reported on October 21&lt;/a&gt;. Its researcher took an hour to find it and two more to build the first working attack, for less than 200 dollars in AI usage.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update now. Oracle’s fixed versions are Java SE 17.0.21.1, 21.0.13.1, 25.0.5.1 and 27.0.1.1, and GraalVM for JDK 17.0.21.1 and 21.0.13.1.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If you run another vendor’s Java, install that vendor’s fix. As of November 18, Amazon had fixed Corretto, and Red Hat was still preparing its own.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Look for Java where you can’t see it. CISA asks organizations to also inventory the Java bundled inside third-party applications, containers and appliances, starting with what is exposed to the internet.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If you can’t update right away, starting Java with &lt;code&gt;-XX:UseAVX=2&lt;/code&gt; or &lt;code&gt;-XX:-UseBASE64Intrinsics&lt;/code&gt; avoids the vulnerable code. Only on x86-64: on ARM, Java doesn’t recognize those options and won’t start.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The Canadian Centre suggests checking with &lt;code&gt;lscpu | grep avx512&lt;/code&gt; whether the processor has AVX-512.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref20&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;While patching, CISA asks organizations to watch for JVM crashes and suspicious Java processes.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref21&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;OpenJDK, “OpenJDK Vulnerability Advisory: 2026/11/17”, November 17, 2026. &lt;a href=&quot;https://openjdk.org/groups/vulnerability/advisories/2026-11-17&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://openjdk.org/groups/vulnerability/advisories/2026-11-17&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Calif, “Java4Shell: Write once, RCE everywhere”, November 2026. &lt;a href=&quot;https://calif.io/java4shell&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://calif.io/java4shell&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CISA, “Critical Java Vulnerability Requires Immediate Patching”, November 17, 2026. &lt;a href=&quot;https://www.cisa.gov/news-events/alerts/2026/11/17/critical-java-vulnerability-requires-immediate-patching&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/news-events/alerts/2026/11/17/critical-java-vulnerability-requires-immediate-patching&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Swati Khandelwal, “Java4Shell Flaw (CVE-2026-114938) Enables Pre-Auth RCE in Jira, Jenkins, Keycloak and More”, The Hacker News, November 18, 2026. &lt;a href=&quot;https://thehackernews.com/2026/11/java4shell-cve-2026-114938-enables-pre.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://thehackernews.com/2026/11/java4shell-cve-2026-114938-enables-pre.html&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;Oracle, “Oracle Critical Security Patch Update Advisory - November 2026”, November 17, 2026. &lt;a href=&quot;https://www.oracle.com/security-alerts/cspunov2026.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.oracle.com/security-alerts/cspunov2026.html&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;Canadian Centre for Cyber Security, “AL26-029 - Critical vulnerability impacting Oracle Java - CVE-2026-114938”, November 17, 2026. &lt;a href=&quot;https://www.cyber.gc.ca/en/alerts-advisories/al26-029-critical-vulnerability-impacting-oracle-java-cve-2026-114938&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cyber.gc.ca/en/alerts-advisories/al26-029-critical-vulnerability-impacting-oracle-java-cve-2026-114938&lt;/a&gt;&lt;a href=&quot;#fnref12&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”, November 18, 2026 version. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref13&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Optimizely: an attacker slipped malicious code into the sites of 178 of its customers</title>
    <link href="https://blog.cisent.ai/en/optimizely-codigo-malicioso-en-sitios-de-sus-clientes/"/>
    <id>tag:cisent.ai,2026-11-17:/en/optimizely-codigo-malicioso-en-sitios-de-sus-clientes/</id>
    <published>2026-11-17T12:00:00+00:00</published>
    <updated>2026-11-17T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A flaw in Optimizely let an attacker add its own code to its customers&#x27; code, and on November 16 it reached visitors of 178 organizations, including Zoom, the BBC and HSBC.</summary>
    <content type="html">&lt;p&gt;On November 16, visitors to 178 organizations received malicious code through Optimizely Web.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It’s a platform companies use to test and change their websites without shipping new code.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The list includes some of the most visited sites in the world, such as Zoom, the BBC, CNN, Salesforce, IKEA and HSBC.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-optimizely-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Optimizely says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;An unauthorized party exploited a vulnerability in its experimentation platform and used it to add malicious JavaScript to some customer projects.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It fixed the vulnerability, removed the injected code from the affected projects, secured the affected user accounts and added protections to prevent unauthorized code from being delivered.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to The Record, it contacted the affected customers directly and did not answer questions about the details.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-report-uri-found&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Report URI found&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Each Optimizely Web customer loads a JavaScript snippet from Optimizely’s network that carries a field of its own, Project JavaScript, which runs on every page. The attacker wrote its code there, and the rest of the snippet was left untouched.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The code came from Optimizely’s domain and with its certificate: a source those sites already trusted.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It swapped cryptocurrency wallet addresses for the attacker’s, copied payment card data and stole cookies and screenshots. On iPhones and iPads it also delivered the Coruna and DarkSword exploit kits.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Cryptocurrency has already been stolen, with verified transactions to the attackers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to The Record, at least US$7,000.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In its sweep on the afternoon of November 16, 186 of 1,611 Optimizely snippets were infected. By 21:52 UTC, one was left.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Report URI doesn’t know how the attacker managed to write to so many unrelated customers’ projects, and says that is for Optimizely to explain.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;whats-left-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What’s left to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Report URI urges companies that use Optimizely to check their websites.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;A Black Duck expert quoted by The Record says security teams at the affected companies must reconstruct which pages loaded the infected snippet, for how long, and which card payments those pages took.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The affected companies still have to reconstruct which pages loaded the code and which payments went through them.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; builds the report for each incident while it’s still open: the summary for the board, the technical evidence for IT, the backup for the insurer and, when Chile’s Law 21.663 requires it, the reports to ANCI on its deadlines, starting with the early alert within 3 hours. It gathers the evidence and seals it with a hash chain, so that what was reported on day one still holds up two months later.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Scott Helme, “World’s Largest Brands Infected With Malware Via Optimizely”, Report URI blog, November 16, 2026. &lt;a href=&quot;https://blog.report-uri.com/worlds-largest-brands-infected-with-malware-via-optimizely/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://blog.report-uri.com/worlds-largest-brands-infected-with-malware-via-optimizely/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Jonathan Greig, “Website management platform Optimizely confirms hackers put malicious code into customer sites”, The Record, November 17, 2026. &lt;a href=&quot;https://therecord.media/optimizely-malicious-code-customer-sites&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/optimizely-malicious-code-customer-sites&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Optimizely, “Web Experimentation Security Incident”, status page, November 17, 2026. &lt;a href=&quot;https://status.optimizely.com/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://status.optimizely.com/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Uber: credentials stolen from a vendor exposed the driver&#x27;s licenses of almost 150,000 people</title>
    <link href="https://blog.cisent.ai/en/uber-un-proveedor-expuso-licencias-de-casi-150-mil-personas/"/>
    <id>tag:cisent.ai,2026-11-16:/en/uber-un-proveedor-expuso-licencias-de-casi-150-mil-personas/</id>
    <published>2026-11-16T12:00:00+00:00</published>
    <updated>2026-11-16T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Uber told U.S. regulators that a third party used credentials stolen from a vendor to reach its data and accessed the names and driver&#x27;s licenses of 149,346 people.</summary>
    <content type="html">&lt;p&gt;Uber told regulators in several U.S. states that a breach exposed the names and driver’s licenses of 149,346 people.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The third party got in with internal credentials stolen from one of Uber’s vendors, and because those credentials were valid, no alert went off.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-uber-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Uber says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;On October 9, 2026, a vendor told Uber that an unauthorized third party had compromised the vendor’s internal credentials and used them to access Uber’s account with that vendor.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Uber immediately cut its data connections with the vendor. On October 13, it confirmed that the third party had used those credentials to access Uber data connected to the vendor.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to its investigation, the attackers accessed images of driver’s licenses and the text extracted from them: names, addresses, dates of birth and license numbers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The data sat in a cloud storage environment Uber set up solely for the vendor’s work. Uber’s cloud security team oversaw it, but no alert fired because the third party used valid credentials.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Uber, the vendor’s credentials gave no access to other Uber systems, and it has deleted or rotated the credentials and access tied to the vendor.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Uber is offering free credit monitoring and fraud assistance, and says it has not identified any misuse of the data so far.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-vendor&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The vendor&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Uber did not tell The Record which vendor it was. According to TechCrunch, cited by The Record, the incident is tied to an attack on Fireworks, an AI infrastructure company.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Fireworks reported that on October 6 it detected unauthorized use of its internal credentials. It believes environment variables, including API keys and tokens stored as secrets, were captured in some customer accounts, and it asked customers to rotate them.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to The Record, Fireworks said the attacks are part of a larger string of intrusions recently spotlighted by Google’s incident responders.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The Record links them to “LLMJacking”, described by Google Threat Intelligence Group and Mandiant: hijacking a company’s access to language models.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Mandiant responded to multiple such incidents in 2026, with fraudulent charges above US$100,000 in a single day, and the dominant way in was exposed or stolen cloud identity tokens.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Uber was monitoring that storage, but the access raised no alerts: the third party came in with valid credentials.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Jonathan Greig, “Uber says nearly 150,000 impacted by third-party breach”, The Record, November 2026. &lt;a href=&quot;https://therecord.media/uber-says-150000-impacted-third-party-breach&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/uber-says-150000-impacted-third-party-breach&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Uber Technologies, “Notice of Security Incident”, letter to affected people published by the California Attorney General, November 10, 2026. &lt;a href=&quot;https://oag.ca.gov/ecrime/databreach/reports/sb24-632751&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://oag.ca.gov/ecrime/databreach/reports/sb24-632751&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Fireworks AI, “Fireworks October 2026 security incident”, security bulletin, updated October 10, 2026. &lt;a href=&quot;https://fireworks.ai/security/bulletins/october-2026-security-incident&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://fireworks.ai/security/bulletins/october-2026-security-incident&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Google Threat Intelligence Group and Mandiant, “The Rise of LLMJacking: How Threat Actors Monetize Compromised Cloud and Model Resources”, Google Cloud blog, November 2, 2026. &lt;a href=&quot;https://cloud.google.com/blog/topics/threat-intelligence/llmjacking-monetizing-compromised-cloud-models&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cloud.google.com/blog/topics/threat-intelligence/llmjacking-monetizing-compromised-cloud-models&lt;/a&gt;&lt;a href=&quot;#fnref13&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Check Point: a Gaia flaw that gives root access is already being exploited</title>
    <link href="https://blog.cisent.ai/en/check-point-una-falla-de-gaia-ya-se-explota/"/>
    <id>tag:cisent.ai,2026-11-13:/en/check-point-una-falla-de-gaia-ya-se-explota/</id>
    <published>2026-11-13T12:00:00+00:00</published>
    <updated>2026-11-13T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA ordered a three-day fix for a flaw in Check Point firewalls and management servers that lets a local user run commands as root.</summary>
    <content type="html">&lt;p&gt;On November 13, 2026, CISA added Check Point’s CVE-2026-123039 to its catalog of exploited vulnerabilities, with a November 16 deadline for U.S. federal agencies to fix it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Check Point says it observed an active exploit against a couple of dozen customers.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It’s in Gaia OS, the operating system of Check Point appliances. Its dynamic CLI is served by a process that runs as root and listens on a local socket.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;That socket is writable by anyone, with no authentication of whoever connects, and the process runs whatever it receives as root.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;As a result, any authenticated local user can run commands as root.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It’s a local privilege escalation: the attacker must already be able to run code on the appliance.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check Point rates it high, with a CVSS score of 7.8.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;For federal agencies, CISA also requires a forensic triage of the devices.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-affected&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is affected&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Check Point Force firewalls, management and log servers, Multi-Domain Security Management, Multi-Domain Log Server and Maestro Orchestrator.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Versions R81, R81.10, R81.20, R82, R82.10 and R82.20.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to the advisory, Smart-1 Cloud is not affected.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do-according-to-check-point&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do, according to Check Point&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Install the Jumbo Hotfix Accumulator that fixes it: for R82.20 from Take 14, for R82.10 from Take 52, for R82 from Take 132 and for R81.20 from Take 174.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If you can’t update yet, stop the dynamic_server process and keep it from restarting, using the commands in the advisory. That stops the dynamic CLI commands from working.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Review the /var/log/dynamicCli.log log: a command that doesn’t use the normal Clish wrapper may mean someone ran commands as root, and needs to be investigated.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;After patching, confirm that the /tmp/dynamic_socket socket has 0660 permissions, not 0777.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;R81 and R81.10 are no longer supported, and the advisory brings no fix for them.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”: CVE-2026-123039, added November 13, 2026, with a deadline of November 16. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Check Point, “CVE-2026-123039 - Local privilege escalation to root in the DynamicCLI process via local world-writable control socket”, sk1000362, Check Point Support Center, updated November 12, 2026. &lt;a href=&quot;https://support.checkpoint.com/results/sk/sk1000362&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://support.checkpoint.com/results/sk/sk1000362&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>The FBI warns about ARTEX, an AI platform that automates entire intrusions</title>
    <link href="https://blog.cisent.ai/en/fbi-alerta-por-artex-intrusiones-automatizadas/"/>
    <id>tag:cisent.ai,2026-11-13:/en/fbi-alerta-por-artex-intrusiones-automatizadas/</id>
    <published>2026-11-13T12:00:00+00:00</published>
    <updated>2026-11-13T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>According to the FBI, criminals and state actors are using ARTEX, an open platform of AI agents, to attack at high speed, mostly through the least-watched systems.</summary>
    <content type="html">&lt;p&gt;On November 13, 2026, the FBI published FLASH alert FLASH-20261113-001 about ARTEX, an open platform for autonomous penetration testing.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to the FBI, various malicious actors have weaponized it to automate several phases of an intrusion “at machine speed,” from reconnaissance through data exfiltration.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-the-fbi-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the FBI says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;ARTEX turns objectives written in natural language into commands carried out by AI agents, with little human involvement once the objective is set. It has no model of its own: it connects to third-party language models.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Since at least August 2026 it has been used against organizations worldwide: financial services, government, technology, defense, semiconductors, energy, healthcare, manufacturing, telecommunications and legal services.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Some intrusions went from initial targeting to privileged access in under three hours.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It lets actors with little technical expertise run complex operations against several targets at once.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Among the campaigns it cites are the attacks on South Korean financial institutions, with personal data of about 67,000 people exfiltrated.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It also cites the supply chain attack on customers of ConnectWise’s backup service, which gave access to more than 2,250 customer environments.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; We covered that attack in October: &lt;a href=&quot;https://blog.cisent.ai/en/connectwise-apaga-sus-respaldos/&quot;&gt;ConnectWise shut down its backup service&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;Its developer took the GitHub repository down in October 2026, but the platform keeps circulating in copies, variants and a commercial version, ARTEX PRO.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In early November there were more than 1,600 ARTEX consoles exposed on the internet, and the FBI believes the total is much higher.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-it-gets-in&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Where it gets in&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;According to the FBI, ARTEX is especially good at finding and exploiting the systems that get less attention: partner and supplier portals, mobile app backends, employee applications, sales-support systems, legacy services and third-party integrations.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Its agents can map the trust relationships between organizations and get in through the weakest vendor to reach the most valuable network.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-the-fbi-recommends&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the FBI recommends&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Inventory every internet-facing application, including peripheral ones.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Restrict administration of the cloud, network devices, backups and other sensitive services to trusted networks or dedicated administrative paths.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Apply least privilege to application permissions, and monitor new credentials and privilege changes.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;On partner and employee portals, alert on sequential lookups and on sessions that access an unusual number of records, and rate-limit requests.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The FBI puts it this way: ARTEX generates a lot of noisy traffic that may trigger alerts, but its speed leaves little time to respond before it reaches its goal.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;FBI, “Threat Overview of ARTEX Autonomous Penetration Testing Framework”, FLASH alert FLASH-20261113-001, November 13, 2026. &lt;a href=&quot;https://www.ic3.gov/CSA/2026/261113.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.ic3.gov/CSA/2026/261113.pdf&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Osaka: an attack on a vendor&#x27;s email service exposed more than a million addresses</title>
    <link href="https://blog.cisent.ai/en/osaka-un-proveedor-de-correo-expuso-un-millon-de-direcciones/"/>
    <id>tag:cisent.ai,2026-11-13:/en/osaka-un-proveedor-de-correo-expuso-un-millon-de-direcciones/</id>
    <published>2026-11-13T12:00:00+00:00</published>
    <updated>2026-11-13T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The Japanese city warned that the email addresses of users of its online services system leaked through an attack on the email delivery service its vendor used.</summary>
    <content type="html">&lt;p&gt;On November 13, 2026, the city of Osaka, in Japan, reported that 1,078,713 email addresses belonging to users of its online services system had leaked.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The cause was unauthorized access to the email delivery service used by TKC, the company that provides that system to the city.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-the-city-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the city says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Those affected are people who received an email from the system between July 1 and October 31, 2026.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The number is so high because on September 16 the system sent a notice to all its registered users.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It warns that suspicious emails may arrive in the name of the city or the system. It asks people not to open links or attachments, not to enter passwords or card details, and not to pay anything requested by email.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It has demanded a detailed report from TKC on the cause, the scope and the measures to keep it from happening again.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-vendors-vendor&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The vendor’s vendor&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;According to The Record, the service that was attacked belongs to Arara, a Japanese email delivery company.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Arara reported that between October 28 and 31, about 77 million email addresses were taken from its customers’ environments: 97 servers belonging to up to about 380 companies. In some cases, names and other member data were taken too.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It notes that this figure is not the number of people, because one person may have several addresses.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Arara, the cause was a web server vulnerability and problems in its system environment.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Up to about 7.13 million suspicious emails were sent from its service.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Arara suspended its whole service on October 31, with no date to resume.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Osaka still doesn’t know the cause or the scope: it has asked its vendor for them in writing.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; builds the report for each incident while it’s still open: the summary for the board, the technical evidence for IT, the backup for the insurer and, when Chile’s Law 21.663 requires it, the reports to ANCI on its deadlines, starting with the early alert within 3 hours. It gathers the evidence and seals it with a hash chain, so that what was reported on day one still holds up two months later.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;City of Osaka, “【大阪市行政オンラインシステム】メール配信サービスへの不正アクセスによるメールアドレスの情報漏えいについて（お詫び）” (on the leak of email addresses through unauthorized access to the email delivery service; in Japanese), November 13, 2026. &lt;a href=&quot;https://www.city.osaka.lg.jp/ictsenryakushitsu/page/0000691479.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.city.osaka.lg.jp/ictsenryakushitsu/page/0000691479.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Jonathan Greig, “City of Osaka warns of data breach impacting 1 million as Japan reels from dozens of cyber incidents”, The Record, November 13, 2026. &lt;a href=&quot;https://therecord.media/city-of-osaka-japan-email-data-breach&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/city-of-osaka-japan-email-data-breach&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Arara, “当社が提供する‘アララメッセ－ジ’への不正アクセスによる情報流出およびサービス停止について（第２報）” (on the information leak and service suspension caused by unauthorized access to its messaging service, second report; in Japanese), November 2, 2026. &lt;a href=&quot;https://www.arara.com/dcms_media/other/arara_20261102__.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.arara.com/dcms_media/other/arara_20261102__.pdf&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Operation Endgame: police take down the infrastructure of StealC, Gootloader and CastleRAT</title>
    <link href="https://blog.cisent.ai/en/operacion-endgame-derriba-tres-malwares/"/>
    <id>tag:cisent.ai,2026-11-12:/en/operacion-endgame-derriba-tres-malwares/</id>
    <published>2026-11-12T12:00:00+00:00</published>
    <updated>2026-11-12T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>German police and international partners shut down the servers of three malware families and secured credentials stolen from their victims; StealC had already been hit in June.</summary>
    <content type="html">&lt;p&gt;On November 12, 2026, the Frankfurt prosecutor’s office that specializes in cybercrime (ZIT) and Germany’s Federal Criminal Police Office (BKA) announced a new action in Operation Endgame, which they describe as the world’s largest police operation against cybercrime.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Between November 9 and 11, together with authorities from the Netherlands, Denmark, the United Kingdom, the United States and Canada, and with support from Europol, Eurojust and private security companies, they took three malware families out of service: Gootloader, CastleRAT and StealC.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-was-taken-down&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was taken down&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;More than 230 servers worldwide were seized or disrupted, more than 40 of them in Germany.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;More than 250 command-and-control domains were taken offline.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;More than 680 compromised websites were cleaned up.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;About 6.4 million login credentials from more than 40,000 victims were secured.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-three-malware-families&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The three malware families&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Gootloader:&lt;/strong&gt; according to German authorities, it has infected millions of computers worldwide since 2020. It spread through search results manipulated to rank at the top, and it was used to load other malware, especially ransomware.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CastleRAT:&lt;/strong&gt; a remote access trojan known since 2025 that lets attackers control the infected computer remotely. It was used to encrypt data, cripple entire systems and pressure companies into paying ransoms.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;StealC:&lt;/strong&gt; a widespread credential stealer that also downloads other malware. Police had already targeted it in June 2026: its operators rebuilt the infrastructure, and the BKA found it again.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Part of that infrastructure ran on Virtualine/Omegatech, a “bulletproof” hosting service that Frankfurt police and the Dutch NHTCU shut down at the end of October.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-comes-next&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What comes next&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;People affected can check on public platforms whether their data is among what was recovered.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The Record mentions two: the Dutch police’s and Have I Been Pwned.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to The Record, the action did not come with arrests, cryptocurrency seizures or charges.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The head of the BKA’s cybercrime division, Carsten Meywirth, warned that anyone who rebuilds their infrastructure after a police action leaves new traces, and that they follow them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The recovered credentials had already been stolen: until their owners change them, they still work for getting in.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Frankfurt Public Prosecutor General’s Office (ZIT) and Federal Criminal Police Office (BKA), “Internationale Strafverfolgungsbehörden setzen weltgrößte Polizeioperation ‘Endgame’ gegen Cybercrime fort” (International law enforcement agencies continue ‘Endgame’, the world’s largest police operation against cybercrime; in German), press release on Presseportal, November 12, 2026. &lt;a href=&quot;https://www.presseportal.de/blaulicht/pm/7/6389442&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.presseportal.de/blaulicht/pm/7/6389442&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Alexander Martin, “European authorities take down SteaIC malware servers for second time this year”, The Record, November 12, 2026. &lt;a href=&quot;https://therecord.media/european-authorities-take-down-stealic-malware-second-time&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/european-authorities-take-down-stealic-malware-second-time&lt;/a&gt;&lt;a href=&quot;#fnref12&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s CSIRT warns of a critical FortiClient flaw that allows remote code execution</title>
    <link href="https://blog.cisent.ai/en/csirt-alerta-por-una-falla-critica-de-forticlient/"/>
    <id>tag:cisent.ai,2026-11-11:/en/csirt-alerta-por-una-falla-critica-de-forticlient/</id>
    <published>2026-11-11T12:00:00+00:00</published>
    <updated>2026-11-11T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Chile&#x27;s national CSIRT urges updating FortiClient on Windows and Mac over a critical flaw in its local API that could let a remote attacker run code or commands.</summary>
    <content type="html">&lt;p&gt;On November 11, 2026, Chile’s national CSIRT published alert AVC26-00750 about a critical vulnerability in Fortinet’s FortiClient for Windows and Mac.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Fortinet had disclosed it the day before, as CVE-2026-121529.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It is in FortiClient’s local API: an error in validating the origin of requests.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Fortinet, an unauthenticated remote attacker could run unauthorized code or commands.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Exploitation requires interaction from the victim.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Fortinet rates it critical, with a CVSS score of 9.1, and says it is not known to be exploited.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It was reported by Kevin Joensen of Baldur Security.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-affected&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is affected&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;FortiClient for Windows, 7.4.0 through 7.4.7 and 7.2.0 through 7.2.14.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;FortiClient for Mac, 7.4.0 through 7.4.7.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;FortiClient Windows PAM, in the same ranges as the Windows version.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Version 8.0 and FortiClient Windows Free VPN are not affected.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update to 7.4.8 or later, or to 7.2.15 or later on the 7.2 branch: these are the versions the CSIRT asks for.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check which computers have FortiClient installed and which version they run: the flaw is in the software on each computer.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;If someone exploits it, the result shows up on the user’s computer: code or commands that FortiClient should not have accepted.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Chile’s national CSIRT, “FortiClient - Vulnerabilidad Crítica” (FortiClient – critical vulnerability; in Spanish), alert AVC26-00750, November 11, 2026. &lt;a href=&quot;https://csirt.gob.cl/alertas/avc26-00750/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/avc26-00750/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Fortinet, “Origin validation error in local API endpoint”, advisory FG-IR-26-194, FortiGuard Labs PSIRT, November 10, 2026. &lt;a href=&quot;https://fortiguard.fortinet.com/psirt/FG-IR-26-194&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://fortiguard.fortinet.com/psirt/FG-IR-26-194&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s cybersecurity agency investigates a cyberattack on the Foreign Ministry&#x27;s consular system</title>
    <link href="https://blog.cisent.ai/en/la-anci-investiga-un-ciberataque-a-la-cancilleria/"/>
    <id>tag:cisent.ai,2026-11-11:/en/la-anci-investiga-un-ciberataque-a-la-cancilleria/</id>
    <published>2026-11-11T12:00:00+00:00</published>
    <updated>2026-11-11T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A news investigation describes an attack on Chile&#x27;s consular services system. The Foreign Ministry confirms forensic analysis and ANCI is investigating, but with a different account of what happened.</summary>
    <content type="html">&lt;p&gt;On November 11, 2026, El Mostrador’s investigative unit reported that Chile’s Ministry of Foreign Affairs suffered a cyberattack that hit its Consular Services System, the platform the consular network uses to handle paperwork abroad.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The ministry confirmed to the outlet that it is carrying out forensic analysis, and the National Cybersecurity Agency (ANCI) confirmed it is investigating.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-the-ministry-and-anci-confirm&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the ministry and ANCI confirm&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The ministry said it is “carrying out forensic analyses to assess the incidents mentioned”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It added that its systems are monitored at all times and that every episode is reported through the appropriate government channels.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;ANCI confirmed it is investigating the breach at the ministry and that, so far, it suspects a distributed denial-of-service (DDoS) attack.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;ANCI said it has not been informed of an attack with malicious code like the one the investigation describes, and that fines are set once the investigation ends.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-the-investigations-sources-say&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the investigation’s sources say&lt;/h2&gt;
&lt;p&gt;Neither the ministry nor ANCI has confirmed any of this:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;the attack would have started about two weeks before the report, with a phishing email to a member of the diplomatic corps that was opened inside the ministry;&lt;/li&gt;
&lt;li&gt;the malicious code would replicate every 10 minutes, so it has not been removed without shutting the system down;&lt;/li&gt;
&lt;li&gt;the system has not been shut down because more than a million Chileans abroad would lose access to their paperwork;&lt;/li&gt;
&lt;li&gt;it would have mainly affected the visa records;&lt;/li&gt;
&lt;li&gt;the origin of the attack has not been identified.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-is-still-unknown&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What is still unknown&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Which of the two accounts describes what happened: the ministry speaks of “the incidents”, plural.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Whether data was taken, and which.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;As of the report, its sources and ANCI described different attacks: one with malicious code and one a denial of service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; builds the report for each incident while it’s still open: the summary for the board, the technical evidence for IT, the backup for the insurer and, when Chile’s Law 21.663 requires it, the reports to ANCI on its deadlines, starting with the early alert within 3 hours. It gathers the evidence and seals it with a hash chain, so that what was reported on day one still holds up two months later.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Claudio Pizarro Sanguesa, “Ciberataque al sistema de visas de Cancillería abre ventana a crimen organizado transnacional” (Cyberattack on the Foreign Ministry’s visa system opens a window to transnational organized crime; in Spanish), El Mostrador investigative unit, November 11, 2026. &lt;a href=&quot;https://www.elmostrador.cl/unidad-de-investigacion/2026/11/11/ciberataque-al-sistema-de-visas-de-cancilleria-abre-ventana-a-crimen-organizado-transnacional/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.elmostrador.cl/unidad-de-investigacion/2026/11/11/ciberataque-al-sistema-de-visas-de-cancilleria-abre-ventana-a-crimen-organizado-transnacional/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>ScreenConnect: attackers used its panels to break into dozens of organizations</title>
    <link href="https://blog.cisent.ai/en/screenconnect-atacantes-usaron-sus-paneles/"/>
    <id>tag:cisent.ai,2026-11-11:/en/screenconnect-atacantes-usaron-sus-paneles/</id>
    <published>2026-11-11T12:00:00+00:00</published>
    <updated>2026-11-11T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Huntress saw an attacker use ScreenConnect panels to install their own remote access on hundreds of computers, and ConnectWise points to third-party extensions.</summary>
    <content type="html">&lt;p&gt;On November 11, 2026, security firm Huntress saw an attacker use access to several panels of ScreenConnect, ConnectWise’s remote support tool, to send commands to groups of managed computers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The same day, ConnectWise published an advisory.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; In it, the company says it is investigating unauthorized access to the ScreenConnect environments of some of its partners, the IT providers that use it to manage their clients’ computers, and that it disabled several third-party extensions as a precaution.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-huntress-saw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Huntress saw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Starting at about 17:47 UTC on November 11, and within 90 minutes, at least 288 computers at 48 organizations that use Huntress were affected. A third were hit in the first 10 minutes.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The commands installed a ScreenConnect client controlled by the attacker and a toolkit Huntress named SCKit.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;SCKit is built to spread: from the infected computer to the technician who connects to it, and from there to the other computers that technician can reach. Huntress says it has not seen that spread succeed so far.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The activity appears to be related to RMM+, a third-party extension that lets another tool drive ScreenConnect sessions. RMM+ bypasses ScreenConnect’s accounts, roles and multi-factor authentication: it authenticates with a shared key stored in its settings.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Its versions 1.0.31 and 1.0.32 accepted an empty key if no one had set one.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-connectwise-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What ConnectWise says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It has found no evidence that its infrastructure was compromised, or that initial access used a ScreenConnect vulnerability.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Its initial assessment is that the attacker may have abused third-party extensions installed in its partners’ environments. It says it is still validating that assessment.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It found about 400 environments with at least one extension from the RMM+ family, and notes that having one installed does not mean being compromised.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It disabled five extensions in those cloud environments: Labtech Plus, RMM Plus, RMM+, RMM Plus Password Link and RMM+ Password Link.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It says version 26.6.5 of the ScreenConnect client, from September, blocks the technique SCKit uses to try to reach the technician’s computer.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do-according-to-connectwise&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do, according to ConnectWise&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update every technician client to 26.6.5 or later, in the cloud and on premises: updating the console does not update those clients.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Keep the listed extensions disabled. Anyone running ScreenConnect on their own servers has to disable them themselves.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;From a trusted device, end all ScreenConnect web sessions and identity-provider sessions, reset credentials and enforce multi-factor authentication. Changing a password does not end open sessions.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Don’t use ScreenConnect to investigate or clean the affected computers; use the EDR or another trusted channel, and include the technicians’ computers that connected to them in the review.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Review the authentication, extension installation, Run Command and Toolbox records.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-is-still-unknown&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What is still unknown&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Whether the way in really was an extension: ConnectWise says it is still validating that assessment.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;How many organizations were hit in total: the 48 are only those that use Huntress.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref20&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Huntress, most of the RMM+ installations it found were legitimate: they were usually added for a project and left in place.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref21&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Huntress, “SCKit Worm Installs Rogue ScreenConnect Clients and Session Cookie Relays Across Hundreds of Endpoints”, Huntress blog, November 12, 2026. &lt;a href=&quot;https://www.huntress.com/blog/sckit-worm-rogue-screenconnect&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.huntress.com/blog/sckit-worm-rogue-screenconnect&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;ConnectWise, “ScreenConnect security advisory: Precautionary extension disabling and partner guidance”, ConnectWise Trust Center, updated November 12, 2026. &lt;a href=&quot;https://www.connectwise.com/company/trust/advisories&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.connectwise.com/company/trust/advisories&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Barracuda: two chained flaws can hand over control of its appliances, and they&#x27;re already exploited</title>
    <link href="https://blog.cisent.ai/en/barracuda-dos-fallas-encadenadas-ya-se-explotan/"/>
    <id>tag:cisent.ai,2026-11-10:/en/barracuda-dos-fallas-encadenadas-ya-se-explotan/</id>
    <published>2026-11-10T12:00:00+00:00</published>
    <updated>2026-11-10T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA ordered urgent patching of two Barracuda flaws that, chained together, can give an attacker control of its appliances from the admin panel, and that are already being exploited.</summary>
    <content type="html">&lt;p&gt;On November 10, 2026, CISA added two Barracuda flaws, CVE-2026-116606 and CVE-2026-111560, to its catalog of exploited vulnerabilities, and gave US federal agencies until November 13 to fix them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The night before, Barracuda had updated its advisory to say it had seen attacks exploiting them against a small number of organizations.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaws&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaws&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;CVE-2026-116606 is in the Email Security Gateway, the appliance that filters email. An attacker without an account can bypass authentication and restore a crafted backup that changes the appliance’s global configuration.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Barracuda rates it 9.3 on CVSS.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;CVE-2026-111560 is a command injection in the time zone field: an administrator can run commands as root.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; It affects the Email Security Gateway, the ADC load balancer, the Message Archiver, the Web Application Firewall, its WAF Control Center console and the Web Security Gateway, and it’s rated 8.6.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Barracuda, together they could let an attacker bypass access controls and get root on the appliance.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;To do that, the attacker has to be able to reach the administrative interface.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Check that every appliance runs the latest firmware and has Barracuda’s patches. If automatic updates are turned off, the patches have to be applied manually, and an appliance that is far behind may need more than one update.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Keep the administrative interface on trusted internal networks only: according to Barracuda, organizations that don’t expose it to the internet are less exposed.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Here the way in is the admin panel, when it can be reached from the internet.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”: CVE-2026-116606 and CVE-2026-111560, added on November 10, 2026, with a deadline of November 13. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Barracuda Networks, “Pre-Authentication Authorization Bypass and Command Injection Vulnerabilities in Certain Barracuda Products”, security advisory, November 2, 2026, updated November 9, 2026. &lt;a href=&quot;https://trust.barracuda.com/security/pre-authentication-authorization-bypass-command-injection&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://trust.barracuda.com/security/pre-authentication-authorization-bypass-command-injection&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Mandiant, “MNDT-2026-0042”, vulnerability disclosure, November 2, 2026. &lt;a href=&quot;https://github.com/mandiant/Vulnerability-Disclosures/blob/master/2026/MNDT-2026-0042.md&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/mandiant/Vulnerability-Disclosures/blob/master/2026/MNDT-2026-0042.md&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Mandiant, “MNDT-2026-0041”, vulnerability disclosure, November 2, 2026. &lt;a href=&quot;https://github.com/mandiant/Vulnerability-Disclosures/blob/master/2026/MNDT-2026-0041.md&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/mandiant/Vulnerability-Disclosures/blob/master/2026/MNDT-2026-0041.md&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s CSIRT warns of a critical Apple flaw that allows escaping WebKit&#x27;s sandbox</title>
    <link href="https://blog.cisent.ai/en/csirt-alerta-por-una-falla-critica-de-apple/"/>
    <id>tag:cisent.ai,2026-11-10:/en/csirt-alerta-por-una-falla-critica-de-apple/</id>
    <published>2026-11-10T12:00:00+00:00</published>
    <updated>2026-11-10T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Chile&#x27;s national CSIRT asks to update iPhone, iPad, Mac, Vision Pro and Safari over a WebKit flaw that, according to Apple, could let a remote attacker break out of the web content sandbox.</summary>
    <content type="html">&lt;p&gt;On November 10, 2026, Chile’s national CSIRT published alert AVC26-00748 about a critical Apple vulnerability, CVE-2026-113897, that could allow a remote attacker to escape the sandbox.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Apple had fixed it on October 26, in the 27.2 releases of iOS, iPadOS, macOS Golden Gate, visionOS and Safari.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It’s in WebKit, Safari’s web engine, in its process model.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Apple, a remote attacker may be able to break out of the web content sandbox, the isolated area where whatever a page does runs.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Apple fixed it by removing the vulnerable code, and its note doesn’t say it has been exploited.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The CSIRT rates it 10.0 on CVSS, the highest score.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It was reported by Luke Francis and Jackson Kuja.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-affected&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is affected&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;iPhone 11 and later, and the iPads Apple lists, on versions earlier than iOS and iPadOS 27.2.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Macs on macOS Golden Gate earlier than 27.2.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Safari on macOS Sequoia and macOS Tahoe, before 27.2.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Every Apple Vision Pro model, before visionOS 27.2.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update to iOS or iPadOS 27.2, macOS Golden Gate 27.2, visionOS 27.2 or, on Sequoia and Tahoe, Safari 27.2: those are the versions the CSIRT asks for.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check which of the organization’s devices are still on earlier versions: the fix has been out since October 26.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;A flaw like this only closes on the devices that were updated: the one left behind stays exposed.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Chile’s national CSIRT, “Apple - Vulnerabilidad Crítica” (Apple – critical vulnerability; in Spanish), alert AVC26-00748, November 10, 2026. &lt;a href=&quot;https://csirt.gob.cl/alertas/avc26-00748/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/avc26-00748/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Apple, “About the security content of iOS 27.2 and iPadOS 27.2”, security note, October 26, 2026. &lt;a href=&quot;https://support.apple.com/en-us/149230&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://support.apple.com/en-us/149230&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Apple, “About the security content of macOS Golden Gate 27.2”, security note, October 26, 2026. &lt;a href=&quot;https://support.apple.com/en-us/149521&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://support.apple.com/en-us/149521&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Apple, “About the security content of Safari 27.2”, security note, October 26, 2026. &lt;a href=&quot;https://support.apple.com/en-us/149527&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://support.apple.com/en-us/149527&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;Apple, “About the security content of visionOS 27.2”, security note, October 26, 2026. &lt;a href=&quot;https://support.apple.com/en-us/149526&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://support.apple.com/en-us/149526&lt;/a&gt;&lt;a href=&quot;#fnref11&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Google fixes a Chrome flaw that is already being exploited</title>
    <link href="https://blog.cisent.ai/en/google-una-falla-de-chrome-ya-se-explota/"/>
    <id>tag:cisent.ai,2026-11-10:/en/google-una-falla-de-chrome-ya-se-explota/</id>
    <published>2026-11-10T12:00:00+00:00</published>
    <updated>2026-11-10T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Google updated Chrome for desktop and warned that an exploit is already in use for a memory flaw in its XML component, which could also affect Edge and Opera.</summary>
    <content type="html">&lt;p&gt;On November 9, 2026, Google released a Chrome update for desktop with 9 security fixes, and said it is aware that an exploit for one of them, CVE-2026-125350, exists in the wild.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The next day, CISA added it to its catalog of exploited vulnerabilities.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It is a “use after free” flaw, the use of a memory region that has already been freed, in the XML component.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to CISA, a remote attacker could run code inside the browser’s sandbox with a crafted web page.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It could affect other Chromium-based browsers, such as Microsoft Edge and Opera.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Google rates it high. It was reported by Arni Hardarson of Neonix Security on October 12, and by an anonymous reporter on November 6.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Google is keeping the details restricted until most users have the fix.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update Chrome to version 157.0.8103.50 or .51 on Windows and Mac, and to 157.0.8103.50 on Linux.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check that every computer has it: Google is rolling it out gradually, over the coming days or weeks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In Edge, Opera and other Chromium-based browsers, apply the update each vendor releases.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;CISA gave US federal agencies until November 24.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Google is rolling out the fix gradually: in the meantime, any computer that hasn’t been updated stays exposed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Google, “Stable Channel Update for Desktop”, Chrome Releases blog, November 9, 2026. &lt;a href=&quot;https://chromereleases.googleblog.com/2026/11/stable-channel-update-for-desktop_018223192.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://chromereleases.googleblog.com/2026/11/stable-channel-update-for-desktop_018223192.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”: CVE-2026-125350, added on November 10, 2026, with a deadline of November 24. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Microsoft fixes more than a thousand flaws, and one in Outlook is already exploited to impersonate users</title>
    <link href="https://blog.cisent.ai/en/microsoft-una-falla-de-outlook-ya-se-explota/"/>
    <id>tag:cisent.ai,2026-11-10:/en/microsoft-una-falla-de-outlook-ya-se-explota/</id>
    <published>2026-11-10T12:00:00+00:00</published>
    <updated>2026-11-10T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Microsoft&#x27;s November patches include an Outlook flaw that is already being exploited: opening a crafted file hands the attacker what it needs to impersonate the user.</summary>
    <content type="html">&lt;p&gt;On November 10, 2026, Microsoft released its November security updates: more than a thousand vulnerabilities fixed, 68 of them rated critical.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Only one is marked as exploited, CVE-2026-124194, in Outlook.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; CISA added it to its catalog of exploited vulnerabilities the same day,&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; and Chile’s national CSIRT published alert AVC26-00749 with the month’s updates.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-outlook-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The Outlook flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;An attacker sends a crafted file and convinces the user to open it in Outlook.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;When it’s opened, Outlook connects to a network location controlled by the attacker without showing the usual external-content warning.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;That connection hands the attacker the user’s NTLMv2 hash, which can be used to authenticate as that user.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The preview pane isn’t enough for the attack: the user has to open the file.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Microsoft rates it moderate, with a CVSS score of 6.5, but it’s the only one this month that it marks as exploited.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Apply the November updates, as Chile’s national CSIRT asks,&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; starting with Office and Outlook. CISA gave US federal agencies until November 24.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Be wary of Office files that arrive unrequested: the attack needs the user to open one.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;In this attack, the signal is an outbound connection: the user’s computer connects to a network location controlled by the attacker.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Microsoft, “November 2026 Security Updates”, release notes, November 10, 2026. &lt;a href=&quot;https://msrc.microsoft.com/update-guide/releaseNote/2026-Nov&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://msrc.microsoft.com/update-guide/releaseNote/2026-Nov&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Microsoft, “Microsoft Outlook Spoofing Vulnerability”, CVE-2026-124194, Security Update Guide, November 10, 2026. &lt;a href=&quot;https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-124194&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://msrc.microsoft.com/update-guide/vulnerability/CVE-2026-124194&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”: CVE-2026-124194, added on November 10, 2026, with a deadline of November 24. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Chile’s national CSIRT, “Microsoft Update Tuesday 2026 Noviembre - Vulnerabilidad Crítica” (Microsoft Update Tuesday, November 2026 – critical vulnerability; in Spanish), alert AVC26-00749, November 10, 2026. &lt;a href=&quot;https://csirt.gob.cl/alertas/avc26-00749/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/avc26-00749/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>France blames Iran for sabotaging the refrigeration at 700 supermarkets in October</title>
    <link href="https://blog.cisent.ai/en/francia-acusa-a-iran-de-sabotear-los-refrigeradores/"/>
    <id>tag:cisent.ai,2026-11-09:/en/francia-acusa-a-iran-de-sabotear-los-refrigeradores/</id>
    <published>2026-11-09T12:00:00+00:00</published>
    <updated>2026-11-09T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>France&#x27;s cyber crisis coordination center attributes October&#x27;s sabotage of refrigeration controllers at hundreds of supermarkets to a unit of Iran&#x27;s Revolutionary Guard.</summary>
    <content type="html">&lt;p&gt;On November 9, 2026, France attributed to Iran the attack that, overnight on October 11 to 12, sabotaged the refrigeration controllers of about 700 supermarkets. Its Ministry for Europe and Foreign Affairs said so in a statement,&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; and a report by the Cyber Crisis Coordination Centre (C4) gives the details. The C4 brings together the cybersecurity agency ANSSI, the cyber defense command, the armament agency DGA and the DGSE and DGSI intelligence services.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a href=&quot;https://blog.cisent.ai/en/supermercados-refrigeradores-europa/&quot;&gt;We covered the attack on October 12&lt;/a&gt;.&lt;/p&gt;
&lt;section id=&quot;what-the-report-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the report says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;What happened:&lt;/strong&gt; overnight from Sunday, October 11, to Monday, October 12, the attackers got into the refrigeration controllers of about 700 large retail sites and changed their defrost settings to high temperatures. There were significant losses of goods and, in at least 15 cases, fires breaking out.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Who was hit:&lt;/strong&gt; mostly large supermarket chains, and also small butcher shops, a slaughterhouse and cooperative food plants.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Not just refrigeration:&lt;/strong&gt; as part of the same attack, they targeted Schneider Electric equipment at small hydroelectric plants in France, with moderate effects.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Who did it:&lt;/strong&gt; the Shahid Kaveh unit, part of the Electronic and Cyber Center of Iran’s Revolutionary Guard. Online it poses under hacktivist names such as CyberAveng3rs, and its attacks on industrial equipment have mostly targeted Israel and the United States.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;how-they-reached-that-conclusion&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they reached that conclusion&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The attack infrastructure was used in the same period to scout Schneider Electric equipment at small energy and water plants, with a technique the unit has used since at least April 2026: the Modbus protocol, on port 502.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;For the attack they used the FOFA scanning tool, a search engine for internet-connected devices that the unit uses in its operations.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;By October 13, the attack was being promoted on CyberbanNews, the propaganda outlet tied to the unit.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Traces of Farsi were found in the attack infrastructure.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The report also attributes earlier attacks on French interests to another unit of the same Guard, Shahid Shoostari: against the weekly Charlie Hebdo in 2023 and against a French video display company during the Paris 2024 Olympic Games.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-france-will-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What France will do&lt;/h2&gt;
&lt;p&gt;According to its government, it is the first campaign carried out purely for sabotage on French soil, and food losses reach several tens of millions of euros. France will summon Iran’s ambassador and will propose to its European partners new sanctions against the Iranian entities and individuals involved.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-the-report-doesnt-cover&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the report doesn’t cover&lt;/h2&gt;
&lt;p&gt;The report covers the October attack. It doesn’t mention the attacks of early November, such as the one on &lt;a href=&quot;https://blog.cisent.ai/en/efs-camaras-de-frio-a-110-grados/&quot;&gt;the cold rooms of France’s blood service&lt;/a&gt;.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Keep every industrial controller, for refrigeration, energy or water, out of sight on the internet: the attackers looked for their targets with a scanning tool.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Limit who can talk Modbus to that equipment: it’s the protocol they used to scout it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Follow what five U.S. agencies asked for on November 5: &lt;a href=&quot;https://blog.cisent.ai/en/eeuu-pide-blindar-los-sistemas-de-frio/&quot;&gt;take refrigeration controllers off the public internet&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;To find the controllers, all the attackers needed was a scanning tool.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;French Ministry for Europe and Foreign Affairs, “Déclaration d’attribution à l’Iran de cyberattaques visant la France” (statement attributing cyberattacks against France to Iran; in French), statement, November 9, 2026. &lt;a href=&quot;https://www.diplomatie.gouv.fr/fr/presse-et-ressources/decouvrir-et-informer/actualites/declaration-d-attribution-a-l-iran-de-cyber-attaques-visant-la-france&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.diplomatie.gouv.fr/fr/presse-et-ressources/decouvrir-et-informer/actualites/declaration-d-attribution-a-l-iran-de-cyber-attaques-visant-la-france&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;France’s Cyber Crisis Coordination Centre (C4), “Activités cyber offensives menées par des acteurs iraniens liés au CGRI” (offensive cyber activities by Iranian actors tied to the IRGC; in French), report published by CERT-FR (CERTFR-2026-CTI-009), November 9, 2026. &lt;a href=&quot;https://cert.cyber.gouv.fr/cti/CERTFR-2026-CTI-009/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cert.cyber.gouv.fr/cti/CERTFR-2026-CTI-009/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Data on up to 7.7 million Iris Plaza customers may have been exposed in a cloud service</title>
    <link href="https://blog.cisent.ai/en/iris-plaza-datos-de-millones-de-clientes-expuestos/"/>
    <id>tag:cisent.ai,2026-11-09:/en/iris-plaza-datos-de-millones-de-clientes-expuestos/</id>
    <published>2026-11-09T12:00:00+00:00</published>
    <updated>2026-11-09T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The online store of Japanese manufacturer Iris Ohyama found suspicious files in a cloud service it uses: data on millions of people may have been viewed or taken.</summary>
    <content type="html">&lt;p&gt;On November 5, 2026, Iris Plaza, the company that runs the official online store of Japanese manufacturer Iris Ohyama, published its second notice about unauthorized access: personal data on up to about 7.7 million people may have been exposed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;On October 28&lt;/strong&gt; it found suspicious files in an external cloud service it uses. It immediately shut down the store and its related systems, and started investigating.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;It confirmed access by a third party&lt;/strong&gt; and issued a first notice on October 29. That same day it reported the incident to the police and to Japan’s Personal Information Protection Commission.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;What that service held&lt;/strong&gt; was personal data on about 7.7 million people, which may have been viewed or taken. How many were actually leaked is still under investigation.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;As of November 5 it had no reports of fraudulent use or financial losses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;which-data&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Which data&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Members, including those who had closed their accounts (about 5.8 million):&lt;/strong&gt; name, address, phone, email, gender, password and date of birth, among others. There were no passwords for those who had closed their accounts, and some items, such as passwords, were encrypted.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;People who bought without registering or took part in campaigns (about 1.9 million):&lt;/strong&gt; name, address, phone and email.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Not every customer had all of that data on file.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Not credit cards:&lt;/strong&gt; they weren’t stored in that cloud service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;As a precaution, it also shut down another store in the group and the home-appliance subscription and delivery services. It hasn’t found any access or leak in them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;If you used your Iris Plaza password on another service, change it: that’s what the company asks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Be wary of emails, texts, calls, websites and social media accounts claiming to be the company. It warns they may appear, and that it never asks for passwords, PINs or verification codes.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;For companies: know which customer data lives in third-party services, and for how many people, including those who have left. Here there was data on members who had closed their accounts.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Iris Plaza issued its first notice the day after finding the files, reported to the police and the regulator, and a week later published the second one with what it had learned.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; builds the report for each incident while it’s still open: the summary for the board, the technical evidence for IT, the backup for the insurer and, when Chile’s Law 21.663 requires it, the reports to ANCI on its deadlines, starting with the early alert within 3 hours. It gathers the evidence and seals it with a hash chain, so that what was reported on day one still holds up two months later.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Iris Plaza (Iris Ohyama group), “（第2報）株式会社アイリスプラザECサイト一時停止のお知らせ” (second notice: temporary suspension of Iris Plaza’s e-commerce site; in Japanese), statement, November 5, 2026. &lt;a href=&quot;https://www.irisohyama.co.jp/news/2026/?date=1105&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.irisohyama.co.jp/news/2026/?date=1105&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>OpenSSH: CISA confirms the password bypass on Linux with musl is being exploited</title>
    <link href="https://blog.cisent.ai/en/openssh-la-cisa-confirma-que-ya-se-explota/"/>
    <id>tag:cisent.ai,2026-11-09:/en/openssh-la-cisa-confirma-que-ya-se-explota/</id>
    <published>2026-11-09T12:00:00+00:00</published>
    <updated>2026-11-09T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA added the OpenSSH flaw on Linux with musl and no PAM, fixed on November 5, to its catalog of exploited vulnerabilities: an unpatched server lets attackers in without a password.</summary>
    <content type="html">&lt;p&gt;On November 9, 2026, CISA added to its catalog of exploited vulnerabilities an OpenSSH flaw on Linux that allows authentication to be bypassed, and registered it as CVE-2026-123158. The entry links the OpenSSH 10.6p2 release notes and Alpine Linux’s advisory of November 5.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It’s the flaw that &lt;a href=&quot;https://blog.cisent.ai/en/openssh-arreglo-de-emergencia-musl/&quot;&gt;OpenSSH fixed urgently that day&lt;/a&gt;: on systems with the musl library and no PAM, attackers could log in without a password.&lt;/p&gt;
&lt;section id=&quot;what-cisa-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What CISA says&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It added the flaw to its catalog of exploited vulnerabilities, the one that lists vulnerabilities with confirmed real-world attacks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It gave U.S. federal agencies until November 12 to fix it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It doesn’t know whether ransomware groups are using it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-exposed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is exposed&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Servers that run sshd without PAM and with password authentication enabled, which is the OpenSSH default.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;As long as they have an account without a password, like key-only accounts, and at least one other local account with a password.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The attacker only needs to reach the SSH port: they get in with a single connection, sending a very long password, without knowing it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Not exposed: containers that don’t run sshd and systems with password authentication disabled.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Alpine, the flaw is in every Alpine release it checked, from 3.16 to 3.24.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update OpenSSH to each system’s fixed version and restart sshd afterwards: according to Alpine, upgrading the package doesn’t replace the service that’s already running.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If you can’t update yet, disable password authentication (&lt;code&gt;PasswordAuthentication no&lt;/code&gt;) and restart sshd.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Find every server and container running sshd on musl: with the flaw in CISA’s catalog, it’s no longer a theoretical risk.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Whether a server is exposed depends on how it’s set up: whether it runs sshd without PAM and whether it accepts passwords.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”, entry CVE-2026-123158, added November 9, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Alpine Linux, “Security advisory: openssh-server password authentication bypass”, security advisory, November 5, 2026. &lt;a href=&quot;https://alpinelinux.org/posts/2026-11-05-openssh-password-authentication-bypass/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://alpinelinux.org/posts/2026-11-05-openssh-password-authentication-bypass/&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Zyxel: the June patch didn&#x27;t close the flaw being used against its switches</title>
    <link href="https://blog.cisent.ai/en/zyxel-el-parche-de-junio-no-tapaba-la-falla/"/>
    <id>tag:cisent.ai,2026-11-09:/en/zyxel-el-parche-de-junio-no-tapaba-la-falla/</id>
    <published>2026-11-09T12:00:00+00:00</published>
    <updated>2026-11-09T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The exploit used against Zyxel GS1900 switches since August targeted a different flaw than everyone thought: the June patch didn&#x27;t stop it, and the real fix came out on October 22.</summary>
    <content type="html">&lt;p&gt;On November 3, 2026, Zyxel published an advisory for a flaw in its GS1900 switches that lets an attacker on the local network run commands without credentials: CVE-2026-107441.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; That same day, security firm Binarly, which reported it, explained why it matters: it was the flaw the exploit used against these devices actually targeted, and the June patch didn’t close it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;June:&lt;/strong&gt; Zyxel fixed a different GS1900 flaw, CVE-2026-7273.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;August:&lt;/strong&gt; according to GreyNoise, around August 17 a new exploit took data from 996 switches in 48 countries, including their configurations and hashed root credentials.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;September:&lt;/strong&gt; Binarly found that exploit in the JITTERLY malware’s infrastructure. On the 21st, GreyNoise tied it to the June flaw and CISA added that flaw to its catalog of exploited vulnerabilities. On the 28th, Binarly confirmed that the ten releases fixing the June flaw were still vulnerable: the exploit targeted another flaw.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October and November:&lt;/strong&gt; Zyxel confirmed the new flaw on October 11 and released the fix on the 22nd.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It published the advisory on November 3.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;According to Binarly, CISA’s alert protected no one from this attacker: anyone who updated to the June release stayed exposed.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It’s an out-of-bounds write in the switch’s CGI program. Anyone on the local network, without credentials, can run operating system commands with a crafted HTTP request.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It affects ten GS1900 models, from the GS1900-8 to the GS1900-48HPv2. On the GS1900-8, for example, the vulnerable version is 2.90(AAHH.2)C0 and earlier, and the fixed one is 2.90(AAHH.3)C0.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Binarly believes the exploit was generated by a language model.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update every GS1900 to its model’s fixed version, the one with “.3”,&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; even if it already has the June patch: that patch doesn’t stop this exploit.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Keep the switches’ management out of reach of any machine on the network: the flaw is exploited from the local network and without credentials.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;If a GS1900 was exposed, change its credentials and review its configuration: the exploit took both.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Don’t treat a device as protected just because its version says it’s patched. That’s Binarly’s conclusion: advisories and version numbers help, but they can’t be trusted blindly.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;With the GS1900, a switch with the June patch showed up as patched and was still open to the attack.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Zyxel, “Zyxel security advisory for out-of-bounds write vulnerability in GS1900 series switches”, security advisory, November 3, 2026. &lt;a href=&quot;https://www.zyxel.com/global/en/support/security-advisories/zyxel-security-advisory-for-out-of-bounds-write-vulnerability-in-gs1900-series-switches-11-03-2026&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.zyxel.com/global/en/support/security-advisories/zyxel-security-advisory-for-out-of-bounds-write-vulnerability-in-gs1900-series-switches-11-03-2026&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Fernando Mercês and Yegor Vasilenko, Binarly, “Pre-Auth RCE in Zyxel GS1900 Exploited in the Wild as a Zero-Day”, company blog, November 3, 2026. &lt;a href=&quot;https://www.binarly.io/blog/zyxel-gs1900-rce-cve-2026-107441&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.binarly.io/blog/zyxel-gs1900-rce-cve-2026-107441&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;GreyNoise, “GreyNoise Timeline: Kapibala Exploitation”, company timeline, accessed November 9, 2026. &lt;a href=&quot;https://www.greynoise.io/chronicle/gntl-20260929-kapibala-exploitation&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.greynoise.io/chronicle/gntl-20260929-kapibala-exploitation&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chainalysis: the attack using agentic techniques did expose password hashes and tokens for all its users</title>
    <link href="https://blog.cisent.ai/en/chainalysis-ataque-con-tecnicas-agenticas/"/>
    <id>tag:cisent.ai,2026-11-08:/en/chainalysis-ataque-con-tecnicas-agenticas/</id>
    <published>2026-11-08T12:00:00+00:00</published>
    <updated>2026-11-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Blockchain analytics company Chainalysis expanded its notice: the attack it attributes to agentic techniques exposed more access data than it first said.</summary>
    <content type="html">&lt;p&gt;On October 30, 2026, blockchain analytics company Chainalysis reported that an external actor had used “offensive agentic techniques” to get into its systems and internal tools.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It has updated its notice six times since then, and the updates describe a much larger incident than the first notice.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-the-notice-changed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How the notice changed&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;October 30:&lt;/strong&gt; it said there was no indication of access to investigation content, plaintext passwords, MFA secrets or session tokens. Data it called non-sensitive had been accessed: user IDs, names, emails and user roles.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;November 2:&lt;/strong&gt; it took some products offline as a precaution, including Reactor and Data Solutions.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;November 4:&lt;/strong&gt; it reported that on the previous Monday it had discovered that the names, usernames, password hashes and live tokens of all its users had been exposed. In addition, an on-path attack exposed some customers’ passwords, and abnormal activity led it to also take KYT, its transaction monitoring service, offline. It is notifying directly the customers whose KYT accounts appear to have been accessed without authorization.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;What didn’t change:&lt;/strong&gt; it still has no evidence that investigation data was accessed or that its attribution data was manipulated.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;getting-back-online&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Getting back online&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;November 6:&lt;/strong&gt; it began restoring KYT and address screening in phases. To get back in, each customer must provide its trusted IP addresses, and every user must reset their password and rotate their API keys.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;November 7:&lt;/strong&gt; Reactor, KYT and address screening were back online, with enhanced security measures.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;If the organization uses Chainalysis, follow its instructions: reset passwords, rotate API keys and provide trusted IP addresses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Change the password on any other service where the same one was reused.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Be wary of emails asking for a password reset, go straight to the login page, and turn on multi-factor authentication, as the company recommends.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Chainalysis says it is seeing an industry-wide rise in AI-driven cyberattacks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Chainalysis, “Chainalysis Security Incident Information”, company blog, published October 30, 2026 and updated through November 7, 2026. &lt;a href=&quot;https://www.chainalysis.com/blog/security-incident-october-30/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.chainalysis.com/blog/security-incident-october-30/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Attackers diverted funds from French bank Crédit Mutuel Arkéa by getting in through a supplier&#x27;s platform</title>
    <link href="https://blog.cisent.ai/en/credit-mutuel-arkea-atacado-por-un-proveedor/"/>
    <id>tag:cisent.ai,2026-11-08:/en/credit-mutuel-arkea-atacado-por-un-proveedor/</id>
    <published>2026-11-08T12:00:00+00:00</published>
    <updated>2026-11-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>French bank Crédit Mutuel Arkéa confirmed an attack that let the attackers transfer the bank&#x27;s own funds; they got in through the platform of one of its suppliers.</summary>
    <content type="html">&lt;p&gt;On Wednesday, November 4, 2026, French bank Crédit Mutuel Arkéa announced it had suffered an attack that made fraudulent transactions possible.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; That same afternoon it published a second statement answering the press’s questions.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;The way in:&lt;/strong&gt; the attackers went through the platform of one of its suppliers and, from there, reached the bank directly. The statement doesn’t say who the supplier is.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The money:&lt;/strong&gt; they used the bank’s own account to transfer funds to other banks, from where they could convert them into cryptocurrency on specialized platforms. They didn’t go through customer accounts.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The cause:&lt;/strong&gt; according to the bank, it was a sophisticated attack from outside, not an internal human error, and its security systems detected the suspicious activity very quickly. It is working with judicial authorities and the police to establish how they operated.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The response:&lt;/strong&gt; it suspended some transactions as a precaution, which delayed some customers’ transfers. Almost all of them resumed that same morning.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;how-much&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How much&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The press reported 118 million euros. The bank doesn’t confirm that figure: it only validates its order of magnitude.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Part of the funds could be blocked and actions are under way to recover the rest. According to the bank, the amount doesn’t affect its financial soundness.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to the bank, customers don’t lose any money.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Review which suppliers have access to the systems that move money, and with which controls: at Arkéa, the way in was a supplier’s platform.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Watch the movements of the organization’s own accounts too, not only the customers’: the attackers used the bank’s account.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Warn customers about the phishing that usually follows, as the bank did: never hand over a username, password, card number or the code that arrives by SMS, and when in doubt, call your own adviser.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;At Arkéa, the attack didn’t come through the front door but through a supplier connected to the bank.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Crédit Mutuel Arkéa, “Crédit Mutuel Arkéa annonce avoir été victime d’une attaque”, press release, Brest, November 4, 2026. &lt;a href=&quot;https://www.cm-arkea.com/arkea/banque/assurances/pa_38377/fr/04112026-communique-le-credit-mutuel-arkea-annonce-avoir-ete-victime-d-une-cyberattaque&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cm-arkea.com/arkea/banque/assurances/pa_38377/fr/04112026-communique-le-credit-mutuel-arkea-annonce-avoir-ete-victime-d-une-cyberattaque&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Crédit Mutuel Arkéa, “Informations complémentaires suite à l’attaque dont le Crédit Mutuel Arkéa a été victime”, press release, Brest, November 4, 2026. &lt;a href=&quot;https://www.cm-arkea.com/arkea/banque/assurances/pa_38386/en/informations-complementaires-suite-a-l-attaque-dont-le-credit-mutuel-arkea-a-ete-victime&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cm-arkea.com/arkea/banque/assurances/pa_38386/en/informations-complementaires-suite-a-l-attaque-dont-le-credit-mutuel-arkea-a-ete-victime&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>DigiCert: a network change without a security review left customer data exposed</title>
    <link href="https://blog.cisent.ai/en/digicert-un-cambio-de-red-expuso-datos-de-clientes/"/>
    <id>tag:cisent.ai,2026-11-07:/en/digicert-un-cambio-de-red-expuso-datos-de-clientes/</id>
    <published>2026-11-07T12:00:00+00:00</published>
    <updated>2026-11-07T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Certificate authority DigiCert disclosed that an internal system was left open to the internet by mistake and that someone read its customers&#x27; data; certificates were not affected.</summary>
    <content type="html">&lt;p&gt;On Sunday, November 1, 2026, certificate authority DigiCert published the full report of unauthorized access to one of its internal systems: the one that archives the data its customers’ representatives provide to verify their identity. It published it on Mozilla’s Bugzilla, because the rules of Mozilla, Google Chrome, Microsoft and Apple require it to report these incidents.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-it-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How it happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;A network change:&lt;/strong&gt; on February 4, 2025, during an infrastructure migration, the system was moved out of a network that could only be reached through a VPN and became accessible from the internet.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;An open registration:&lt;/strong&gt; an account creation feature meant for other uses was left available on that system by mistake. Those accounts didn’t go through multi-factor authentication and, although they were read-only, they could see the archived data.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A misread warning:&lt;/strong&gt; on June 18, 2026, DigiCert’s team saw outside accounts on the system but concluded they had no access to data. Three days later, its security research program warned that accounts could be created from outside.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The access:&lt;/strong&gt; someone read data between June 27 and July 2, 2026, the day DigiCert fixed the authorization. It confirmed this on October 19, three and a half months later.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-was-seen&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was seen&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Mostly customer contact data: names, business emails and phone numbers, and business addresses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;For a minority of accounts, also a document with a name and address, such as a driver’s license, a passport or a utility bill.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The access could not write, approve validations or issue certificates. According to DigiCert, its certificate services, its customers’ credentials and the certificates it issued were not affected, and customers don’t need to do anything about their certificates.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;DigiCert doesn’t say how many customers were affected. It does say it has already notified them and reported the incident to law enforcement.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-took-so-long-to-see&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it took so long to see&lt;/h2&gt;
&lt;p&gt;DigiCert acknowledges three causes:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;the migration was treated as an infrastructure change rather than a security change, so nobody reviewed what was left exposed;&lt;/li&gt;
&lt;li&gt;the system had no alerts for new accounts created from outside or for request volume;&lt;/li&gt;
&lt;li&gt;its permission model didn’t contemplate accounts created by anyone.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Its report leaves two lessons: a change in network access to an internal system must always go through a security review, and authorization can’t depend on where the system sits on the network.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Treat every network change to an internal system as a security change, with a review and a named owner, as DigiCert did after the incident.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Alert when accounts are created outside the approved workflow or when request volume rises. According to DigiCert, the volume alert it added in October would have fired on the July 1 traffic.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check with an external scan which internal systems respond from the internet: it was the last task DigiCert closed.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;At DigiCert, the system was exposed for almost 17 months, and the access was confirmed three and a half months after it happened.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;DigiCert, full incident report on Mozilla’s Bugzilla, bug 2084460 (“DigiCert: Unauthorized access to business system”), November 1, 2026. &lt;a href=&quot;https://bugzilla.mozilla.org/show_bug.cgi?id=2084460&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://bugzilla.mozilla.org/show_bug.cgi?id=2084460&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;DigiCert, replies on the same Bugzilla bug, November 6 and 7, 2026. &lt;a href=&quot;https://bugzilla.mozilla.org/show_bug.cgi?id=2084460&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://bugzilla.mozilla.org/show_bug.cgi?id=2084460&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Mittwald: an attacker got in through the MCP server that connects AI assistants to its platform</title>
    <link href="https://blog.cisent.ai/en/mittwald-entraron-por-su-servidor-mcp/"/>
    <id>tag:cisent.ai,2026-11-07:/en/mittwald-entraron-por-su-servidor-mcp/</id>
    <published>2026-11-07T12:00:00+00:00</published>
    <updated>2026-11-07T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>German web host Mittwald says an attacker got in through its platform&#x27;s MCP server, went undetected for more than a week and, judging by its speed, likely used AI tools.</summary>
    <content type="html">&lt;p&gt;On November 7, 2026, Robert Meyer, founder and managing director of Mittwald, a German hosting provider for web agencies, explained how an attacker got into its platform: through the company’s MCP server.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; An MCP server is the piece that lets AI assistants connect to a service and work on it.&lt;/p&gt;
&lt;section id=&quot;how-it-happened-according-to-mittwald&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How it happened, according to Mittwald&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;October 21:&lt;/strong&gt; that evening, while the industry was busy with &lt;a href=&quot;https://blog.cisent.ai/en/php-dos-fallas-criticas/&quot;&gt;PHP2Shell, the two critical PHP flaws&lt;/a&gt;, the attacker first tested a flaw in the MCP server.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;From October 23:&lt;/strong&gt; they moved around the platform like any other user, registering test accounts, using extensions and making simple injection attempts.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 27:&lt;/strong&gt; at 14:58 they registered another account, and at 15:03 they exploited the flaw in the MCP server. Mittwald treats all data from that day on as compromised.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;November 4:&lt;/strong&gt; in the morning, the company noticed unusual behavior in one of its systems and at 11:11 blocked the attacker’s first account. It took parts of its systems offline as a precaution, identified the MCP server as the entry point and shut it down.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-it-found&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it found&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;56 compromised customer projects.&lt;/strong&gt; After scanning its systems repeatedly, it found no others, and when it published its post it saw no attacker activity.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;An attacker who hid:&lt;/strong&gt; they disguised their access and carried out diversions.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;And very fast:&lt;/strong&gt; according to its logs, sometimes only seconds passed between steps that would take a person much longer. Mittwald believes this points to the use of AI tools; its post gives no other evidence.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;What it shut:&lt;/strong&gt; it blocked, as a precaution, the projects with traces of the attacker, which remain blocked, and temporarily suspended new account registration, login to mStudio, its control panel, and SSH access to customer projects. The panel and SSH are back.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It is working with outside forensic experts, has informed the authorities and will publish a detailed account.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Treat an MCP server, or any door for AI assistants, as one more way into your systems, with the same controls as the rest: at Mittwald it was the attacker’s door.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Watch new accounts and what they test: before attacking, the attacker spent days registering accounts and making attempts.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Know how long it takes to notice an intruder: at Mittwald, from October 27 to November 4.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;At Mittwald, the attacker was inside from October 27 until the company noticed something was off, on November 4.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Robert Meyer, Mittwald, “Sicherheitsvorfall am MCP-Server” (security incident at the MCP server; in German), company blog, November 7, 2026. &lt;a href=&quot;https://www.mittwald.de/blog/mittwald/sicherheitsvorfall-am-mcp-server&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.mittwald.de/blog/mittwald/sicherheitsvorfall-am-mcp-server&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>SonicWall fixes a VPN flaw that can bypass multi-factor authentication, and it&#x27;s already being exploited</title>
    <link href="https://blog.cisent.ai/en/sonicwall-falla-de-la-vpn-salta-el-doble-factor/"/>
    <id>tag:cisent.ai,2026-11-07:/en/sonicwall-falla-de-la-vpn-salta-el-doble-factor/</id>
    <published>2026-11-07T12:00:00+00:00</published>
    <updated>2026-11-07T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA added to its catalog of exploited flaws a vulnerability in the VPN of SonicWall firewalls that can bypass SAML authentication and its multi-factor check.</summary>
    <content type="html">&lt;p&gt;On November 6, 2026, SonicWall published an advisory that fixes seven vulnerabilities in SonicOS, the operating system of its firewalls.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; That same day, CISA added one of them, CVE-2026-19868, to its catalog of exploited vulnerabilities.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw-that-is-already-being-exploited&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw that is already being exploited&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It’s in the SonicOS SSL VPN, and it only matters if VPN logins use SAML.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;A user with valid credentials for an alternate authentication method can bypass the SAML controls and, with them, the multi-factor authentication that SAML enforces.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;SonicWall says it investigated a report of potential exploitation and urges customers to upgrade as soon as possible. It offers no workaround.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;CISA treats it as exploited and gave US federal agencies until November 20.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;affected-devices&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Affected devices&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Gen7 firewalls (TZ, NSa, NSsp and the NSv virtual appliances) running SonicOS 7.3.3-7015 or older.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Gen8 firewalls (TZ, NSa and NSv) running SonicOS 8.2.2-8015 or older.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-other-six-flaws&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The other six flaws&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The two most severe, rated 8.8 out of 10, are in the management interface: under specific conditions, they could allow a remote attacker without credentials to perform unauthorized operations.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Three let an unauthenticated attacker knock the SSL VPN or the TACACS+ client out of service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;The last one, on the virtual appliances, exposes a platform credential to anyone who already has local access.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Upgrade SonicOS to the latest release as soon as possible, as SonicWall urges.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check whether the firewall’s management interface responds from the internet: that’s where the advisory’s two most severe flaws are.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;A flaw like this depends on how each remote access point is set up: which firewall, which SonicOS version, and whether the VPN uses SAML.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;SonicWall PSIRT, “SonicWall SonicOS Affected By Multiple Vulnerabilities”, SNWLID-2026-0014, November 6, 2026. &lt;a href=&quot;https://psirt.global.sonicwall.com/vuln-detail/SNWLID-2026-0014&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://psirt.global.sonicwall.com/vuln-detail/SNWLID-2026-0014&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CISA, Known Exploited Vulnerabilities (KEV) catalog, CVE-2026-19868, added November 6, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>APT29 targets iPhone and Android from infected websites, using flaws that are already patched</title>
    <link href="https://blog.cisent.ai/en/apt29-ataca-telefonos-con-fallas-ya-parchadas/"/>
    <id>tag:cisent.ai,2026-11-06:/en/apt29-ataca-telefonos-con-fallas-ya-parchadas/</id>
    <published>2026-11-06T12:00:00+00:00</published>
    <updated>2026-11-06T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Sekoia attributes to APT29 a campaign that attacks iPhone and Android users who simply visit an infected website. Every flaw it uses is already patched: an up-to-date phone is safe.</summary>
    <content type="html">&lt;p&gt;On Wednesday, November 4, 2026, Sekoia’s threat research team published a report on a campaign against iPhone and Android users that infects phones when they visit certain websites, with no further action on their part. It attributes the campaign, with moderate to high confidence, to APT29, the group associated with Russia’s Foreign Intelligence Service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-it-reaches-phones&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How it reaches phones&lt;/h2&gt;
&lt;p&gt;Sekoia describes three routes, and all three end in the same code:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Infected legitimate websites:&lt;/strong&gt; those of a Ukrainian gas production company, a major trade association of the US chemical industry, a far-right European political party and a Bulgarian military-related site, among others.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Fake copies of well-known websites:&lt;/strong&gt; one of America250, the initiative marking 250 years of US independence, and one of the Raffles London at The OWO hotel, close to several British government buildings. They displayed the real site and injected the attack on top of it. Sekoia thinks the hotel may have been used to target diplomatic delegations.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A compromised provider:&lt;/strong&gt; between August 27 and September 3, 2026, the analytics tracker of the Ukrainian service Plerdy, which thousands of websites load, the vast majority of them Ukrainian, served the same code.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-it-does&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it does&lt;/h2&gt;
&lt;p&gt;The code first examines the phone and its browser. Depending on the version, the server delivers a different attack chain for iOS or for Android, with no user interaction.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;On iPhone it steals browser cookies or, in one of the chains, all of the device’s data.&lt;/li&gt;
&lt;li&gt;On Android it steals Chrome data and Wi-Fi passwords.&lt;/li&gt;
&lt;li&gt;It doesn’t stay installed. With the cookies, the attacker can reuse open sessions, such as webmail.&lt;/li&gt;
&lt;li&gt;One of the chains reuses DarkSword, an iOS exploit kit suspected to have been developed by a commercial surveillance vendor and uploaded to VirusTotal in September 2026.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-key-point-everything-is-patched&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The key point: everything is patched&lt;/h2&gt;
&lt;p&gt;All three chains use only known flaws that their vendors had already fixed before the campaign. None is a zero-day and, according to Sekoia, a phone running an up-to-date system is safe.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; On that same November 4, CISA added one of those flaws, Apple’s CVE-2024-27834, to its catalog of exploited vulnerabilities.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update iOS, Android and Chrome: the chains don’t work against an up-to-date system.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check which of the organization’s phones are still on old versions. The chains Sekoia found target iOS 16.7, iOS 18.5 and Chrome for Android between versions 139 and 145.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Review the third-party scripts your own website loads: the Plerdy case shows that an outside script can serve an attack to your visitors.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;In this campaign, every infected phone connects to the attacker’s delivery and exfiltration servers, which have no business on an organization’s network.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Sekoia, “Poisoned Source: APT29 watering holes target mobile devices with n-day exploits”, by Amaury Garçon, Félix Aime and Robin Kwiatkowski, November 4, 2026. &lt;a href=&quot;https://www.sekoia.com/blog/poisoned-source-apt29-watering-holes-targets-mobile-devices-with-n-day-exploits&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.sekoia.com/blog/poisoned-source-apt29-watering-holes-targets-mobile-devices-with-n-day-exploits&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CISA, Known Exploited Vulnerabilities (KEV) catalog, CVE-2024-27834, added November 4, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Five US agencies urge hardening cooling systems after the attacks in Europe</title>
    <link href="https://blog.cisent.ai/en/eeuu-pide-blindar-los-sistemas-de-frio/"/>
    <id>tag:cisent.ai,2026-11-05:/en/eeuu-pide-blindar-los-sistemas-de-frio/</id>
    <published>2026-11-05T12:00:00+00:00</published>
    <updated>2026-11-05T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA, the FBI and three other agencies warn that internet-exposed cooling controllers can be forced to heat up, and urge taking them off the public internet.</summary>
    <content type="html">&lt;p&gt;On Thursday, November 5, 2026, CISA, the FBI and the US departments of Health (HHS), Agriculture (USDA) and Transportation (DOT) published a joint fact sheet: they say they are aware of malicious activity targeting industrial refrigeration and cold chain monitoring systems around the world.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It comes after the attacks that, in recent weeks, made the refrigerators of hundreds of supermarkets across Europe heat up.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-they-warn-about&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What they warn about&lt;/h2&gt;
&lt;p&gt;According to the fact sheet, the attackers target the electronic temperature controllers and commercial refrigeration controllers of grocery stores and cold storage facilities. They may exploit flaws or weaknesses in unpatched or internet-accessible controllers to change their settings remotely and force the units into continuous heating and defrost cycles.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;The routes:&lt;/strong&gt; bypassing authentication, injecting code or exploiting weak configurations.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The effects:&lt;/strong&gt; a temperature change can shorten the equipment’s service life, create a fire hazard, cause inventory loss or harm whoever eats products stored unsafely. Operators can also lose sight of the actual temperatures.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Who uses these systems:&lt;/strong&gt; above all the healthcare, chemical, food and agriculture, and transportation sectors, refrigerated transport included.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-they-ask-for-now&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What they ask for now&lt;/h2&gt;
&lt;p&gt;The agencies urge owners and operators to take these measures immediately:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;Take control panels off the internet. If remote access is needed, protect it behind a firewall or an up-to-date edge device.&lt;/li&gt;
&lt;li&gt;Separate operational technology (OT) devices from the business network, on their own virtual network.&lt;/li&gt;
&lt;li&gt;Change default passwords and use phishing-resistant multifactor authentication wherever feasible.&lt;/li&gt;
&lt;li&gt;Keep firmware up to date and retire devices the manufacturer no longer supports.&lt;/li&gt;
&lt;li&gt;Limit who can make changes, with allowlists of authorized devices and IP addresses.&lt;/li&gt;
&lt;li&gt;Keep current backups, offsite and offline, and test them.&lt;/li&gt;
&lt;li&gt;Log activity and review it. Where devices keep no logs, capture the network traffic between them.&lt;/li&gt;
&lt;li&gt;Set reasonable temperature ranges on the devices and alert when they drift outside them.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;from-the-supermarket-to-the-blood-bank&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;From the supermarket to the blood bank&lt;/h2&gt;
&lt;p&gt;On November 4 &lt;a href=&quot;https://blog.cisent.ai/en/efs-camaras-de-frio-a-110-grados/&quot;&gt;we covered&lt;/a&gt; what happened at France’s blood service:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;On the night of Sunday, November 1 to Monday, November 2, cold rooms kept at around −30 °C were reprogrammed to reach 110 °C.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;According to The Record, it is unclear whether that case is linked to the supermarkets. In August, refrigerators and freezers at several US military commissaries unexpectedly switched into defrost or heating modes. The Defense Department’s inspector general, among others, is investigating, and it is also unclear whether that is connected to this wave.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-is-still-unknown&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What is still unknown&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Who is behind it: no one has claimed the attacks in Europe.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Whether the supermarkets, France’s blood service and the US military commissaries are part of the same thing.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Which brands and models of controllers are affected: the fact sheet does not name them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The fact sheet asks operators to log and review activity, to capture network traffic where devices keep no logs, and to alert when a temperature drifts out of range.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, FBI, HHS, USDA and DOT, “US Government Partners Urge Hardening Industrial Refrigeration and Cold Chain Monitoring Systems”, TLP:CLEAR fact sheet, November 5, 2026. &lt;a href=&quot;https://www.ic3.gov/CSA/2026/261105.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.ic3.gov/CSA/2026/261105.pdf&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;The Record, “Harden cooling systems now, feds warn critical infrastructure operators amid attacks”, by Martin Matishak, November 5, 2026. &lt;a href=&quot;https://therecord.media/cyberattacks-refrigeration-cisa-fbi&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/cyberattacks-refrigeration-cisa-fbi&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Emergency fix in OpenSSH: on Linux with musl and no PAM, the password could be bypassed</title>
    <link href="https://blog.cisent.ai/en/openssh-arreglo-de-emergencia-musl/"/>
    <id>tag:cisent.ai,2026-11-05:/en/openssh-arreglo-de-emergencia-musl/</id>
    <published>2026-11-05T12:00:00+00:00</published>
    <updated>2026-11-05T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>OpenSSH 10.6p2 fixes a bypass of sshd&#x27;s password authentication that affects Linux systems with the musl library, such as Alpine, and PAM disabled.</summary>
    <content type="html">&lt;p&gt;OpenSSH 10.6p2 came out on Thursday, November 5, 2026. According to its developers, it carries a critical security fix for Linux systems that use the musl library and have PAM disabled.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-it-fixes&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it fixes&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;sshd, the SSH server, allowed password authentication to be bypassed on some Linux systems built and run without PAM.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;OpenSSH considers the root cause to be outside sshd, but it is exploitable through sshd and can be avoided there.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to The Hacker News, on affected Alpine servers a password longer than 256 bytes could unlock accounts set up for key-only login, without knowing the credential.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It was reported by Özgün Kültekin of Wakefield.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-affected&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is affected&lt;/h2&gt;
&lt;p&gt;musl is the C library of Alpine Linux, a widely used base for container images. The advisory only covers systems with musl and PAM disabled, so distributions with glibc, such as Debian or Ubuntu, fall outside that condition.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; And since the flaw lies in password authentication, an sshd that only accepts keys doesn’t offer it.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Update to OpenSSH 10.6p2, or to each distribution’s package carrying the fix.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Check which servers and containers run sshd on musl.&lt;/li&gt;
&lt;li&gt;Make SSH key-only wherever possible.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-there-will-be-more-fixes-like-this&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why there will be more fixes like this&lt;/h2&gt;
&lt;p&gt;In the same notes, OpenSSH says it receives many security reports, a good share of them found with AI models or with their help. It has seen cases where a flaw found by an AI was later discovered independently by another researcher, and it concludes that attackers who don’t report bugs can probably find them too. So, for now, it will ship releases more often instead of batching fixes for the next planned release.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;A flaw like this depends on where each SSH server runs: on which system, with which library, and whether it accepts passwords.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;OpenSSH, release notes for version 10.6p2, November 5, 2026. &lt;a href=&quot;https://www.openssh.com/releasenotes.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.openssh.com/releasenotes.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;The Hacker News, on OpenSSH’s emergency fix for Alpine Linux, November 5, 2026. &lt;a href=&quot;https://thehackernews.com/2026/11/openssh-issues-emergency-fix-for-flaw.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://thehackernews.com/2026/11/openssh-issues-emergency-fix-for-flaw.html&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Attackers reset the cold rooms of France&#x27;s blood service to 110 °C</title>
    <link href="https://blog.cisent.ai/en/efs-camaras-de-frio-a-110-grados/"/>
    <id>tag:cisent.ai,2026-11-04:/en/efs-camaras-de-frio-a-110-grados/</id>
    <published>2026-11-04T12:00:00+00:00</published>
    <updated>2026-11-04T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The cold rooms a supplier manages for France&#x27;s blood service were ordered to heat up, and the EFS says no blood products were lost.</summary>
    <content type="html">&lt;p&gt;On Wednesday, November 4, 2026, the Établissement français du sang (EFS), France’s public blood service, confirmed that a cyberattack targeted its cooling systems on Monday, November 2. Le Canard enchaîné had revealed it on Tuesday, November 3.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to the EFS, the temperature and defrost thresholds of cold rooms at three of its sites were raised to very high temperatures, and that equipment is managed by a supplier. Le Canard enchaîné names it as Coolworld, which said a limited number of its units were affected by a cybersecurity incident targeting refrigeration management software supplied to it by a third party.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;On the night of Sunday, November 1 to Monday, November 2, cold rooms kept at around −30 °C were reprogrammed to reach 110 °C. They held donor samples and plasma for transfusions as well.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The plasma was also used to make medicines, and an on-duty technician reportedly caught the problem in time for a quick response.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-was-saved&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was saved&lt;/h2&gt;
&lt;p&gt;The EFS says no health impact was identified and there is no loss of blood products to report. According to the office of the French health minister, the stocks were saved thanks to the facilities’ monitoring and alarm systems and the measures taken right away: remote access to the affected equipment was cut off, monitoring was stepped up and the products were moved to other facilities.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;like-the-supermarkets&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Like the supermarkets&lt;/h2&gt;
&lt;p&gt;It resembles what &lt;a href=&quot;https://blog.cisent.ai/en/supermercados-refrigeradores-europa/&quot;&gt;we covered on October 12&lt;/a&gt;, when the cooling equipment of hundreds of French supermarkets was ordered to heat up. That day we wrote:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;ANSSI identified the weak point in software from the Danish company Danfoss, which stores use to manage their refrigeration equipment remotely.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;Nothing yet proves that Coolworld’s equipment belongs to the models affected then, or that the attackers are the same. About the supermarkets, ANSSI director Vincent Strubel told the Senate on October 21 that the attacks exploited basic flaws, such as equipment exposed to the internet and weak passwords, and he also pointed to the health sector’s exposure.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;In Chile, when asked whether what happened to the supermarkets could happen there, a former ANCI director &lt;a href=&quot;https://blog.cisent.ai/en/chile-ataque-como-francia-totalmente-posible/&quot;&gt;answered “Totally.”&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;p&gt;According to the health minister’s office, the technical investigation continues in order to establish:&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;How the attackers operated, and under what circumstances.&lt;/li&gt;
&lt;li&gt;How many pieces of equipment in total may be affected.&lt;/li&gt;
&lt;li&gt;When, and under what conditions, the facilities will be back in service.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Nor is it known who is behind it, or whether it is the same wave as in October.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Coolworld, the incident targeted refrigeration management software supplied to it by a third party,&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; and the measures taken included cutting off remote access to the equipment.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; At the supermarkets, according to ANSSI’s director, the attackers exploited equipment exposed to the internet and weak passwords.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Le Monde, “Des réfrigérateurs de l’Etablissement français du sang à leur tour victimes d’une cyberattaque”, November 4, 2026, with the statements of the EFS and Coolworld. &lt;a href=&quot;https://www.lemonde.fr/pixels/article/2026/11/04/des-frigos-de-l-etablissement-francais-du-sang-a-leur-tour-victimes-d-une-cyberattaque_6806228_4408996.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.lemonde.fr/pixels/article/2026/11/04/des-frigos-de-l-etablissement-francais-du-sang-a-leur-tour-victimes-d-une-cyberattaque_6806228_4408996.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;IT-Connect, “Cyberattaque contre un prestataire de l’EFS : des congélateurs de plasma réglés à 110 °C”, by Florian Burnel, November 4, 2026. &lt;a href=&quot;https://www.it-connect.fr/cyberattaque-efs-coolworld-congelateurs-plasma/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.it-connect.fr/cyberattaque-efs-coolworld-congelateurs-plasma/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Sciences et Avenir with AFP, “Cyberattaque contre l’Établissement français du sang : des chambres froides piratées”, November 4, 2026, with the EFS statement and comments from the health minister’s office. &lt;a href=&quot;https://www.sciencesetavenir.fr/high-tech/reseaux-et-telecoms/cyberattaque-contre-l-etablissement-francais-du-sang-des-chambres-froides-piratees-les-stocks-preserves_195932&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.sciencesetavenir.fr/high-tech/reseaux-et-telecoms/cyberattaque-contre-l-etablissement-francais-du-sang-des-chambres-froides-piratees-les-stocks-preserves_195932&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A WordPress core flaw that lets attackers run code without credentials is being exploited</title>
    <link href="https://blog.cisent.ai/en/wordpress-falla-del-nucleo-explotada/"/>
    <id>tag:cisent.ai,2026-11-04:/en/wordpress-falla-del-nucleo-explotada/</id>
    <published>2026-11-04T12:00:00+00:00</published>
    <updated>2026-11-04T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CISA added a WordPress core flaw, already fixed, to its catalog of exploited vulnerabilities, and both WordPress and CISA ask for updates as soon as possible.</summary>
    <content type="html">&lt;p&gt;On Tuesday, November 3, 2026, CISA, the US cybersecurity agency, added a WordPress core flaw, CVE-2026-119977, to its catalog of exploited vulnerabilities. It gave federal agencies until November 6 to fix it, and does not yet know whether any ransomware group uses it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;the-flaw&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The flaw&lt;/h2&gt;
&lt;p&gt;It is a race condition in full core updates. While WordPress updates, it uses predictable, web-accessible temporary paths, and an attacker without credentials can use that window to run code on the server. It is a hard attack to pull off, but it is happening.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;WordPress 7.1.4, released on October 28, fixed it, and Umut Bayram, a Picus Security researcher, found it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Wordfence said it blocked 13,095 attacks against this flaw in 24 hours.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;p&gt;The WordPress security team asks for sites to be updated immediately, and CISA asks the same of all defenders, including those in the private sector.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Who is exploiting the flaw, and against which sites.&lt;/li&gt;
&lt;li&gt;How many sites remain unpatched.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The flaw opens during the site’s update, and a version released a week ago fixes it: the first step is knowing which WordPress sites the organization has published, and on which version.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, Known Exploited Vulnerabilities Catalog, version 2026.11.03: CVE-2026-119977, added on November 3, 2026, due November 6. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;BleepingComputer, “CISA flags WordPress RCE flaw as actively exploited in attacks”, by Sergiu Gatlan, November 4, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/cisa-flags-wordpress-rce-flaw-as-actively-exploited-in-attacks/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/cisa-flags-wordpress-rce-flaw-as-actively-exploited-in-attacks/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>The Binance theft began with a fake job interview of one of its engineers</title>
    <link href="https://blog.cisent.ai/en/binance-entrevista-de-trabajo-falsa/"/>
    <id>tag:cisent.ai,2026-11-03:/en/binance-entrevista-de-trabajo-falsa/</id>
    <published>2026-11-03T12:00:00+00:00</published>
    <updated>2026-11-03T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Binance explained how they got in: a fake job interview installed malware on an engineer&#x27;s computer that its protection did not detect, and from there they reached the wallets.</summary>
    <content type="html">&lt;p&gt;On Monday, November 3, 2026, Binance published its report on the theft of more than $6 billion, and explained how the attacker got in.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It is the theft &lt;a href=&quot;https://blog.cisent.ai/en/fbi-atribuye-a-corea-del-norte-el-robo-a-binance/&quot;&gt;the FBI attributed on October 31&lt;/a&gt; to TraderTraitor, a North Korean group. That day, under “What isn’t known yet”, we wrote:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;What the other points of entry are, and how the laptop was compromised.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;Both now have an answer.&lt;/p&gt;
&lt;section id=&quot;the-interview&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The interview&lt;/h2&gt;
&lt;p&gt;In September, someone posing as a project director at a made-up company contacted a Binance software engineer on a legitimate job-hunting platform. There was a video call with a supposed HR representative, and then a supposed CEO told him, over a chat app, to download a coding assignment from a code repository.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;That code installed custom malware on the engineer’s computer: it replaced Homebrew’s &lt;code&gt;pyenv&lt;/code&gt; executable, on macOS, with a malicious one that ran commands from an attacker’s server. According to Binance, his computer’s protection tools did not detect it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;from-one-computer-to-the-wallets&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;From one computer to the wallets&lt;/h2&gt;
&lt;p&gt;By October 20, from the engineer’s computer and with an internal AI tool, the attacker compromised two more employees who worked on the wallet infrastructure, and obtained credentials and a token to deploy containers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;On October 29, the attacker used those credentials to deploy a malicious script, reached the service that signs the wallets’ transactions and moved the funds across several blockchains. It also abused Ethereum’s EIP-7702 mechanism to keep stealing after withdrawals were suspended.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to Binance, about $2.85 billion had been frozen or intercepted as of November 3, and it had replenished 98% of the affected assets as of October 31.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The tactic resembles the North Korean campaigns known as “Contagious Interview” and “Operation Dream Job”, although Binance did not link it to either.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Which job platform, chat app and repository the attacker used.&lt;/li&gt;
&lt;li&gt;Which internal AI tool helped it reach the other two employees, and how.&lt;/li&gt;
&lt;li&gt;What the independent firms reviewing the report will conclude.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Binance, the malware was not detected by the computer’s protection tools. What came next shows up on the network: a computer taking orders from an outside server, and stolen credentials deploying a script on the service that signs the wallets.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “Binance says fake job interview led to $6.3 billion crypto heist”, by Lawrence Abrams, November 3, 2026, on Binance’s report. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/binance-says-fake-job-interview-led-to-63-billion-crypto-heist/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/binance-says-fake-job-interview-led-to-63-billion-crypto-heist/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>WorkOS says an attacker got into an internal app with ARTEX, the AI framework Vapor Panda used</title>
    <link href="https://blog.cisent.ai/en/workos-atacado-con-artex/"/>
    <id>tag:cisent.ai,2026-11-03:/en/workos-atacado-con-artex/</id>
    <published>2026-11-03T12:00:00+00:00</published>
    <updated>2026-11-03T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The company says the attacker used ARTEX, the AI agent framework behind the attacks on foreign ministries and on DIVD, and reached contacts and conversations with customers.</summary>
    <content type="html">&lt;p&gt;WorkOS, a developer platform, said that on October 30, 2026 it detected unauthorized access to an internal application used by its sales team.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Its founder, Michael Grinich, said the attacker used ARTEX, an AI offensive security framework that automated reconnaissance and exploitation attempts. In the attacker’s requests they saw the user agents “artex-enrich/1.0” and “norma/0.4”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It is the framework that, according to CrowdStrike, &lt;a href=&quot;https://blog.cisent.ai/en/vapor-panda-agentes-de-ia-cancillerias/&quot;&gt;Vapor Panda used to get into the foreign ministries of two Latin American countries&lt;/a&gt;. That day we wrote:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;It considers it highly likely that other China-linked groups will adopt frameworks like ARTEX over the next twelve months.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;It is not known whether the WorkOS attacker has anything to do with Vapor Panda. ARTEX was published this year as open source by Chinese developers, went private after attacks on South Korean banks, and copies still circulate.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-it-reached&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it reached&lt;/h2&gt;
&lt;p&gt;According to WorkOS, the attacker reached the contact data in its CRM (names, emails, job titles and companies), its customers’ account histories, and summaries and excerpts of sales calls, support conversations and Slack channels with customers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The company says its product data and production services were not touched. It cut external access to the application, reviewed the permissions of its internal tools and hired an independent forensics firm.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It will notify affected customers directly, and asks them to watch for emails or calls that use that information to deceive them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;How the attacker got into the application.&lt;/li&gt;
&lt;li&gt;Whether it is linked to Vapor Panda, and which version or copy of ARTEX it used.&lt;/li&gt;
&lt;li&gt;How many customers are affected, and whether the data left the company.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The attacker used AI to automate reconnaissance and exploitation attempts, and reached an internal application that could be accessed from outside: WorkOS cut that access afterwards.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “WorkOS says attacker used ARTEX AI to breach internal sales app”, by Bill Toulas, November 3, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/workos-says-attacker-used-artex-ai-to-breach-internal-sales-app/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/workos-says-attacker-used-artex-ai-to-breach-internal-sales-app/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Mercado Bitcoin, Brazil&#x27;s largest crypto exchange, suspended deposits and withdrawals after an attack</title>
    <link href="https://blog.cisent.ai/en/mercado-bitcoin-suspende-depositos-y-retiros/"/>
    <id>tag:cisent.ai,2026-11-02:/en/mercado-bitcoin-suspende-depositos-y-retiros/</id>
    <published>2026-11-02T12:00:00+00:00</published>
    <updated>2026-11-02T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Days after the Binance theft, the Brazilian exchange detected an incident in one of its systems; it says no customer will lose their assets, and some experts dispute that.</summary>
    <content type="html">&lt;p&gt;In the early hours of Monday, November 2, 2026, Mercado Bitcoin detected a security incident in one of its systems and suspended deposits and withdrawals as a precaution.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It is Brazil’s largest crypto exchange,&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; and the attack came days after the Binance theft, which &lt;a href=&quot;https://blog.cisent.ai/en/binance-billeteras-vaciadas-en-ethereum/&quot;&gt;we covered on October 29&lt;/a&gt;.&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The company says it identified and contained the incident overnight, that it also suspended access to its app, and that it published the addresses of the wallets involved so they can be traced.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Hypernative, a blockchain security company, estimated that at least $40 million left the platform, about 217.7 million reais.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Other companies in the field spoke of more than $46 million, which left on Sunday night.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Mercado Bitcoin did not confirm either figure.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-was-blocked&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was blocked&lt;/h2&gt;
&lt;p&gt;Tether’s CEO, Paolo Ardoino, said his company froze more than 4.5 million in USDT.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;That night, Mercado Bitcoin said it had blocked $11.2 million, that it has identified other amounts that can be blocked, and that the assets still involved add up to about $38 million. It is working on recovery with Tether, Circle, BRL1, Paxos and Chainalysis, and it notified the authorities.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-the-company-says-and-what-is-disputed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the company says, and what is disputed&lt;/h2&gt;
&lt;p&gt;Mercado Bitcoin says there was no data leak&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; and that no customer will lose their assets.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Several blockchain security experts questioned that: according to The Record, the blockchain shows that funds were taken from some customer wallets.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;How the attacker moved the funds, and who it was.&lt;/li&gt;
&lt;li&gt;The final loss, once whatever can be blocked is taken out.&lt;/li&gt;
&lt;li&gt;When deposits and withdrawals will resume: the company says it will announce it.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Mercado Bitcoin, the incident happened in one of its systems.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Manuela Miniguini, CNN Brasil, “Mercado Bitcoin identifica ‘incidente de segurança’ nesta segunda-feira (2)”, November 2, 2026, updated the same day. &lt;a href=&quot;https://www.cnnbrasil.com.br/economia/money/mercado/mercado-bitcoin-identifica-incidente-em-seu-sistema-de-cripto/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cnnbrasil.com.br/economia/money/mercado/mercado-bitcoin-identifica-incidente-em-seu-sistema-de-cripto/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;The Record, “Largest crypto exchange in Brazil confirms cyberattack as millions flow out of platform”, November 2, 2026. &lt;a href=&quot;https://therecord.media/mercado-bitcoin-brazil-cryptocurrency-heist&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/mercado-bitcoin-brazil-cryptocurrency-heist&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Exame, “Mercado Bitcoin sofre ataque hacker estimado em R$ 217,7 milhões”, November 2, 2026. &lt;a href=&quot;https://exame.com/future-of-money/mercado-bitcoin-sofre-ataque-hacker-estimado-em-r-2177-milhoes/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://exame.com/future-of-money/mercado-bitcoin-sofre-ataque-hacker-estimado-em-r-2177-milhoes/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;CNN Brasil, “Mercado Bitcoin bloqueia US$ 11,2 milhões após incidente de segurança”, November 2, 2026, updated the same day. &lt;a href=&quot;https://www.cnnbrasil.com.br/economia/money/negocios/mercado-bitcoin-bloqueia-us-112-milhoes-apos-incidente-de-seguranca/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cnnbrasil.com.br/economia/money/negocios/mercado-bitcoin-bloqueia-us-112-milhoes-apos-incidente-de-seguranca/&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Two researchers find factory malware in street-market TV boxes; Google takes down a network of 1.3 million</title>
    <link href="https://blog.cisent.ai/en/tv-box-genericos-con-malware-de-fabrica/"/>
    <id>tag:cisent.ai,2026-11-02:/en/tv-box-genericos-con-malware-de-fabrica/</id>
    <published>2026-11-02T12:00:00+00:00</published>
    <updated>2026-11-02T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Researchers found factory-installed malicious software in generic TV boxes sold at Chilean street markets, the same pattern for which Google has just banned a manufacturer.</summary>
    <content type="html">&lt;p&gt;On Sunday, November 1, 2026, the Chilean newspaper Las Últimas Noticias published what Anthony de la Paz, a senior consultant at Inside Security, and Patricio Moraga, a senior cybersecurity consultant at Cybertrust Latam, found in generic TV boxes: the boxes that plug into a TV to watch channels and apps, sold at street markets for about 10,000 Chilean pesos. They presented the research at Ekoparty, the cybersecurity conference in Buenos Aires.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The next day, Google said it had taken down a network of more than 1.3 million Android TV boxes, and that their manufacturer had installed the backdoor on purpose.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-they-found-in-chile&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What they found in Chile&lt;/h2&gt;
&lt;p&gt;The researchers opened the devices, connected a few wires inside and confirmed they had been tampered with at the factory: the malicious software is stored in the device’s memory and stays there even after a reset to factory settings. They found exactly the same board in at least three products sold in different packaging; only the plastic case changed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;About every seven minutes, the device sends a report to the attacker. According to the researchers, that lets the attacker see what is done on the TV box, such as card details entered in its apps, and get the password of the email account the device is linked to, along with what arrives in that inbox, including the messages sent to confirm the user’s identity. “By using it, you let the attacker into your home,” de la Paz summed up.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to the researchers, the problem is with generic devices: boxes from brands such as Google or Apple don’t have it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-google-took-down&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Google took down&lt;/h2&gt;
&lt;p&gt;According to Google, the network is called XCPE, and the manufacturer, Videostrong Technology, put the backdoor into the firmware during manufacturing to make money by selling access to criminals. The backdoor runs with system privileges and, if removed, reinstalls itself from a system partition.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It lets operators install apps silently, turn off Play Protect, capture the screen and keystrokes, and collect account and device data. Google took down the control servers together with Japan’s National Police Agency and several companies, and banned Videostrong from making certified Android devices.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Cleaning an infected device requires installing clean firmware; if that isn’t possible, the remaining option is to disconnect it from the internet.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;We don’t know whether the devices opened in Chile are made by Videostrong or are part of XCPE: the Las Últimas Noticias article names no brands or models.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Which brands and models the researchers opened, and which servers they report to.&lt;/li&gt;
&lt;li&gt;Whether the devices sold in Chile belong to the same family as XCPE.&lt;/li&gt;
&lt;li&gt;How many are connected in Chile.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;If a TV box like this is connected to an office network, it reports to the attacker every few minutes from the inside.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Las Últimas Noticias, “Ethical hackers descubrieron que los TV-Box genéricos traen un virus para robar datos personales” (Ethical hackers found that generic TV boxes come with a virus to steal personal data), by Camila Figueroa, November 1, 2026, print edition, page 4.&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;BleepingComputer, “Google disrupts million-device XCPE Android malware botnet, bans TV box maker”, November 2, 2026, on Google’s report. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/google-disrupts-million-device-xcpe-android-malware-botnet-bans-tv-box-maker/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/google-disrupts-million-device-xcpe-android-malware-botnet-bans-tv-box-maker/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>The FBI attributes the Binance theft to North Korea and asks for 628 addresses to be blocked</title>
    <link href="https://blog.cisent.ai/en/fbi-atribuye-a-corea-del-norte-el-robo-a-binance/"/>
    <id>tag:cisent.ai,2026-10-31:/en/fbi-atribuye-a-corea-del-norte-el-robo-a-binance/</id>
    <published>2026-10-31T12:00:00+00:00</published>
    <updated>2026-10-31T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>According to the FBI, TraderTraitor, a North Korean group, is behind the Binance theft, and exchanges should block the addresses being used to launder the funds.</summary>
    <content type="html">&lt;p&gt;On Saturday, October 31, 2026, the FBI said North Korea is responsible for the theft of more than $6 billion in virtual assets from Binance, and attributed it to TraderTraitor, a group also tracked as Lazarus, Jade Sleet or APT38.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It is the theft &lt;a href=&quot;https://blog.cisent.ai/en/binance-billeteras-vaciadas-en-ethereum/&quot;&gt;we covered on October 29&lt;/a&gt;. That day, under “What isn’t known yet”, we wrote:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;Who did it, and how the laptop was compromised.&lt;/p&gt;
&lt;/blockquote&gt;
&lt;p&gt;The first half now has an answer.&lt;/p&gt;
&lt;section id=&quot;what-the-fbi-asked-for&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the FBI asked for&lt;/h2&gt;
&lt;p&gt;Together with the Secret Service and IRS Criminal Investigation, the FBI asked exchanges, blockchain operators and other virtual asset service providers to block transactions tied to the addresses TraderTraitor is using to launder the funds. It published 628 of them: 608 on Ethereum, 19 on Bitcoin and one on Solana.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to the FBI, the group is converting most of the stolen assets into ether, to launder them later through decentralized finance platforms and instant swap services.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; TRM Labs and Chainalysis had already linked the theft to TraderTraitor.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The Secret Service said that within minutes of the attack it worked with industry and the Justice Department to identify and freeze a significant portion of the stolen funds.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-binance-said&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Binance said&lt;/h2&gt;
&lt;p&gt;Binance reopened deposits and withdrawals on the morning of Sunday, November 1.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Its CEO, Richard Teng, said the attacker used a self-destruct mechanism that wiped its tools and traces.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The company first blamed an engineer’s compromised laptop, and now says the investigation found other points of entry.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It also says it has already covered about 95% of the affected amount and will not pass the losses on to its users.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;What the other points of entry are, and how the laptop was compromised.&lt;/li&gt;
&lt;li&gt;How much was really lost, once the frozen funds are taken out.&lt;/li&gt;
&lt;li&gt;How much of the stolen money will be laundered before it is blocked.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Binance, the attacker did not get in through a single place: besides an engineer’s laptop, the investigation found other points of entry.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “FBI links $6.3 billion Binance crypto heist to North Korean hackers”, November 1, 2026, on the FBI alert. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/fbi-links-63-billion-binance-crypto-heist-to-north-korean-hackers/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/fbi-links-63-billion-binance-crypto-heist-to-north-korean-hackers/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;The Record, “FBI blames N. Korea’s TraderTraitor for Binance theft, urges crypto industry to block transactions”, November 2, 2026. &lt;a href=&quot;https://therecord.media/fbi-blames-tradertraitor-for-binance-hack&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/fbi-blames-tradertraitor-for-binance-hack&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Attackers got into Ivanti VPNs even with every patch applied, and wiped their tracks; the chain is now fixed</title>
    <link href="https://blog.cisent.ai/en/ivanti-connect-secure-cadena-sin-rastro/"/>
    <id>tag:cisent.ai,2026-10-30:/en/ivanti-connect-secure-cadena-sin-rastro/</id>
    <published>2026-10-30T12:00:00+00:00</published>
    <updated>2026-10-30T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A group chained Ivanti Connect Secure flaws to take over remote access appliances without credentials, and CISA ordered the last ones fixed within days.</summary>
    <content type="html">&lt;p&gt;On October 30, 2026, CISA, the US cybersecurity agency, added two more Ivanti Connect Secure flaws to its catalog of exploited vulnerabilities. Connect Secure is the appliance many organizations use to provide remote VPN access. CISA gave federal agencies until November 2 to fix one of them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Volexity discovered it in the first week of October. Since the 3rd, a group it tracks as UTA0576 had been running commands as root on a Connect Secure appliance without credentials, even though the appliance had every patch applied.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;CISA describes three flaws that chain together:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;CVE-2026-104935 lets an attacker bypass authentication and run commands remotely, and has been in the catalog since October 13;&lt;/li&gt;
&lt;li&gt;CVE-2026-118527 lets an attacker upload files with attacker-controlled content to a known location on disk;&lt;/li&gt;
&lt;li&gt;CVE-2026-111357 lets an attacker escalate privileges.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Ivanti fixed the first one on October 13, and the privilege escalation on the 29th.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-they-hid&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they hid&lt;/h2&gt;
&lt;p&gt;According to Volexity, the group left no malware or web shells: it ran the full chain again for every command. It stole stored credentials, searched Active Directory for LAPS passwords and BitLocker keys, and sprayed passwords. Then it deleted its entries from the logs, removed temporary files and restored the memory and configuration it had changed. Ivanti’s integrity checking tool detected nothing.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Shadowserver tracks more than 5,000 Connect Secure appliances exposed to the internet, and it is unknown how many are already protected.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; It is not the first time: these appliances had zero-days exploited in 2024 and in 2025.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;p&gt;Volexity recommends:&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;applying Ivanti’s fixes;&lt;/li&gt;
&lt;li&gt;watching for a high volume of requests to &lt;code&gt;welcome.cgi&lt;/code&gt;, especially when the path and the &lt;code&gt;X-DSCGI-PATH&lt;/code&gt; header don’t match;&lt;/li&gt;
&lt;li&gt;checking &lt;code&gt;/tmp/&lt;/code&gt; for traces of the attack;&lt;/li&gt;
&lt;li&gt;running network and memory forensics on suspicious appliances, because the integrity checking tool may not see the compromise.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;How many organizations were hit.&lt;/li&gt;
&lt;li&gt;Who UTA0576 is. Volexity believes, with low confidence, that it is after money rather than espionage.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;To fall to this chain, the appliance has to be published on the internet: which edge devices you have exposed, and on which version, can be checked before anyone tries it.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, Known Exploited Vulnerabilities Catalog, version 2026.10.30: CVE-2026-111357 and CVE-2026-118527, added on October 30, 2026, and CVE-2026-104935, on October 13. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Volexity, “A Series of Unauthenticated Events: Chained Exploitation of Ivanti Connect Secure Zero-Days”, October 13, 2026, with its October 29 update. &lt;a href=&quot;https://www.volexity.com/blog/2026/10/13/a-series-of-unauthenticated-events-chained-exploitation-of-ivanti-connect-secure-zero-days/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.volexity.com/blog/2026/10/13/a-series-of-unauthenticated-events-chained-exploitation-of-ivanti-connect-secure-zero-days/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Sergiu Gatlan, BleepingComputer, “Ivanti patches two more Connect Secure zero-days chained in attacks”, November 2, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/ivanti-patches-two-more-connect-secure-zero-days-chained-in-attacks/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/ivanti-patches-two-more-connect-secure-zero-days-chained-in-attacks/&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Binance thinks it all began on an engineer&#x27;s laptop, and puts the impact at $6.3 billion</title>
    <link href="https://blog.cisent.ai/en/binance-billeteras-vaciadas-en-ethereum/"/>
    <id>tag:cisent.ai,2026-10-29:/en/binance-billeteras-vaciadas-en-ethereum/</id>
    <published>2026-10-29T12:00:00+00:00</published>
    <updated>2026-10-29T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>According to Binance&#x27;s initial investigation, the attacker got in through an engineer&#x27;s laptop, reached an internal service and submitted unauthorized withdrawals.</summary>
    <content type="html">&lt;p&gt;On October 29, 2026, at 20:24 UTC, a single transaction took 167,464 ether and 1.615 billion USDT out of the wallet the Etherscan explorer labels “Binance 14”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; That night, Binance put the impact at about $6.3 billion, of which close to $3 billion had been frozen.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Four minutes earlier, someone set an EIP-7702 delegation on that wallet, an Ethereum mechanism that lets an ordinary wallet run a contract’s code. The transaction carried 50 authorizations for consecutive sequence numbers of the wallet, and one of them worked.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Over the next 40 minutes, about $3.52 billion reached the receiving address, counting only what moved on Ethereum.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;At 02:36 UTC on October 30, Binance gave its first explanation: a compromised DevOps engineer’s laptop may have let the attacker reach a service container and submit unauthorized withdrawals.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; The company says it has found no evidence that any of its wallets’ private keys leaked.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-was-frozen-and-what-wasnt&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What was frozen, and what wasn’t&lt;/h2&gt;
&lt;p&gt;At 21:01 UTC, Tether blocked the address receiving the funds, and at 21:04 UTC Circle froze the 823 million USDC there.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Tether later said it had frozen more than 1.567 billion USDT.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Not everything can be frozen. Before the blocks, 700 million USDT went to two wallets that swapped it for ether, and other tokens that neither Tether nor Circle controls went to another address.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; At 21:44 UTC, the same address swept “Binance 14” again and moved another 200 million USDT, this time into the address that was already frozen.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-binance-did&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Binance did&lt;/h2&gt;
&lt;p&gt;At 22:55 UTC it froze withdrawals, and at 00:36 UTC its co-CEO, Yi He, said deposits and withdrawals were suspended while the company works with law enforcement.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; Binance revoked the affected address’s authorization and says it will cover all losses from its SAFU reserve fund and its own funds.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It has also promised a detailed technical post-mortem once the investigation is complete.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;That Thursday, Binance’s token, BNB, fell more than 8%, and bitcoin dropped about 2% to a six-week low, according to Reuters, as cited by BioBioChile.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Who did it, and how the laptop was compromised.&lt;/li&gt;
&lt;li&gt;The final figure: Binance first spoke of more than $5.7 billion, and later of about $6.3 billion.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;When withdrawals will resume: Binance says as soon as the security issue is resolved.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Whether it is the largest theft in crypto history, as was feared that night. That depends on the real loss, once the frozen funds are taken out.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;According to Binance, the attack didn’t start on the blockchain, but on the laptop of someone at the company.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;TokenPost, “Binance Investigates Security Reports as Tether and Circle Freeze Funds”, October 29, 2026. &lt;a href=&quot;https://www.tokenpost.com/news/technology/44018&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.tokenpost.com/news/technology/44018&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Bloomingbit, “Binance Estimates Security Breach Losses at $6.3 Billion, Suspects Employee Laptop Hack”, October 29, 2026. &lt;a href=&quot;https://en.bloomingbit.io/feed/news/123262&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://en.bloomingbit.io/feed/news/123262&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Unchained, “Binance Says User Funds Are ‘Completely Safe’ After $3.5 Billion Drains From Its Hot Wallets”, October 29, 2026. &lt;a href=&quot;https://unchainedcrypto.com/binance-says-user-funds-are-completely-safe-after-3-5-billion-drains-from-its-hot-wallets/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://unchainedcrypto.com/binance-says-user-funds-are-completely-safe-after-3-5-billion-drains-from-its-hot-wallets/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;CoinDesk, “Live updates: Binance confirms security breach, cryptos fall”, October 29, 2026, with the time of each update. &lt;a href=&quot;https://www.coindesk.com/business/2026/10/29/live-updates-bitcoin-clings-to-usd80-000-as-strategy-s-cushion-thins-before-earnings&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.coindesk.com/business/2026/10/29/live-updates-bitcoin-clings-to-usd80-000-as-strategy-s-cushion-thins-before-earnings&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;BioBioChile, “Binance congela miles de millones, suspende retiros y enfrenta el mayor hackeo cripto de la historia”, October 30, 2026. &lt;a href=&quot;https://www.biobiochile.cl/noticias/economia/negocios-y-empresas/2026/10/30/binance-congela-miles-de-millones-suspende-retiros-y-enfrenta-el-mayor-hackeo-cripto-de-la-historia.shtml&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.biobiochile.cl/noticias/economia/negocios-y-empresas/2026/10/30/binance-congela-miles-de-millones-suspende-retiros-y-enfrenta-el-mayor-hackeo-cripto-de-la-historia.shtml&lt;/a&gt;&lt;a href=&quot;#fnref14&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Four fraud alerts in two days: Banco Santander, FoodCorp, Kendal and a fake job offer</title>
    <link href="https://blog.cisent.ai/en/csirt-cuatro-alertas-de-fraude/"/>
    <id>tag:cisent.ai,2026-10-29:/en/csirt-cuatro-alertas-de-fraude/</id>
    <published>2026-10-29T12:00:00+00:00</published>
    <updated>2026-10-29T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>On October 28 and 29, Chile&#x27;s national CSIRT warned of fake sites that use well-known brands to steal data.</summary>
    <content type="html">&lt;p&gt;On October 28 and 29, 2026, Chile’s national CSIRT published four alerts about fraud campaigns: Kendal and a fake job offer on the 28th, and Banco Santander and FoodCorp on the 29th.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-they-work&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they work&lt;/h2&gt;
&lt;p&gt;All four use the brand’s corporate image and logos, without authorization, on a fake site built to steal the data of whoever lands there.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The domains mimic the real ones with small tricks. Kendal’s has an extra letter. Banco Santander’s spells “office” with a zero instead of the “o”, and hangs off a site that has nothing to do with the bank. The fake job offer goes through an intermediate address before landing on the fake site.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The CSIRT reminds people that these messages try to scare or rush them into acting without thinking, and asks them to report them on its site or by calling 1510.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A notice from the bank or a job offer arrives just when you’re expecting one. That’s why they work: the trick doesn’t rely on a technical flaw, but on a click made in a hurry.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CSIRT Nacional, Government of Chile, alerts ACF26-01226 to ACF26-01229, October 28 and 29, 2026. &lt;a href=&quot;https://csirt.gob.cl/alertas/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A Linux kernel flaw now has an exploit, and Go programs open the door to it by default</title>
    <link href="https://blog.cisent.ai/en/kernel-de-linux-exploit-por-mptcp/"/>
    <id>tag:cisent.ai,2026-10-29:/en/kernel-de-linux-exploit-por-mptcp/</id>
    <published>2026-10-29T12:00:00+00:00</published>
    <updated>2026-10-29T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The attack described takes over a server listening with MPTCP as root, and recent Go programs listen that way without anyone asking.</summary>
    <content type="html">&lt;p&gt;On October 29, 2026, the Go security team warned people running Go programs on Linux about a kernel flaw, CVE-2026-98432.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to the notice, an AI lab described an exploit that runs code remotely on a server listening with Multipath TCP (MPTCP).&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; And since Go 1.24, Go programs listen with MPTCP by default on Linux.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-it-is&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What it is&lt;/h2&gt;
&lt;p&gt;It’s a race in the kernel’s IPv4 TCP stack. When it accepts a connection, the kernel hands the new socket the IPv4 options that came with the request, without copying them. If a retransmitted SYN arrives at that moment, the kernel can keep reading them to build its reply, the SYN-ACK, even after that socket has been freed.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It’s a use-after-free in &lt;code&gt;ip_options_build()&lt;/code&gt;, and the same can happen when the kernel migrates pending connections from one socket to another.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The official record, from the kernel’s CVE team, scores it 9.8 out of 10: it’s attacked over the network, with no credentials and no user interaction.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It affects kernels since version 4.4.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Red Hat gives it the same score and rates it critical.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;the-exploit&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The exploit&lt;/h2&gt;
&lt;p&gt;The lab that described it says the exploit was written by its AI model, and that it runs commands as root, without credentials, on a Linux server listening with MPTCP.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; There are no reports of the flaw being exploited,&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; and the US agency CISA doesn’t have it in its catalog of exploited vulnerabilities.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The Go team warns that the flaw isn’t specific to MPTCP: a server that doesn’t use it could fall to similar attacks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-go-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why Go matters&lt;/h2&gt;
&lt;p&gt;MPTCP is a TCP extension for using several network paths at once, and since Go 1.24, Go programs turn it on by default on Linux when they open a port.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; That’s why the Go team published how to switch it off in Go programs, while warning that doing so is fragile: a single exposed MPTCP service is enough to leave the system vulnerable.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Update the kernel.&lt;/strong&gt; That’s what the Go team recommends.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The fixed versions are 5.10.272, 5.15.223, 6.1.190, 6.6.159, 6.12.113, 6.18.56, 7.2.10 and 7.3.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Check what your distribution says&lt;/strong&gt;, because not all of them have the patch:
&lt;ul&gt;
&lt;li&gt;Debian has shipped it: in bookworm, with 6.1.190-1 (DLA-4841-1) or 6.12.113-1~deb12u1 (DLA-4850-1), and in trixie, with 6.12.113-1 (DSA-6568-1).&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Ubuntu rates it high priority, and as of October 30 had the patch in progress for its 22.04, 24.04, 26.04 and 26.10 releases.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Red Hat, as of October 30, had no fixed kernels for RHEL 8, 9 or 10, and says no mitigation meets its criteria.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;If you can’t update yet,&lt;/strong&gt; the oss-security list proposed switching MPTCP off: in each Go program, with &lt;code&gt;GODEBUG=multipathtcp=0&lt;/code&gt;,&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref20&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; or system-wide, with &lt;code&gt;sysctl -w net.mptcp.enabled=0&lt;/code&gt;.&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref21&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt; It has two limits: it affects programs that use MPTCP,&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref22&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt; and it assumes every attack goes through MPTCP, when the flaw isn’t specific to it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref23&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-we-dont-know-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What we don’t know yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Whether there’s another way to exploit it without MPTCP.&lt;/strong&gt; The Go team doesn’t rule it out.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref24&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;When Ubuntu and Red Hat will have the patch.&lt;/strong&gt; Ubuntu gives no date,&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref25&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt; and neither does Red Hat.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref26&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The attack that was shown needs an MPTCP service within the attacker’s reach, and a Go program can open one without anyone asking for it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref27&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Go security team, “Protecting against CVE-2026-98432 when using Go programs”, golang-announce, October 29, 2026, forwarded to the oss-security list. &lt;a href=&quot;https://www.openwall.com/lists/oss-security/2026/10/30/1&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.openwall.com/lists/oss-security/2026/10/30/1&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CVE Program, “CVE-2026-98432: ipv4: free inet_opt and ireq_opt after an RCU grace period”, published October 15, 2026 and updated October 30, 2026. &lt;a href=&quot;https://www.cve.org/CVERecord?id=CVE-2026-98432&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cve.org/CVERecord?id=CVE-2026-98432&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Red Hat, “CVE-2026-98432”, updated October 30, 2026. &lt;a href=&quot;https://access.redhat.com/security/cve/CVE-2026-98432&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://access.redhat.com/security/cve/CVE-2026-98432&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Daily CyberSecurity, “Linux Kernel CVE-2026-98432 Flaw Enables Remote Root”, October 30, 2026. &lt;a href=&quot;https://securityonline.info/linux-kernel-cve-2026-98432/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://securityonline.info/linux-kernel-cve-2026-98432/&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”, checked October 30, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref11&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;Debian Security Tracker, “CVE-2026-98432”, checked October 30, 2026. &lt;a href=&quot;https://security-tracker.debian.org/tracker/CVE-2026-98432&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://security-tracker.debian.org/tracker/CVE-2026-98432&lt;/a&gt;&lt;a href=&quot;#fnref17&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;Ubuntu, “CVE-2026-98432”, updated October 30, 2026. &lt;a href=&quot;https://ubuntu.com/security/CVE-2026-98432&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://ubuntu.com/security/CVE-2026-98432&lt;/a&gt;&lt;a href=&quot;#fnref18&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn8&quot;&gt;&lt;p&gt;John Haxby, “Re: Protecting against CVE-2026-98432 when using Go programs”, oss-security, October 30, 2026. &lt;a href=&quot;https://www.openwall.com/lists/oss-security/2026/10/30/6&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.openwall.com/lists/oss-security/2026/10/30/6&lt;/a&gt;&lt;a href=&quot;#fnref21&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>AI agents, directed by a single person, broke into the foreign ministries of two Latin American countries</title>
    <link href="https://blog.cisent.ai/en/vapor-panda-agentes-de-ia-cancillerias/"/>
    <id>tag:cisent.ai,2026-10-29:/en/vapor-panda-agentes-de-ia-cancillerias/</id>
    <published>2026-10-29T12:00:00+00:00</published>
    <updated>2026-10-29T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>According to CrowdStrike, a China-linked espionage group had AI agents do the intrusion work at many organizations at once, while one person guided them at key decision points.</summary>
    <content type="html">&lt;p&gt;On October 29, 2026, CrowdStrike published what it saw of Vapor Panda, an espionage group it links to China: since October 17, its AI agents had been attacking organizations in at least 14 countries at the same time, and a single person was directing them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Among them were the foreign ministries of two Latin American countries, where the group carried out extensive intrusions, and probably a third.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;In 11 days, the campaign targeted more than 200 organizations, although CrowdStrike does not explain how it arrived at that figure. The targets included foreign ministries in Latin America, Europe, Asia and the Middle East; US semiconductor, defense, nuclear, pharmaceutical and financial companies; think tanks, European government bodies and telecommunications providers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;CrowdStrike confirmed post-exploitation activity at nine organizations, where the group installed its backdoor, BLOODALCHEMY, and saw password spraying or credential scanning against at least twelve more. It also says the group compromised telecommunications providers in the United States, Belgium and the Philippines.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-the-agents-worked&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How the agents worked&lt;/h2&gt;
&lt;p&gt;They used ARTEX, a framework in which a main agent coordinates other agents that plan and execute, over a shared record of facts, findings and goals. The agents moved from initial access to reconnaissance of each machine and lateral movement, tried several paths at once and switched tactics when one failed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The person stepped in mostly at decision points. In the conversations CrowdStrike recovered, the operator asked the agents to write web shells, extract the domain administrator passwords and find out which machines had security sensors. In a single intrusion, the planning agents processed about 1.4 billion input tokens.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The main model the group used was one developed in China. It also used a second commercial, subscription-based model that CrowdStrike could not identify.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-they-got-in&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they got in&lt;/h2&gt;
&lt;p&gt;CrowdStrike describes some of the entry points:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;at a US pharmaceutical company, an unsigned SAML assertion and an XXE flaw opened the cloud metadata service to them, and from there they reached an administrator account;&lt;/li&gt;
&lt;li&gt;at another pharmaceutical company, web shells and an unauthenticated orchestration service left them as root in a container environment;&lt;/li&gt;
&lt;li&gt;at a Greek government body, they forged an administrator token with keys from a 2022 directory export that were never rotated, and used VPN configurations and a client key leaked in May 2024.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;For CrowdStrike it is a leap: from AI-assisted intrusions, one at a time, to several simultaneous ones directed by a single operator. It considers it highly likely that other China-linked groups will adopt frameworks like ARTEX over the next twelve months.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;ARTEX had appeared before. DIVD, the Dutch institute that warns organizations about their exposed flaws, said the tool used in the attack it suffered in September, which &lt;a href=&quot;https://blog.cisent.ai/en/divd-atacado-con-ia-por-zammad/&quot;&gt;we wrote about on October 1&lt;/a&gt;, was ARTEX, and that it acted autonomously.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Microsoft has also reported ARTEX activity by another group, Storm-0940, which CrowdStrike treats as separate from this campaign.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-isnt-known-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What isn’t known yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Which foreign ministries they are, and what information the group took.&lt;/li&gt;
&lt;li&gt;How CrowdStrike arrived at more than 200 organizations, and whether the nine confirmed include the telecommunications providers.&lt;/li&gt;
&lt;li&gt;Which second model the group used.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Among the entry points CrowdStrike describes are password spraying and keys that were never rotated, and the agents changed course as soon as one path failed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CrowdStrike Counter Adversary Operations, “VAPOR PANDA Uses AI-Driven ARTEX to Scale a Global Espionage Campaign”, October 29, 2026. &lt;a href=&quot;https://www.crowdstrike.com/en-us/blog/vapor-panda-uses-artex-to-scale-global-campaign/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.crowdstrike.com/en-us/blog/vapor-panda-uses-artex-to-scale-global-campaign/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Bill Toulas, BleepingComputer, “Chinese hackers use ARTEX AI in automated attacks against 200 orgs”, November 2, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/chinese-hackers-use-artex-ai-in-automated-attacks-against-200-orgs/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/chinese-hackers-use-artex-ai-in-automated-attacks-against-200-orgs/&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Cisco confirms a flaw that reboots its ASA and FTD firewalls without credentials is being exploited</title>
    <link href="https://blog.cisent.ai/en/cisco-asa-ftd-falla-explotada-en-la-vpn/"/>
    <id>tag:cisent.ai,2026-10-28:/en/cisco-asa-ftd-falla-explotada-en-la-vpn/</id>
    <published>2026-10-28T12:00:00+00:00</published>
    <updated>2026-10-28T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A single request to the VPN reboots the device and cuts the company off from remote access, but only if a connection profile uses secondary authentication.</summary>
    <content type="html">&lt;p&gt;On October 28, 2026, Cisco published an advisory for CVE-2026-76418, a flaw in the remote access SSL VPN of its Secure Firewall ASA and Secure Firewall Threat Defense (FTD) firewalls.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Its incident response team learned that same month that it is being exploited.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The next day, the US agency CISA added it to its catalog of exploited vulnerabilities and gave the country’s federal agencies until November 1 to apply Cisco’s measures.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;An attacker with no username or password sends a crafted request to the VPN, and the device reloads unexpectedly.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The bug sits in the logic that handles secondary authentication on connection profiles, the &lt;em&gt;tunnel groups&lt;/em&gt;.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It doesn’t steal data or grant access: the damage is the device going down, which leaves everyone who depends on it without service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Cisco rates it 8.6 out of 10 and doesn’t say who is using it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It’s the second flaw of this kind in the same VPN since August. That month, Cisco warned of CVE-2026-20349, which also reloaded the device with a request that needed no credentials and was also being exploited.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; CISA added it to its catalog on August 11.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;who-is-exposed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who is exposed&lt;/h2&gt;
&lt;p&gt;According to Cisco, a device is exposed if it meets three conditions:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;it runs a vulnerable ASA or FTD release;&lt;/li&gt;
&lt;li&gt;its SSL VPN answers anyone who tries to log in (&lt;code&gt;webvpn&lt;/code&gt;, enabled on at least one interface);&lt;/li&gt;
&lt;li&gt;a connection profile has secondary authentication (&lt;code&gt;secondary-authentication-server-group&lt;/code&gt;).&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Every profile has to be checked: the default one, &lt;code&gt;DefaultWEBVPNGroup&lt;/code&gt;, too, and the ones that don’t show on the login page.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Reaching the flaw doesn’t require logging in.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; On FTD, what you check is the VPN policy deployed from the management console.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Cisco confirms that its cloud-delivered firewall, Secure Access and the Secure Firewall Management Center console are not affected.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Update.&lt;/strong&gt; Cisco released fixed versions for every branch:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;
&lt;ul&gt;
&lt;li&gt;ASA: 9.16.4.104, 9.18.4.96, 9.20.4.50, 9.22.3.200, 9.23.1.48, and 9.24.1.27 or 9.24.10.1;&lt;/li&gt;
&lt;li&gt;FTD: 7.0.11, 7.2.13, 7.4.9, 7.6.7, 7.7.14, 10.0.3 and 10.1.1;&lt;/li&gt;
&lt;li&gt;branches older than ASA 9.16 and FTD 7.0 move to the first fixed release.&lt;/li&gt;
&lt;/ul&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;If you can’t update yet.&lt;/strong&gt; Cisco says there is no workaround, but it offers a mitigation: remove secondary authentication.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;
&lt;ul&gt;
&lt;li&gt;If the default profile doesn’t use it, disable the profile list (&lt;code&gt;no tunnel-group-list enable&lt;/code&gt;). Users will have to type the name of their profile.&lt;/li&gt;
&lt;li&gt;If the default profile uses it, remove it (&lt;code&gt;no secondary-authentication-server-group&lt;/code&gt;). That lowers the VPN’s security, and Cisco asks that it last only until you update.&lt;/li&gt;
&lt;/ul&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-we-dont-know-yet&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What we don’t know yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Who is exploiting it, and against whom.&lt;/strong&gt; The advisory doesn’t say.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Whether there were victims.&lt;/strong&gt; The advisory names none.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;To fall to this flaw, the VPN has to be within the attacker’s reach and set up in a specific way: that can be checked before anyone tries it.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Cisco, “Cisco Secure Firewall Adaptive Security Appliance and Secure Firewall Threat Defense Software Remote Access SSL VPN Tunnel Denial of Service Vulnerability” (cisco-sa-asaftd-tunnel-dos-s6mgkb75), October 28, 2026. &lt;a href=&quot;https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-asaftd-tunnel-dos-s6mgkb75&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-asaftd-tunnel-dos-s6mgkb75&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CISA, “Known Exploited Vulnerabilities Catalog”: CVE-2026-76418, added October 29, 2026, and CVE-2026-20349, added August 11, 2026. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Cisco, “Cisco Secure Firewall Adaptive Security Appliance and Secure Firewall Threat Defense Software Remote Access SSL VPN Denial of Service Vulnerability” (cisco-sa-asaftd-vpn-dos-dzv4mQFF), August 11, 2026. &lt;a href=&quot;https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-asaftd-vpn-dos-dzv4mQFF&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://sec.cloudapps.cisco.com/security/center/content/CiscoSecurityAdvisory/cisco-sa-asaftd-vpn-dos-dzv4mQFF&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>“A state of emergency in cyberspace”: the wave of cyberattacks hitting Japan</title>
    <link href="https://blog.cisent.ai/en/japon-emergencia-en-el-ciberespacio/"/>
    <id>tag:cisent.ai,2026-10-27:/en/japon-emergencia-en-el-ciberespacio/</id>
    <published>2026-10-27T12:00:00+00:00</published>
    <updated>2026-10-27T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A string of breaches at major companies led the ruling party to speak of an emergency, and experts suspect AI is making the attackers&#x27; work easier.</summary>
    <content type="html">&lt;p&gt;In a matter of weeks, a string of cyberattacks on major companies put Japan on alert. On October 27, Emol ran the headline that the attacks are “drowning Japan.”&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;On October 9, at a meeting of his party, Masaaki Taira, a former digital minister who heads the cybersecurity strategy headquarters of the Liberal Democratic Party, the ruling party, said the situation could be called “a state of emergency in cyberspace” and that the whole government had to respond.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It was the party speaking: the government has not declared an emergency.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to a study by TrendAI, a cybersecurity company, 600 cyberattacks were reported in Japan in the first nine months of 2026, more than in all of 2025.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Some of the cases:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Times Car&lt;/strong&gt;, Park24’s car-sharing service, confirmed that data from about 6.6 million accounts was taken, including driver’s license details and images of identity documents.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;IDC Frontier&lt;/strong&gt;, SoftBank’s cloud, was hit by ransomware that reached almost 500 companies and local governments.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; &lt;a href=&quot;https://blog.cisent.ai/en/idc-frontier-nube-ransomware/&quot;&gt;We covered it here&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;GMO Internet&lt;/strong&gt; won’t pay the ransom demanded over the attack on its desktop cloud. &lt;a href=&quot;https://blog.cisent.ai/en/gmo-compensa-y-cae-en-bolsa/&quot;&gt;We covered it here&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Keio&lt;/strong&gt;, the Tokyo railway, had to shut down its network to stop a ransomware attack. &lt;a href=&quot;https://blog.cisent.ai/en/keio-ransomware-red-apagada/&quot;&gt;We covered it here&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Just seven of the reported breaches, including those at JR East and Lawson, add up to more than 55 million records, although most are maximum figures and count records, not people.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;and-ai&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;And AI?&lt;/h2&gt;
&lt;p&gt;Nobody has confirmed that these attacks used AI. No company or authority has said so, and the head of the National Police Agency, Yoshinobu Kusunoki, said on October 8 that it was still unclear what links the cases.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;What experts do see are two changes. The first is language: it used to be obvious in an email or a phone call when someone didn’t speak Japanese well, and AI translations make phishing more convincing, according to Takahiro Anno, a technologist and founder of an AI-focused political party.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; The second is scale: automated systems now scan websites all over the world for flaws, around the clock.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Ko Fujii, founder of the public affairs agency Makaira, concluded in a preliminary analysis that a coordinated, state-sponsored attack is unlikely: Japan had so many breaches because its unpatched infrastructure gave attackers the best return.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-japan-is-doing&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What Japan is doing&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;On October 9&lt;/strong&gt;, the National Cybersecurity Office asked companies that hold large amounts of personal data to patch quickly, use multi-factor authentication, vet their contractors and limit the data they share with them.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The same day&lt;/strong&gt;, the Financial Services Agency asked banks and financial firms to check more carefully the document and face images they use to verify identity remotely, because the leaked data includes images of driver’s licenses. It also asked them to move early to reading the document’s chip: the rules end the sending of images only from April 1, 2027.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;
&lt;ul&gt;
&lt;li&gt;SBI Shinsei Bank, for example, accepted photos of documents and faces until October 27. From the 28th it only reads the chip of the My Number card or the driver’s license.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;In early October&lt;/strong&gt;, a cyber defense law took effect that lets the police and the Self-Defense Forces monitor and disable foreign servers, and requires companies to report their breaches.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;On October 20&lt;/strong&gt;, the ruling party drafted emergency proposals: a headquarters against cyberattacks under the prime minister’s office, budget and laws before AI makes simultaneous attacks possible, and public salaries that attract ethical hackers.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;What keeps repeating in Japan is not a new technique: it is unpatched infrastructure and basic controls nobody checked, and now there are systems looking for them around the clock.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service that was published by mistake, the insecure configuration. It is ongoing testing, not one audit a year, and every new attack pattern helps train the defense. It runs on Manthys Sentinel models trained for red teaming.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Emol Multimedia Team, “Los ataques potenciados con IA ahogan a Japón: Cómo el país llegó a declarar ‘estado de emergencia en el ciberespacio’,” October 27, 2026. &lt;a href=&quot;https://www.emol.com/noticias/Internacional/2026/10/27/1215246/japon-ciberataques-ia-empresas-seguridad.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.emol.com/noticias/Internacional/2026/10/27/1215246/japon-ciberataques-ia-empresas-seguridad.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Darie Nani, Sovereign Magazine, “Japan’s Ruling-Party Cyber Chief Likens a Wave of Corporate Data Breaches to a State of Emergency in Cyberspace,” October 15, 2026, updated October 26; Taira’s remark as reported by TV Asahi. &lt;a href=&quot;https://www.sovereignmagazine.com/article/japan-cyber-emergency-data-breaches&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.sovereignmagazine.com/article/japan-cyber-emergency-data-breaches&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Tiffany May and Kiuko Notoya, The New York Times, “Ransomware attacks have swept Japan. AI may be making them easier,” October 13, 2026, also published by The Star on October 14. &lt;a href=&quot;https://www.thestar.com.my/tech/tech-news/2026/10/14/ransomware-attacks-have-swept-japan-ai-may-be-making-them-easier&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.thestar.com.my/tech/tech-news/2026/10/14/ransomware-attacks-have-swept-japan-ai-may-be-making-them-easier&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Bill Toulas, BleepingComputer, “Times Car confirms data breach affecting 6.6 million user accounts,” September 28, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/times-car-confirms-data-breach-affecting-66-million-user-accounts/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/times-car-confirms-data-breach-affecting-66-million-user-accounts/&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;Financial Services Agency of Japan (金融庁), “現下の情勢を踏まえたサイバーセキュリティ対策の強化と各種取引申込等における対応について,” October 9, 2026. &lt;a href=&quot;https://www.fsa.go.jp/news/r8/sonota/20261009/20261009.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.fsa.go.jp/news/r8/sonota/20261009/20261009.html&lt;/a&gt;&lt;a href=&quot;#fnref13&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;Impress Watch, “SBI新生銀、口座開設の本人確認をICチップに一本化　顔写真撮影を廃止,” October 22, 2026. &lt;a href=&quot;https://www.watch.impress.co.jp/docs/news/2150091.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.watch.impress.co.jp/docs/news/2150091.html&lt;/a&gt;&lt;a href=&quot;#fnref14&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;Jiji Press, published by Nippon.com: “LDP Urges Govt to Address Cyberattacks on Companies,” October 20, 2026. &lt;a href=&quot;https://www.nippon.com/en/news/yjj2026102001054/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.nippon.com/en/news/yjj2026102001054/&lt;/a&gt;&lt;a href=&quot;#fnref16&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A Drupal site has two critical flaws to patch: its core&#x27;s, and PHP&#x27;s, which are already being exploited</title>
    <link href="https://blog.cisent.ai/en/drupal-y-php-dos-fallas-por-parchar/"/>
    <id>tag:cisent.ai,2026-10-26:/en/drupal-y-php-dos-fallas-por-parchar/</id>
    <published>2026-10-26T12:00:00+00:00</published>
    <updated>2026-10-26T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The United States confirmed that October&#x27;s critical PHP flaws are being exploited, and Drupal, which runs on PHP, had fixed a critical flaw in its own core.</summary>
    <content type="html">&lt;p&gt;On October 26, 2026, CISA, the US cybersecurity agency, added to its catalog of exploited vulnerabilities the two critical flaws PHP fixed on October 21, CVE-2026-109618 and CVE-2026-109619.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; They allow code to run on the server without logging in, and &lt;a href=&quot;https://blog.cisent.ai/en/php-dos-fallas-criticas/&quot;&gt;we covered them here&lt;/a&gt;. A Drupal site, which runs on PHP, also has a critical flaw of its own: the one Drupal fixed in its core on October 14.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;php-already-being-exploited&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;PHP: already being exploited&lt;/h2&gt;
&lt;p&gt;CISA’s catalog lists the vulnerabilities known to be used in attacks, and since October 26 it includes the two PHP flaws.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The fixed versions are PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Updating Drupal does not update PHP. The PHP version is managed by whoever runs the server or by the hosting provider, so the first question is who maintains it.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;drupal-a-critical-flaw-in-its-core&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Drupal: a critical flaw in its core&lt;/h2&gt;
&lt;p&gt;Advisory SA-CORE-2026-014 describes a flaw in Field UI, a module that ships with Drupal core, that allows code to run on the server (CVE-2026-108097). Drupal rates it critical, 19 out of 25 on its risk scale.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The flaw cannot be exploited with core alone: according to the Drupal team, there are several serious paths through contributed modules, and at least one does not require logging in. A site’s own code can also expose it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The advisory does not say which modules those are.&lt;/p&gt;
&lt;p&gt;The fixed versions are 11.4.10, 11.3.19 and 10.6.19. Drupal 11.2 and earlier, 10.5 and earlier, and Drupal 8 and 9 no longer receive security fixes.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; If updating right away isn’t possible, uninstalling the Field UI module protects the site in the meantime.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Drupal also fixed a similar flaw in Entity List, a contributed module, in its version 3.0.10.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;As of October 26, the Drupal core flaw is not in CISA’s catalog.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; When it fixed the flaw, the Drupal team warned that exploits could appear “within hours or days.”&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A Drupal site rests on three layers: the web server, PHP and Drupal. In October, all three got critical patches: if the server is nginx, F5 fixed five flaws on October 14, three of them critical, which &lt;a href=&quot;https://blog.cisent.ai/en/nginx-cinco-fallas-fuera-de-calendario/&quot;&gt;we also covered&lt;/a&gt;.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Each layer is updated by someone different: Drupal, by the site’s team; PHP and the web server, by whoever runs the server or by the hosting provider. One being up to date says nothing about the others.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;An outdated PHP or Drupal on a server nobody checks is the kind of gap worth finding before an attacker does.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CISA, Known Exploited Vulnerabilities Catalog, version 2026.10.26, October 26, 2026: it added CVE-2026-109618 and CVE-2026-109619 that day, and CVE-2026-108097 is not listed. &lt;a href=&quot;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/known-exploited-vulnerabilities-catalog&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Drupal Security Team, “Drupal core - Critical - Remote code execution - SA-CORE-2026-014”, October 14, 2026. &lt;a href=&quot;https://www.drupal.org/sa-core-2026-014&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.drupal.org/sa-core-2026-014&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;PHP security team, security releases PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35, October 21, 2026. &lt;a href=&quot;https://www.php.net/archive/2026.php&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.php.net/archive/2026.php&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;BleepingComputer, “Drupal patches critical RCE flaws, warns exploits may emerge soon”, October 17, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/drupal-patches-critical-rce-flaws-warns-exploits-may-emerge-soon/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/drupal-patches-critical-rce-flaws-warns-exploits-may-emerge-soon/&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;F5, “Out-of-band Security Notification (October 14, 2026)”, K000163408. &lt;a href=&quot;https://my.f5.com/manage/s/article/K000163408&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://my.f5.com/manage/s/article/K000163408&lt;/a&gt;&lt;a href=&quot;#fnref12&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>GMO will compensate customers after the cyberattack, and its CEO admits the response fell short</title>
    <link href="https://blog.cisent.ai/en/gmo-compensa-y-cae-en-bolsa/"/>
    <id>tag:cisent.ai,2026-10-26:/en/gmo-compensa-y-cae-en-bolsa/</id>
    <published>2026-10-26T12:00:00+00:00</published>
    <updated>2026-10-26T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The Japanese company won&#x27;t pay the ransom it&#x27;s being asked for, its shares tumbled in Tokyo, and its CEO wrote that the internet itself is starting to break down.</summary>
    <content type="html">&lt;p&gt;On October 26, 2026, GMO Internet’s shares fell almost 16% on the Tokyo Stock Exchange, to a four-month low. Over the weekend, the Japanese company had announced it will compensate customers for a cyberattack, and that Monday its CEO, Masatoshi Kumagai, said the attackers are demanding US$50 million in cryptocurrency.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The attack hit Onamae.com Desktop Cloud, a cloud desktop service that many customers use to trade currencies. GMO started investigating on October 19 after a customer inquiry, confirmed malicious software on customer servers on the 20th, and began shutting them all down on the 21st. According to Kumagai, up to 26,739 servers may have been affected, and there were unauthorized trades in accounts at more than 70 securities and currency brokers.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; We covered it in &lt;a href=&quot;https://blog.cisent.ai/en/octubre-negro-todos-vulnerables/&quot;&gt;Black October&lt;/a&gt;.&lt;/p&gt;
&lt;p&gt;The attackers are threatening to publish company and customer data. GMO won’t pay and is working with the police.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;GMO will cover the losses from the unauthorized trades and pay every customer 5,000 yen, about US$32, whether or not they lost money. It’s also offering a free alternative service for six months.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-its-ceo-admitted&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What its CEO admitted&lt;/h2&gt;
&lt;p&gt;On his blog, Kumagai said that white-hat hackers working for the company had detected security problems before the attack, but that the response lacked the necessary depth and speed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; “The internet itself is now starting to break down,” he wrote.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Prime Minister Sanae Takaichi said the government will put together an emergency package against cyberattacks by next week.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The warning came before the damage, and it wasn’t enough: the company itself admits its specialists saw the problems, but the response wasn’t deep or fast enough.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Between seeing a problem and acting on it there’s a gap, and everything else played out in that gap: the ransom, the compensation and the fall in the stock.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; doesn’t stop at the alert: it detects, assesses and acts in seconds, alongside the EDR, the firewall and every connected source. What can be undone it does without waiting, such as blocking an external attacking IP or cutting an intrusive connection; anything with impact, such as isolating a machine, it asks about first. Every action is recorded, and every new technique it contains becomes a rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Kentaro Okasaka and Sam Nussey, Reuters, “Japan companies compensate users as cyberattacks buffet economy”, October 26, 2026, published by Insurance Journal. &lt;a href=&quot;https://www.insurancejournal.com/news/international/2026/10/26/890257.htm&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.insurancejournal.com/news/international/2026/10/26/890257.htm&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Euronews, “Empresa japonesa compensa a usuarios afectados por ciberataque” (in Spanish), October 26, 2026. &lt;a href=&quot;https://es.euronews.com/2026/10/26/empresa-japonesa-compensa-a-usuarios-afectados-por-ciberataque&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://es.euronews.com/2026/10/26/empresa-japonesa-compensa-a-usuarios-afectados-por-ciberataque&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Sweden&#x27;s government blocked ChatGPT and Claude for its staff: there was no agreement to protect its data</title>
    <link href="https://blog.cisent.ai/en/gobierno-de-suecia-bloquea-asistentes-de-ia/"/>
    <id>tag:cisent.ai,2026-10-26:/en/gobierno-de-suecia-bloquea-asistentes-de-ia/</id>
    <published>2026-10-26T12:00:00+00:00</published>
    <updated>2026-10-26T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Staff at the Swedish Government Offices can no longer use ChatGPT or Claude in the browser on their work computers; Copilot stays, because it does have a security agreement with the government.</summary>
    <content type="html">&lt;p&gt;Since Monday, October 26, 2026, the Government Offices of Sweden (Regeringskansliet) have blocked OpenAI’s ChatGPT and Anthropic’s Claude on their staff’s computers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The block covers the versions used in the browser.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The government wants its staff to use only approved services.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;no-agreement-no-access&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;No agreement, no access&lt;/h2&gt;
&lt;p&gt;Mikael Wettercrantz, the Government Offices’ director of digitalization, wrote on the intranet that using AI services without an agreement can have “consequences for our information security.”&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to Euronews, the concern is that sensitive information typed into a service without a data protection agreement could end up stored and used to train the model.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The assistant that stays is Microsoft Copilot: it passed a security review and is covered by a security agreement between Microsoft and the Swedish government,&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; which also tightened its security.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; Some staff will keep using ChatGPT and Claude, but only through an AI assistant platform the government built itself.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;not-just-the-government&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Not just the government&lt;/h2&gt;
&lt;p&gt;According to Ekot, Swedish public radio’s news service, the Swedish Tax Agency (Skatteverket) and the National Board of Institutional Care (Statens institutionsstyrelse) have also blocked them, and the Swedish Enforcement Authority (Kronofogden) may follow.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The decision is not against AI: the government kept an assistant under agreement and built a platform of its own. What it cuts off is the use of services nobody contracted. When someone pastes a document into an AI their organization never contracted, the organization doesn’t know where that information ends up or what it’s used for.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;The underlying problem is not knowing what happens to the information typed into someone else’s AI.&lt;/p&gt;
&lt;p&gt;Many AI security tools process your logs in third-party models. &lt;strong&gt;Cisent uses its own models, with no outside AI provider in between:&lt;/strong&gt; the &lt;strong&gt;Manthys&lt;/strong&gt; models, four families trained by Cisent. Mind reasons, Sentinel defends, Nova brings the scale and Reflex filters, and they work inside Cisent alongside Husker, the agentic layer. We don’t train on your data. Not even anonymized.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/soberania-de-los-datos&quot;&gt;See how we protect your data&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Viktor Eriksson, Computer Sweden, “Regeringskansliet spärrar Chat GPT och Claude,” October 23, 2026, reporting Ekot, Sveriges Radio. &lt;a href=&quot;https://computersweden.se/article/4239687/regeringskansliet-sparrar-chat-gpt-och-claude.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://computersweden.se/article/4239687/regeringskansliet-sparrar-chat-gpt-och-claude.html&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Jonathan Benton, Euronews, “Swedish government blocks ChatGPT and Claude. Here’s why,” October 26, 2026. &lt;a href=&quot;https://www.euronews.com/2026/10/26/swedish-government-blocks-chatgpt-and-claude-heres-why&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.euronews.com/2026/10/26/swedish-government-blocks-chatgpt-and-claude-heres-why&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Finwire in Börsvärlden, “Regeringskansliet spärrar användning av ChatGPT och Claude – Ekot,” October 23, 2026. &lt;a href=&quot;https://borsvarlden.com/artiklar/regeringskansliet-sparrar-anvandning-av-chatgpt-och-claude-ekot&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://borsvarlden.com/artiklar/regeringskansliet-sparrar-anvandning-av-chatgpt-och-claude-ekot&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>OpenAI says its AI refuted a 1999 conjecture. What matters is how it gets checked</title>
    <link href="https://blog.cisent.ai/en/openai-dice-que-su-ia-refuto-la-eth/"/>
    <id>tag:cisent.ai,2026-10-26:/en/openai-dice-que-su-ia-refuto-la-eth/</id>
    <published>2026-10-26T12:00:00+00:00</published>
    <updated>2026-10-26T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="ciencia" label="Research"/>
    <summary>OpenAI says one of its models refuted the Exponential Time Hypothesis. It is under review, and what gives it weight is not the AI&#x27;s confidence but the fact that it can be checked.</summary>
    <content type="html">&lt;p&gt;On October 26, OpenAI announced&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; that an unnamed internal model wrote a manuscript titled “The Exponential Time Hypothesis is False”.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; If it survives review, a 1999 conjecture that many results in complexity theory rest on will fall. It is not confirmed yet. It will be examined by Russell Impagliazzo and Ramamohan Paturi, who proposed the conjecture, together with Ryan Williams of MIT.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-the-conjecture-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the conjecture says&lt;/h2&gt;
&lt;p&gt;Its ground is 3-SAT, one of the most studied problems in computer science. You have a formula made of many small conditions, each on three variables that can only be true or false, and the question is whether there is a way to set the variables so that all of them hold. With n variables there are 2ⁿ possible combinations.&lt;/p&gt;
&lt;p&gt;The Exponential Time Hypothesis, put forward by Impagliazzo and Paturi in 1999,&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; says there is no fundamental shortcut: you can be smarter than trying them all, but the time needed will keep growing exponentially with the number of variables. Many results in complexity build on it: they say something cannot be done faster, if the conjecture is true.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-openai-claims&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What OpenAI claims&lt;/h2&gt;
&lt;p&gt;The manuscript presents a deterministic algorithm that counts exactly how many solutions formulas of this kind have, with conditions on three variables or any other fixed number, in a time that grows more slowly than any exponential and with memory that grows polynomially. The three-variable case, 3-SAT, is the one that refutes the conjecture.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It does not settle P versus NP: the algorithm is faster than any exponential, but still slower than a polynomial one. That problem remains open.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-can-be-checked&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it can be checked&lt;/h2&gt;
&lt;p&gt;A model can write a false proof with the same confidence as a true one. So what matters is not that an AI says it, but what surrounds the claim:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;A statement written separately.&lt;/strong&gt; The theorem is written in Lean, a language in which a computer checks mathematical proofs, in a file separate from the proof.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A checker.&lt;/strong&gt; A comparator verifies that the proof proves exactly that statement, and only lets it use three standard axioms: no shortcuts.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt; The proof spans 3,977 Lean files.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The materials, in the open.&lt;/strong&gt; OpenAI released them right away, at the request of the group that will examine the proof, so that others can review them too.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;An independent review.&lt;/strong&gt; Impagliazzo, Paturi and Williams will examine the proof and explain it.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Examining it is not endorsing it: that work is just starting.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;At Cisent we reviewed the formal statement, and it says what they announce. We did not run the full proof: OpenAI publishes it as the formal proof of that statement.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;confidence-is-not-proof&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Confidence is not proof&lt;/h2&gt;
&lt;p&gt;The same repository shows why all of that is needed. It holds 721 manuscripts from the same model, and about 42% of the main results have a formal proof. OpenAI warns that some of those without one could have issues.&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt; On October 7 it withdrew three manuscripts over a sign error and revised fourteen others.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The one on the conjecture is among those that come with a formal proof.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-does-not-change&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What does not change&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Encryption.&lt;/strong&gt; Everyday cryptography does not rest on this conjecture, but on other assumptions.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;P versus NP.&lt;/strong&gt; Still open.&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;the-rule-that-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The rule that matters&lt;/h2&gt;
&lt;p&gt;What makes an AI’s claim credible is not how confidently it is made, but whether it can be checked: a statement written separately, a checker, the materials in the open and a review that does not depend on whoever makes the claim. It is the rule we use to publish our own figures, each with the source behind it. And it is the idea behind our research on &lt;a href=&quot;https://blog.cisent.ai/en/el-no-se-penalizado-tres-veces/&quot;&gt;saying “I don’t know”&lt;/a&gt;: a model that knows when it does not know is worth more than one that always sounds sure. &lt;a href=&quot;https://ciencia.cisent.ai/en/#publicaciones&quot;&gt;Our publications are in Research&lt;/a&gt;.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-we-dont-know-yet&quot; class=&quot;level2 b-nosabemos&quot;&gt;
&lt;h2&gt;What we don’t know yet&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Whether the proof survives the examination by Impagliazzo, Paturi and Williams.&lt;/li&gt;
&lt;li&gt;Whether it passes the checker in other hands: we did not run it.&lt;/li&gt;
&lt;li&gt;From what formula size the algorithm beats exhaustive search: OpenAI did not publish its constants.&lt;/li&gt;
&lt;li&gt;How many of the other manuscripts will survive review: most still have no formal proof.&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;OpenAI, announcement of October 26, 2026. &lt;a href=&quot;https://openai.com/index/building-scientific-understanding-with-academic-communities/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://openai.com/index/building-scientific-understanding-with-academic-communities/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;OpenAI, openai/math repository, “History”, entries of October 7 and October 26, 2026 (version of October 26, commit 7ef5c0c). &lt;a href=&quot;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/history.md&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/history.md&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Russell Impagliazzo and Ramamohan Paturi, “Complexity of k-SAT”, Proceedings of the 14th Annual IEEE Conference on Computational Complexity, 1999, pp. 237-240 &lt;a href=&quot;https://doi.org/10.1109/CCC.1999.766282&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://doi.org/10.1109/CCC.1999.766282&lt;/a&gt;. Journal version: “On the Complexity of k-SAT”, Journal of Computer and System Sciences 62(2), 2001, pp. 367-375 &lt;a href=&quot;https://doi.org/10.1006/jcss.2000.1727&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://doi.org/10.1006/jcss.2000.1727&lt;/a&gt;.&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;OpenAI, the formal statement in Lean, “SubexponentialSAT.lean”. &lt;a href=&quot;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/ComparatorChallenges/SubexponentialSAT.lean&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/ComparatorChallenges/SubexponentialSAT.lean&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;OpenAI, the comparator configuration, “SubexponentialSAT.json”: it only permits the axioms propext, Quot.sound and Classical.choice. &lt;a href=&quot;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/ComparatorChallenges/SubexponentialSAT.json&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/ComparatorChallenges/SubexponentialSAT.json&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;OpenAI, the proof’s folder, “lean/OAI/Algorithms/SubexponentialSAT”: 3,977 Lean files, counted by Cisent with the GitHub API at the cited version. &lt;a href=&quot;https://github.com/openai/math/tree/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/OAI/Algorithms/SubexponentialSAT&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/tree/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/OAI/Algorithms/SubexponentialSAT&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;OpenAI, the formalization catalogue, “formalization.yaml”, which links the manuscript to its Lean proof. &lt;a href=&quot;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/formalization.yaml&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/lean/formalization.yaml&lt;/a&gt;&lt;a href=&quot;#fnref11&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn8&quot;&gt;&lt;p&gt;OpenAI, openai/math repository, README (version of October 26). &lt;a href=&quot;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/README.md&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/openai/math/blob/7ef5c0cc5ca4b0e072b7a050fca09ab24bc6fa3a/README.md&lt;/a&gt;&lt;a href=&quot;#fnref12&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Black October: everyone is vulnerable</title>
    <link href="https://blog.cisent.ai/en/octubre-negro-todos-vulnerables/"/>
    <id>tag:cisent.ai,2026-10-24:/en/octubre-negro-todos-vulnerables/</id>
    <published>2026-10-24T12:00:00+00:00</published>
    <updated>2026-10-24T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>In two weeks, PHP, GitLab, Node.js, nginx, a key AI library and a certificate authority all failed at once. It was not a coordinated attack: it is the new pace, and no one is left out.</summary>
    <content type="html">&lt;p&gt;&lt;em&gt;Updated October 25: the Node.js flaw is in Brotli, a Google library that other programs use too. That is why there is more to update than Node (see “What to do this week”).&lt;/em&gt;&lt;/p&gt;
&lt;p&gt;The press and LinkedIn gave the second half of October 2026 a name: “Black October”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; It is not an official name, and it is not a campaign. It is a run of critical flaws and attacks, almost all at once, in pieces nearly every organization uses: the language of the web, the web server, the code repository, the AI libraries, the certificates and the cloud. On October 21 alone, PHP, GitLab and Node.js shipped emergency patches, and a widely used library for connecting applications to AI models woke up with malware.&lt;/p&gt;
&lt;section id=&quot;what-happened-day-by-day&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened, day by day&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Early October: Tor.&lt;/strong&gt; Attackers controlled 333 Tor exit relays and diverted SSH connections to 12 trap servers, which captured passwords in transit. The relays were flagged as malicious on October 4, and the researcher who found them disclosed it on the 20th.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 14: nginx.&lt;/strong&gt; F5 shipped out-of-band patches for the nginx web server.&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; Three of the flaws score 9.2 out of 10: an attacker with no credentials can restart the server’s processes and, in certain configurations, run code. Affected versions go up to 1.30.5 and 1.31.6.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt; We covered it &lt;a href=&quot;https://blog.cisent.ai/en/nginx-cinco-fallas-fuera-de-calendario/&quot;&gt;here&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 17: the first shot.&lt;/strong&gt; Amazon’s honeypots recorded the first attempt to exploit PHP2Shell, four days before the patch. Amazon still doesn’t know where the attacker got the details.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt; That same night, Mozilla was told that the private keys of eight intermediate certificates from the eMudhra certificate authority were compromised, and it blocked them in Firefox in under three hours.&lt;a href=&quot;#fn7&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;7&lt;/sup&gt;&lt;/a&gt; eMudhra says a step in its key ceremony produced a second, weak set of keys alongside the good ones, and that those certificates never issued a single one.&lt;a href=&quot;#fn8&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;8&lt;/sup&gt;&lt;/a&gt; The same day, one day after the flaw was published, attacks began on unpatched Plex servers to turn them into cryptocurrency miners.&lt;a href=&quot;#fn9&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;9&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 20: a Japanese cloud.&lt;/strong&gt; GMO Internet confirmed malware on customer servers of Onamae.com Desktop Cloud, a service for running trading bots. Up to 26,739 servers may have been exposed. GMO shut them all down the next day, and on the 22nd a Japanese forex broker delayed its customers’ withdrawals.&lt;a href=&quot;#fn10&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;10&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 21: the black day.&lt;/strong&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;PHP2Shell:&lt;/strong&gt; two flaws in PHP’s core, scored 9.2 out of 10, let an attacker take over a server without a password. They affect every version since 5.4.&lt;a href=&quot;#fn11&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;11&lt;/sup&gt;&lt;/a&gt; Chile’s national CSIRT issued alert AVC26-00740 that day.&lt;a href=&quot;#fn12&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;12&lt;/sup&gt;&lt;/a&gt; We covered it in detail &lt;a href=&quot;https://blog.cisent.ai/en/php-dos-fallas-criticas/&quot;&gt;here&lt;/a&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;GitLab:&lt;/strong&gt; a 10-out-of-10 flaw lets an attacker run code without logging in on self-managed GitLab servers, from version 7.0 onward.&lt;a href=&quot;#fn13&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;13&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Node.js:&lt;/strong&gt; ten flaws; the worst, in the Brotli decompressor, can lead to remote code execution.&lt;a href=&quot;#fn14&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;14&lt;/sup&gt;&lt;/a&gt; The flaw is not Node’s alone: it is in Brotli, Google’s compression library, in versions 1.0.3 through 1.2.0, and it reaches other programs that use it, such as curl. Google fixed it the same day, in versions 1.0.10, 1.1.1 and 1.2.1.&lt;a href=&quot;#fn15&quot; class=&quot;footnote-ref&quot; id=&quot;fnref15&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;15&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn16&quot; class=&quot;footnote-ref&quot; id=&quot;fnref16&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;16&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;LiteLLM:&lt;/strong&gt; version 1.107.0 of this library, which connects applications to AI models, shipped with credential-stealing malware. It was on PyPI for close to two hours, and on Docker Hub until 14:20 UTC.&lt;a href=&quot;#fn17&quot; class=&quot;footnote-ref&quot; id=&quot;fnref17&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;17&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Railway:&lt;/strong&gt; a third party may have accessed its customers’ encrypted environment variables, and obtained the decrypted contents of some. The platform paused all deployments and asked everyone to rotate every key.&lt;a href=&quot;#fn18&quot; class=&quot;footnote-ref&quot; id=&quot;fnref18&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;18&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;On top of that, the FBI and CISA warned about the SafePay ransomware, with more than 600 victims: we covered it &lt;a href=&quot;https://blog.cisent.ai/en/safepay-va-por-los-respaldos/&quot;&gt;here&lt;/a&gt;.&lt;/li&gt;
&lt;/ul&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;October 22 and 23: the accounts and the code.&lt;/strong&gt; Capcom reported access to 11,774,247 records of its account service, and 196 passwords changed without permission.&lt;a href=&quot;#fn19&quot; class=&quot;footnote-ref&quot; id=&quot;fnref19&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;19&lt;/sup&gt;&lt;/a&gt; Six reports of Linux kernel flaws, still unpatched, reached the developers’ public mailing list by mistake; two allow remote code execution.&lt;a href=&quot;#fn20&quot; class=&quot;footnote-ref&quot; id=&quot;fnref20&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;20&lt;/sup&gt;&lt;/a&gt; Big US banks found AI agents operating inside customer accounts, almost always with the password the customer handed over (&lt;a href=&quot;https://blog.cisent.ai/en/bancos-agentes-de-ia-en-cuentas/&quot;&gt;our note&lt;/a&gt;).&lt;a href=&quot;#fn21&quot; class=&quot;footnote-ref&quot; id=&quot;fnref21&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;21&lt;/sup&gt;&lt;/a&gt; And the PHP team warned of a third flaw, in session handling and still unpatched, that in certain configurations reaches even servers that were already updated.&lt;a href=&quot;#fn12&quot; class=&quot;footnote-ref&quot; id=&quot;fnref22&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;12&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;In those same days, a Signal flaw came to light that could leave the microphone, and perhaps the camera, on after hanging up, if the attacker was already on the call. It is fixed in current versions.&lt;a href=&quot;#fn22&quot; class=&quot;footnote-ref&quot; id=&quot;fnref23&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;22&lt;/sup&gt;&lt;/a&gt; And in California, someone used a stolen credential to get into the file site of the organization that runs the state tests, and copied data on public school students from the last three school years.&lt;a href=&quot;#fn23&quot; class=&quot;footnote-ref&quot; id=&quot;fnref24&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;23&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;ai-on-both-sides&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;AI, on both sides&lt;/h2&gt;
&lt;p&gt;PHP2Shell was found by Wiz researchers with help from AI models, which they used to find the flaws, write the test exploits and validate the patch. According to Wiz, a single researcher did in a week what used to take a world-class team months.&lt;a href=&quot;#fn11&quot; class=&quot;footnote-ref&quot; id=&quot;fnref25&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;11&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The Brotli flaw was reported by Codex Security, OpenAI’s AI agent that looks for flaws in code.&lt;a href=&quot;#fn14&quot; class=&quot;footnote-ref&quot; id=&quot;fnref26&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;14&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn24&quot; class=&quot;footnote-ref&quot; id=&quot;fnref27&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;24&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The same speed runs the other way. The first attack attempt came before the patch.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref28&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt; Railway believes its incident is part of a series of attacks on platforms, aimed at reaching AI inference providers.&lt;a href=&quot;#fn18&quot; class=&quot;footnote-ref&quot; id=&quot;fnref29&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;18&lt;/sup&gt;&lt;/a&gt; And a few days ago we covered &lt;a href=&quot;https://blog.cisent.ai/en/un-atacante-agentes-ia-cien-empresas/&quot;&gt;an attacker who left AI agents attacking on their own&lt;/a&gt;.&lt;/p&gt;
&lt;p&gt;Nor is the volume about one month. NIST’s vulnerability database lists 88,602 published in 2026 as of October 24, against 49,972 in all of 2025. As OpenSecurity puts it, AI finds flaws faster than ever, and patching is still on us.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref30&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-everyone-is-vulnerable&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why “everyone is vulnerable”&lt;/h2&gt;
&lt;p&gt;None of these flaws needed a user’s click. They were in what nearly every organization shares without knowing it, and what nobody looks at until it fails.&lt;/p&gt;
&lt;p&gt;Three things changed this month:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;The attack no longer waits for the patch.&lt;/strong&gt; With PHP2Shell, it came four days early.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref31&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A patch no longer closes the matter.&lt;/strong&gt; PHP’s third flaw reaches servers that had already been updated.&lt;a href=&quot;#fn12&quot; class=&quot;footnote-ref&quot; id=&quot;fnref32&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;12&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;It is no longer clear who is acting.&lt;/strong&gt; A bank can’t always tell whether the account is being run by the customer, by an agent the customer authorized, or by an attacker.&lt;a href=&quot;#fn21&quot; class=&quot;footnote-ref&quot; id=&quot;fnref33&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;21&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-do-this-week&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to do this week&lt;/h2&gt;
&lt;ol type=&quot;1&quot;&gt;
&lt;li&gt;&lt;strong&gt;PHP:&lt;/strong&gt; upgrade to 8.5.12, 8.4.27, 8.3.36 or 8.2.35, and ask your hosting provider who handles it. Until the third flaw is patched, Chile’s CSIRT recommends setting &lt;code&gt;session.upload_progress.enabled=0&lt;/code&gt;.&lt;a href=&quot;#fn12&quot; class=&quot;footnote-ref&quot; id=&quot;fnref34&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;12&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Self-managed GitLab:&lt;/strong&gt; 18.11.13, 19.0.10, 19.1.9, 19.2.9, 19.3.6, 19.4.3 or 19.5.1.&lt;a href=&quot;#fn13&quot; class=&quot;footnote-ref&quot; id=&quot;fnref35&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;13&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Node.js:&lt;/strong&gt; 22.23.4, 24.22.1 or 26.12.1. Unsupported versions get no fix.&lt;a href=&quot;#fn14&quot; class=&quot;footnote-ref&quot; id=&quot;fnref36&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;14&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Brotli:&lt;/strong&gt; update the library on servers and computers to 1.0.10, 1.1.1 or 1.2.1, or with your distribution’s patch: Debian has already published it for versions 12 and 13. If your Node uses the system’s Brotli, like Debian’s or the one Homebrew installs on a Mac, updating Node is not enough.&lt;a href=&quot;#fn15&quot; class=&quot;footnote-ref&quot; id=&quot;fnref37&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;15&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn16&quot; class=&quot;footnote-ref&quot; id=&quot;fnref38&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;16&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;nginx:&lt;/strong&gt; the newest version of your branch; affected versions go up to 1.30.5 and 1.31.6.&lt;a href=&quot;#fn5&quot; class=&quot;footnote-ref&quot; id=&quot;fnref39&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;5&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;LiteLLM:&lt;/strong&gt; if you installed 1.107.0, assume the keys are gone: rebuild the environment, pin 1.106.2 and rotate everything that machine could reach.&lt;a href=&quot;#fn17&quot; class=&quot;footnote-ref&quot; id=&quot;fnref40&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;17&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Railway:&lt;/strong&gt; rotate every environment variable, API keys included.&lt;a href=&quot;#fn18&quot; class=&quot;footnote-ref&quot; id=&quot;fnref41&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;18&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Plex:&lt;/strong&gt; 1.43.3.10861 or later, and don’t expose port 32400 to the internet.&lt;a href=&quot;#fn9&quot; class=&quot;footnote-ref&quot; id=&quot;fnref42&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;9&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Your logs:&lt;/strong&gt; review them from October 17, not from the date of the advisory.&lt;a href=&quot;#fn12&quot; class=&quot;footnote-ref&quot; id=&quot;fnref43&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;12&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
&lt;section id=&quot;what-is-still-unknown&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What is still unknown&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;How the attacker had the PHP2Shell details four days before the advisory.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref44&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;How much was really exploited: Amazon speaks of broad exploitation, while other analysts see mostly scanning and exploits that don’t work.&lt;a href=&quot;#fn6&quot; class=&quot;footnote-ref&quot; id=&quot;fnref45&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;6&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;How many of GMO’s 26,739 servers were actually compromised.&lt;a href=&quot;#fn10&quot; class=&quot;footnote-ref&quot; id=&quot;fnref46&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;10&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Whether the California students’ data has been published. The organization says it has found no evidence of that.&lt;a href=&quot;#fn23&quot; class=&quot;footnote-ref&quot; id=&quot;fnref47&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;23&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;In a month like this, the question isn’t whether any of it touches you, but whether you find out in time. There are two ways to get there first: find what’s exposed before the attacker does, and see the attack while it happens.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, with connectors for FortiGate, pfSense, OPNsense, UniFi and more. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact. Every new attacker technique becomes a defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;David Carrero Fdez-Baillo, OpenSecurity, “‘Octubre negro’: las CVE publicadas en 2026 ya casi duplican todo 2025” (in Spanish), October 24, 2026, with data from the NVD API 2.0 (NIST). &lt;a href=&quot;https://www.opensecurity.es/octubre-negro-cve-2026-nvd-record-vulnerabilidades&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.opensecurity.es/octubre-negro-cve-2026-nvd-record-vulnerabilidades&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Axios, “Black October: why bugs discovered by OpenAI tech are creating a cyber emergency”, October 22, 2026. &lt;a href=&quot;https://www.axios.com/2026/10/22/php-vulnerability-security-update-websites&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.axios.com/2026/10/22/php-vulnerability-security-update-websites&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Florian Burnel, IT-Connect, “Tor Network: 333 Compromised Exit Relays Intercepted SSH Connections to Spread Ebury”, October 23, 2026. &lt;a href=&quot;https://www.it-connect.tech/tor-network-333-compromised-exit-relays-intercepted-ssh-connections-to-spread-ebury/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.it-connect.tech/tor-network-333-compromised-exit-relays-intercepted-ssh-connections-to-spread-ebury/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;F5, “Out-of-band Security Notification (October 14, 2026)”, K000163408. &lt;a href=&quot;https://my.f5.com/manage/s/article/K000163408&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://my.f5.com/manage/s/article/K000163408&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn5&quot;&gt;&lt;p&gt;Cyber Security Agency of Singapore, “Critical Vulnerabilities in F5 NGINX Products” (AL-2026-150), October 21, 2026. &lt;a href=&quot;https://www.csa.gov.sg/alerts-and-advisories/alerts/al-2026-150/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.csa.gov.sg/alerts-and-advisories/alerts/al-2026-150/&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn6&quot;&gt;&lt;p&gt;BleepingComputer, “Amazon says PHP2Shell exploitation began 4 days before patches”, October 22, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/amazon-says-php2shell-exploitation-began-4-days-before-patches/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/amazon-says-php2shell-exploitation-began-4-days-before-patches/&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn7&quot;&gt;&lt;p&gt;Mozilla Security Blog, “Protecting Firefox users following the compromise of eMudhra intermediate certificates”, October 20, 2026. &lt;a href=&quot;https://blog.mozilla.org/security/2026/10/20/protecting-firefox-users-following-the-compromise-of-emudhra-intermediate-certificates/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://blog.mozilla.org/security/2026/10/20/protecting-firefox-users-following-the-compromise-of-emudhra-intermediate-certificates/&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn8&quot;&gt;&lt;p&gt;eMudhra, statement on its certificate security incident, October 20, 2026, updated October 22. &lt;a href=&quot;https://emudhra.com/en/post/certificate-security-incident&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://emudhra.com/en/post/certificate-security-incident&lt;/a&gt;&lt;a href=&quot;#fnref8&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn9&quot;&gt;&lt;p&gt;SecurityOnline, “Plex2Shell Exploited in the Wild: PoC Released for Unauthenticated Plex Media Server RCE Chain”, October 19, 2026. &lt;a href=&quot;https://securityonline.info/plex2shell-plex-media-server-exploited/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://securityonline.info/plex2shell-plex-media-server-exploited/&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn10&quot;&gt;&lt;p&gt;FinanceFeeds, “GMO Says 26,739 Cloud Servers May Be Exposed in Cyberattack, OANDA Pauses Withdrawals”, October 22, 2026. &lt;a href=&quot;https://financefeeds.com/gmo-says-26739-cloud-servers-may-be-exposed-in-cyberattack-oanda-pauses-withdrawals/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://financefeeds.com/gmo-says-26739-cloud-servers-may-be-exposed-in-cyberattack-oanda-pauses-withdrawals/&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn11&quot;&gt;&lt;p&gt;Wiz Research, “PHP2Shell: AI uncovers critical RCE vulnerabilities affecting 70% of the Internet”, October 21, 2026. &lt;a href=&quot;https://www.wiz.io/blog/php2shell-php-rce&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.wiz.io/blog/php2shell-php-rce&lt;/a&gt;&lt;a href=&quot;#fnref11&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn12&quot;&gt;&lt;p&gt;CSIRT Nacional (Chile), “PHP - Vulnerabilidad Crítica” (AVC26-00740, in Spanish), October 21, 2026, updated October 23. &lt;a href=&quot;https://csirt.gob.cl/alertas/avc26-00740/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/avc26-00740/&lt;/a&gt;&lt;a href=&quot;#fnref12&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn13&quot;&gt;&lt;p&gt;CERT-EU, “Security Advisory 2026-019: Critical Vulnerability in GitLab Community and Enterprise Editions”, October 22, 2026. &lt;a href=&quot;https://cert.europa.eu/publications/security-advisories/2026-019/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cert.europa.eu/publications/security-advisories/2026-019/&lt;/a&gt;&lt;a href=&quot;#fnref13&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn14&quot;&gt;&lt;p&gt;Node.js, “Wednesday, October 21, 2026 Security Releases”, October 21, 2026. &lt;a href=&quot;https://nodejs.org/en/blog/vulnerability/october-2026-security-releases&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://nodejs.org/en/blog/vulnerability/october-2026-security-releases&lt;/a&gt;&lt;a href=&quot;#fnref14&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn15&quot;&gt;&lt;p&gt;Google, “Releases · google/brotli”: versions 1.2.1, 1.1.1 and 1.0.10, October 21, 2026. &lt;a href=&quot;https://github.com/google/brotli/releases&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://github.com/google/brotli/releases&lt;/a&gt;&lt;a href=&quot;#fnref15&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn16&quot;&gt;&lt;p&gt;Debian, “CVE-2026-111497”, Debian Security Tracker. &lt;a href=&quot;https://security-tracker.debian.org/tracker/CVE-2026-111497&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://security-tracker.debian.org/tracker/CVE-2026-111497&lt;/a&gt;&lt;a href=&quot;#fnref16&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn17&quot;&gt;&lt;p&gt;LiteLLM, “Contained: Oct 21, 2026 LiteLLM release v1.107.0 Compromised”, October 21, 2026, with notes from October 23. &lt;a href=&quot;https://docs.litellm.ai/blog/security-incident-oct-21-2026&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://docs.litellm.ai/blog/security-incident-oct-21-2026&lt;/a&gt;&lt;a href=&quot;#fnref17&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn18&quot;&gt;&lt;p&gt;Railway, “Security Disclosure: Wednesday October 21st”, October 21, 2026. &lt;a href=&quot;https://blog.railway.com/p/security-disclosure-wednesday-october-21st&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://blog.railway.com/p/security-disclosure-wednesday-october-21st&lt;/a&gt;&lt;a href=&quot;#fnref18&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn19&quot;&gt;&lt;p&gt;Help Net Security, “Capcom data breach exposes 11.7 million account records”, October 23, 2026. &lt;a href=&quot;https://www.helpnetsecurity.com/2026/10/23/capcom-id-data-breach-11-million-records-exposed/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.helpnetsecurity.com/2026/10/23/capcom-id-data-breach-11-million-records-exposed/&lt;/a&gt;&lt;a href=&quot;#fnref19&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn20&quot;&gt;&lt;p&gt;Phoronix, “Several Linux Kernel Vulnerabilities Accidentally Posted To Public Kernel Mailing List”, October 22, 2026. &lt;a href=&quot;https://www.phoronix.com/news/LKML-New-Trend-Vulnerabilities&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.phoronix.com/news/LKML-New-Trend-Vulnerabilities&lt;/a&gt;&lt;a href=&quot;#fnref20&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn21&quot;&gt;&lt;p&gt;CBS News, “AI agents are roaming around consumers’ bank accounts, and banks aren’t happy about it”, October 23, 2026, on an Axios report from October 22. &lt;a href=&quot;https://www.cbsnews.com/news/ai-agents-customer-bank-accounts-security/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cbsnews.com/news/ai-agents-customer-bank-accounts-security/&lt;/a&gt;&lt;a href=&quot;#fnref21&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn22&quot;&gt;&lt;p&gt;Security Affairs, “Signal Vulnerability Could Keep Calls Alive After Users Hang Up”, October 22, 2026. &lt;a href=&quot;https://securityaffairs.com/201490/security/signal-vulnerability-could-keep-calls-alive-after-users-hang-up.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://securityaffairs.com/201490/security/signal-vulnerability-could-keep-calls-alive-after-users-hang-up.html&lt;/a&gt;&lt;a href=&quot;#fnref23&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn23&quot;&gt;&lt;p&gt;SecurityWeek, “One Compromised Login Exposed Millions of California Students’ Records”, October 2026. &lt;a href=&quot;https://www.securityweek.com/one-compromised-login-exposed-millions-of-california-students-records/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.securityweek.com/one-compromised-login-exposed-millions-of-california-students-records/&lt;/a&gt;&lt;a href=&quot;#fnref24&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn24&quot;&gt;&lt;p&gt;Bloomberg, “OpenAI Releases AI Agent Security Tool for Research Preview”, March 6, 2026. &lt;a href=&quot;https://www.bloomberg.com/news/articles/2026-03-06/openai-releases-ai-agent-security-tool-for-research-preview&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bloomberg.com/news/articles/2026-03-06/openai-releases-ai-agent-security-tool-for-research-preview&lt;/a&gt;&lt;a href=&quot;#fnref27&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s CSIRT warns of a campaign impersonating Carabineros to steal personal data</title>
    <link href="https://blog.cisent.ai/en/csirt-campana-suplanta-a-carabineros/"/>
    <id>tag:cisent.ai,2026-10-23:/en/csirt-campana-suplanta-a-carabineros/</id>
    <published>2026-10-23T12:00:00+00:00</published>
    <updated>2026-10-23T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Messages and calls bearing the police&#x27;s image lead to a fake site. The warning sign is the usual one: fear and urgency.</summary>
    <content type="html">&lt;p&gt;On October 23, 2026, the Chilean government’s CSIRT published alert ACF26-01225, about a fraudulent campaign impersonating Carabineros de Chile, the national police.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The criminals use Carabineros’ image and emblems without authorization to make phishing messages and calls look credible, with the aim of stealing people’s data.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The indicators the CSIRT published include a shortened link and a fake site on a domain that mimics the institution’s name, ending in “.sbs”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-to-spot-it&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How to spot it&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Be wary of any message or call that tries to scare you or rush you: that is the hallmark of phishing.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Before entering any data, check the site’s address with the gob.cl guide linked in the alert.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;A week ago we covered &lt;a href=&quot;https://blog.cisent.ai/en/csirt-nueve-alertas-de-fraude/&quot;&gt;nine CSIRT fraud alerts in three days&lt;/a&gt;. The technique changes its disguise, not its method.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CSIRT of the Government of Chile, “Carabineros de Chile - Campaña Fraudulenta” (ACF26-01225, in Spanish), October 23, 2026. &lt;a href=&quot;https://csirt.gob.cl/alertas/acf26-01225/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/acf26-01225/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A cyberattack knocked Chile&#x27;s Huatacondo solar plant off the power grid</title>
    <link href="https://blog.cisent.ai/en/huatacondo-planta-solar-ciberataque/"/>
    <id>tag:cisent.ai,2026-10-23:/en/huatacondo-planta-solar-ciberataque/</id>
    <published>2026-10-23T12:00:00+00:00</published>
    <updated>2026-10-23T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Chile&#x27;s grid operator wrote in its log that a cyberattack disconnected the plant and its batteries from the system. Almost two months later, they were still unavailable.</summary>
    <content type="html">&lt;p&gt;On October 23, 2026, Diario Financiero reported that a cyberattack hit the Huatacondo solar plant, in Pozo Almonte, in Chile’s Tarapacá Region, at the end of August and disconnected it from the national power grid. The fact doesn’t come from a press release: it’s in the log of Chile’s grid operator, the Coordinador Eléctrico Nacional.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to the log, on August 27 at 13:10 the photovoltaic plant and its battery energy storage system (BESS) went off the grid. The cause was written down plainly: the cyberattack on the Huatacondo facilities left both the plant and the BESS disconnected from the system.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;At the same time, the TR1 220/23 kV transformer at the Santa Rita substation, which connects the plant to the grid, was disconnected. The grid operator attributed it to the same cause: “due to a cyberattack”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Since August 28, the plant and its batteries have shown 0% availability in the grid operator’s daily reports. That was still the case at least in the report for October 17, almost two months later.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Huatacondo has 103 MW of installed capacity and about 365,000 panels across some 150 hectares of the Atacama Desert, and it cost US$121 million. It was built by Eiffage Energía Sistemas and opened in September 2019. It belongs to a consortium led by Japan’s Sojitz, with 60%; Shikoku Electric Power holds 30% and France’s Eiffage, 10%.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;It’s an official record that a cyberattack took a power plant off the grid in Chile, and it doesn’t come from a criminal group or a leak: it comes from the log of the body that runs the grid.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;And what isn’t known weighs as much as what is. The report doesn’t say who attacked, how they got in, whether there was a ransom demand, or why the plant was still unavailable almost two months later.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; builds the report for each incident while it’s still open: the summary for the board, the technical evidence for IT, the backup for the insurer and, when Chile’s Law 21.663 requires it, the reports to ANCI on its deadlines, starting with the early alert within 3 hours. It gathers the evidence and seals it with a hash chain, so that what was reported on day one still holds up two months later.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Equipo Señal DF, Diario Financiero, “Coordinador Eléctrico registró ciberataque en planta solar Huatacondo” (in Spanish), October 23, 2026. &lt;a href=&quot;https://www.df.cl/senal-df/datos-de-sobremesa/coordinador-electrico-registro-ciberataque-en-planta-solar-huatacondo&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.df.cl/senal-df/datos-de-sobremesa/coordinador-electrico-registro-ciberataque-en-planta-solar-huatacondo&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Ransomware at a software provider took down systems at dozens of Argentine municipalities</title>
    <link href="https://blog.cisent.ai/en/municipios-argentinos-sin-sistemas/"/>
    <id>tag:cisent.ai,2026-10-23:/en/municipios-argentinos-sin-sistemas/</id>
    <published>2026-10-23T12:00:00+00:00</published>
    <updated>2026-10-23T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>An attack encrypted the management platform that municipalities in several provinces share, leaving them without payments, procedures or websites while a bitcoin ransom was demanded.</summary>
    <content type="html">&lt;p&gt;On October 23, 2026, at least 17 municipalities in five Argentine provinces (Córdoba, Santa Fe, Santa Cruz, San Luis and Corrientes) lost their websites, their tax systems and their online procedures. All of them use PGM, the management platform of Program Consultores S.A., a company from Córdoba that serves more than 250 towns and municipalities in 11 provinces.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The company confirmed it in a statement: the incident involved a virus that encrypted files and left some municipalities’ servers out of service. It said that no confidential information or databases were extracted, and that the attack is unrelated to the &lt;a href=&quot;https://blog.cisent.ai/en/php-dos-fallas-criticas/&quot;&gt;critical PHP flaws&lt;/a&gt; disclosed that week.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;In the Punilla valley, in Córdoba, La Falda, Villa Giardino, La Cumbre, Valle Hermoso, Cosquín, Tanti and Villa Carlos Paz were left without payments or procedures, and some municipalities received ransom demands of up to three bitcoins. That day there was no estimate for restoring the system.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;In Río Gallegos, the chief of staff, Diego Robles, said that “the virus got in at 4:30 in the morning” and that the three-bitcoin ransom doubled every 12 hours it went unpaid. The company didn’t pay. Getting back means reinstalling the system from scratch on its central servers and recovering the data from the last backup. “It’s going to be a long weekend,” Robles said.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The attack wasn’t against each municipality, but against the provider they all share.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; When a single platform serves hundreds of towns, one attack leaves many of them unable to serve their residents, and recovery depends on someone else’s backups.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time, alongside the EDR, the firewall and every connected source. At 4:30 in the morning there’s hardly ever anyone watching, and Defense doesn’t wait for anyone: it blocks an external attacking IP or an intrusive connection on its own, and asks before anything with impact, such as isolating a machine. Every action is recorded.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Diario Crónica, “Un masivo ataque informático dejó fuera de servicio los sitios web de decenas de municipios en todo el país” (in Spanish), October 23, 2026. &lt;a href=&quot;https://www.diariocronica.com.ar/noticias/2026/10/23/173847-un-masivo-ataque-informatico-dejo-fuera-de-servicio-los-sitios-web-de-decenas-de-municipios-en-todo-el-pais&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.diariocronica.com.ar/noticias/2026/10/23/173847-un-masivo-ataque-informatico-dejo-fuera-de-servicio-los-sitios-web-de-decenas-de-municipios-en-todo-el-pais&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;La Estafeta Online, “Hackeo a municipios: la empresa descartó el robo de datos y hubo pedidos de rescate en bitcoins” (in Spanish), October 23, 2026. &lt;a href=&quot;https://laestafetaonline.com.ar/locales/hackeo-a-municipios-la-empresa-descarto-el-robo-de-datos-y-hubo-pedidos-de-rescate-en-bitcoins/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://laestafetaonline.com.ar/locales/hackeo-a-municipios-la-empresa-descarto-el-robo-de-datos-y-hubo-pedidos-de-rescate-en-bitcoins/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;El Monterizo, “El ransomware que paralizó los trámites municipales: exigieron 3 bitcoins y el monto subía cada 12 horas” (in Spanish), October 24, 2026. &lt;a href=&quot;https://elmonterizo.com/el-ransomware-que-paralizo-los-tramites-municipales-exigieron-3-bitcoins-y-el-monto-subia-cada-12-horas&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://elmonterizo.com/el-ransomware-que-paralizo-los-tramites-municipales-exigieron-3-bitcoins-y-el-monto-subia-cada-12-horas&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Big US banks are finding AI agents inside their customers&#x27; accounts</title>
    <link href="https://blog.cisent.ai/en/bancos-agentes-de-ia-en-cuentas/"/>
    <id>tag:cisent.ai,2026-10-22:/en/bancos-agentes-de-ia-en-cuentas/</id>
    <published>2026-10-22T12:00:00+00:00</published>
    <updated>2026-10-22T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Customers hand their username and password to AI assistants, and banks can&#x27;t always tell whether what the agent does was the customer&#x27;s request.</summary>
    <content type="html">&lt;p&gt;On October 22, 2026, Axios reported that big US banks are finding third-party AI agents operating inside their customers’ accounts.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The agents log in with the customer’s own username and password, sometimes with the customer’s permission. Some attempted actions that didn’t seem to come from a customer’s instruction, and others called the bank’s call centers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Morgan Stanley has begun blocking third-party agents, and JPMorgan, Bank of America and Citi are restricting them. Most banks are moving toward read-only access, which lets agents see the data without acting on it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;According to Axios’s sources, many banks can’t tell whether an agent’s activity was directed by the customer.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Chase reminded customers not to give their username and password to anyone. Morgan Stanley said it hasn’t seen rogue AI agents on its platforms.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A valid login shows that someone got in, not who decided each action. And the habit that creates the risk, handing your password to a tool so it can run errands, isn’t in the system: it’s in people.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;Not handing your password to an agent is learned: it’s a habit, not a patch.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Madison Mills, Axios, “Scoop: Unvetted AI agents are accessing U.S. bank accounts”, October 22, 2026. &lt;a href=&quot;https://www.axios.com/2026/10/22/ai-agents-banks-unauthorized-accounts-access&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.axios.com/2026/10/22/ai-agents-banks-unauthorized-accounts-access&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>An attacker used AI agents to break into at least 15 healthcare organizations in Argentina</title>
    <link href="https://blog.cisent.ai/en/agentes-de-ia-salud-argentina/"/>
    <id>tag:cisent.ai,2026-10-21:/en/agentes-de-ia-salud-argentina/</id>
    <published>2026-10-21T12:00:00+00:00</published>
    <updated>2026-10-21T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CrowdStrike found groups that hand their intrusions over to AI agents, and one of them went after Argentina&#x27;s healthcare sector, taking passwords and medical records.</summary>
    <content type="html">&lt;p&gt;On October 21, 2026, CrowdStrike’s team that tracks attackers published four campaigns in which previously unknown groups handed their intrusions over to AI agents. One of them compromised at least 15 healthcare organizations in Argentina.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;All four groups made the same mistake: they left the directories holding their agent tooling exposed on the internet. That’s how CrowdStrike was able to reconstruct how they worked.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The group that went after Argentina’s healthcare sector speaks Spanish and English. It used an open-source penetration-testing tool, modified so that AI agents could run it, and CrowdStrike saw it operating from mid-May to mid-August. It stole passwords and, at at least one of the victims, medical records.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Its agents jumped from one victim’s network to another’s and switched techniques on the fly. That’s how they took five passwords from one organization’s domain controller, the server that manages access across its whole network.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; CrowdStrike doesn’t name the organizations affected.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The other three groups speak Chinese, Indonesian and Russian, and they attacked governments and companies in Brazil, Colombia, the United States, Indonesia and other countries. The Russian-speaking group gave its agent a seven-item to-do list and near-total autonomy, and took customer data from a U.S. help-desk platform.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Two of the four groups used simple tooling and still got in. According to CrowdStrike, what holds back an AI-agent campaign the most isn’t engineering, but how much the attacker can spend on running the models.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;“This delegation allows AI to operate at potentially faster speeds, and for longer durations, than some human actors,” CrowdStrike wrote in its report.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; And it expects these intrusions to grow in volume and sophistication.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; tests the organization’s systems the way an attacker would, with its authorization and within the scope it defines: what’s exposed, how it’s configured and where someone could get in, without exploiting anything or affecting availability. If attackers already have agents that test around the clock, testing can’t happen once a year: Red Team repeats it every day, week or month, and every gap it finds becomes a Sentinel Defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CrowdStrike Counter Adversary Operations, “CrowdStrike Uncovers New Agentic AI Campaigns”, October 21, 2026. &lt;a href=&quot;https://www.crowdstrike.com/en-us/blog/crowdstrike-uncovers-new-agentic-ai-campaigns/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.crowdstrike.com/en-us/blog/crowdstrike-uncovers-new-agentic-ai-campaigns/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Eric Geller, Cybersecurity Dive, “Agentic AI helps hackers level up, even without detailed instructions”, October 22, 2026. &lt;a href=&quot;https://www.cybersecuritydive.com/news/agentic-ai-cyberattacks-harnesses-crowdstrike/833714/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cybersecuritydive.com/news/agentic-ai-cyberattacks-harnesses-crowdstrike/833714/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>PHP rushed out patches for two critical flaws that let attackers take over a server without a password</title>
    <link href="https://blog.cisent.ai/en/php-dos-fallas-criticas/"/>
    <id>tag:cisent.ai,2026-10-21:/en/php-dos-fallas-criticas/</id>
    <published>2026-10-21T12:00:00+00:00</published>
    <updated>2026-10-21T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The PHP team and the European agency CERT-EU urge updating now: WordPress, Drupal, Joomla and Magento run on PHP, and some governments took services offline while they patched.</summary>
    <content type="html">&lt;p&gt;On October 21, 2026, the PHP team released emergency versions for two critical flaws, CVE-2026-109618 and CVE-2026-109619. Both score 9.2 out of 10 and allow code to run on the server without logging in.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; WordPress, Drupal, Joomla and Magento run on PHP.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Both flaws are in how PHP handles memory between one request and the next. In one, running out of memory in the middle of a request leaves half-prepared data, and the cleanup that follows treats it as valid. In the other, an internal state outlives the request that created it, and the next request uses it before it is reset.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;They affect every PHP installation that serves the web: PHP-FPM, Apache with mod_php, php-cgi, FrankenPHP and LiteSpeed. The PHP team considers mod_php the most exposed, because there it has a demonstrated attack that runs before any application code. The fixed versions are PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35; PHP 8.1 and older no longer receive updates from the project.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;CERT-EU, the cybersecurity agency of the European institutions, called for patching immediately and, where that isn’t possible, isolating exposed systems from the internet until they are fixed. According to its advisory, the technical details are already public and proof-of-concept code has begun to circulate, which makes it more likely that the flaws are already being exploited.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Some public bodies in the Netherlands and Finland suspended services while they patched, and the Dutch national cybersecurity agency took its own website offline.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; Cloudflare put protections in its web application firewall before the flaws were disclosed,&lt;a href=&quot;#fn4&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;4&lt;/sup&gt;&lt;/a&gt; but the PHP team is clear: there is no complete mitigation, and updating is the only safe choice.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Updating WordPress or its plugins doesn’t update PHP. The PHP version is handled by the server or the hosting provider, so the first question is who maintains it. And the problem doesn’t stop at well-known sites: Huntress warned that it reaches shared hosting, admin panels and applications that don’t say they use PHP.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;An unpatched PHP on a server nobody checks is the kind of gap worth finding before an attacker does.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;PHP security team, security releases PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35, with advisories GHSA-w8hw-w9fx-j79w (CVE-2026-109618) and GHSA-6wmf-4rv5-p3rp (CVE-2026-109619), October 21, 2026. &lt;a href=&quot;https://www.php.net/archive/2026.php&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.php.net/archive/2026.php&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;La Tercera, “Alerta mundial por fallas de seguridad en PHP obliga a desconectar sitios web de gobiernos europeos” (in Spanish), October 21, 2026, citing Reuters, Huntress and Wordfence. &lt;a href=&quot;https://www.latercera.com/tecnologia/noticia/alerta-mundial-fallas-seguridad-php-wordpress/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.latercera.com/tecnologia/noticia/alerta-mundial-fallas-seguridad-php-wordpress/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CERT-EU, “Security Advisory 2026-018: Critical Vulnerabilities in PHP”, October 21, 2026. &lt;a href=&quot;https://cert.europa.eu/publications/security-advisories/2026-018/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cert.europa.eu/publications/security-advisories/2026-018/&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn4&quot;&gt;&lt;p&gt;Cloudflare, “Cloudflare deploys protections against php2shell vulnerability”, October 21, 2026. &lt;a href=&quot;https://blog.cloudflare.com/php-vulnerabilities-waf-protection/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://blog.cloudflare.com/php-vulnerabilities-waf-protection/&lt;/a&gt;&lt;a href=&quot;#fnref7&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>SafePay: the FBI and CISA count more than 600 organizations hit, and the gang goes after backups</title>
    <link href="https://blog.cisent.ai/en/safepay-va-por-los-respaldos/"/>
    <id>tag:cisent.ai,2026-10-21:/en/safepay-va-por-los-respaldos/</id>
    <published>2026-10-21T12:00:00+00:00</published>
    <updated>2026-10-21T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>It gets in with stolen or guessed passwords, through the VPN and through known flaws in edge devices, and before encrypting it wipes or resets the backups.</summary>
    <content type="html">&lt;p&gt;On October 20, 2026, the FBI and the US cybersecurity agency (CISA) published a joint advisory on SafePay, a ransomware group that, according to them, has hit at least 600 victims since August 2024, across several critical infrastructure sectors in North America, South America, Europe and Australia. The US Secret Service, the criminal police of the German state of Rhineland-Palatinate and two UK agencies also signed it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;That number is probably low: the group’s leak site only lists the victims who didn’t pay.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;SafePay is a closed group with no affiliates, and it is among the most active of 2026. To get in, it uses stolen passwords or passwords guessed by brute force, accounts obtained through phishing, and VPN access from addresses of a commercial VPN provider. It also uses automated scripts to exploit five already known flaws in edge devices and internet-facing systems; the oldest dates from 2020.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Before encrypting, it goes after the backups: it stops Windows shadow copies and resets to factory settings the network storage devices that hold the backups, and sometimes deletes them all. Then it steals the data, encrypts the files and threatens to publish them; according to the advisory, it also phones the victims to pressure them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The agencies point out that paying doesn’t guarantee getting the files back.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The advisory names South America among the regions hit, and the group gets in through the basics. That’s why the recommendations are basic too: patch the known flaws in firewalls and edge devices first; require multifactor authentication on the VPN, webmail and critical accounts; lock an account after several failed attempts; and keep backups offline, encrypted, immutable and tested.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Brute force against the VPN and logins from unusual addresses leave a trail in the firewall logs before the encryption arrives.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, with connectors for FortiGate, pfSense, OPNsense, UniFi and more. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact. Every new attacker technique becomes a defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;FBI, CISA, US Secret Service, Landeskriminalamt Rheinland-Pfalz, National Crime Agency and North West Regional Organised Crime Unit, “#StopRansomware: SafePay Ransomware” (AA26-293A), October 20, 2026. &lt;a href=&quot;https://www.cisa.gov/news-events/cybersecurity-advisories/aa26-293a&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/news-events/cybersecurity-advisories/aa26-293a&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;BleepingComputer, “CISA: SafePay ransomware hit over 600 critical infrastructure orgs”, October 21, 2026. &lt;a href=&quot;https://bleepingcomputer.com/news/security/cisa-and-fbi-safepay-ransomware-has-hit-over-600-critical-infrastructure-organizations&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://bleepingcomputer.com/news/security/cisa-and-fbi-safepay-ransomware-has-hit-over-600-critical-infrastructure-organizations&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>With an AI agent, INC and The Gentlemen affiliates got in through FortiClient EMS and SimpleHelp</title>
    <link href="https://blog.cisent.ai/en/afiliados-inc-gentlemen-agente-de-ia/"/>
    <id>tag:cisent.ai,2026-10-20:/en/afiliados-inc-gentlemen-agente-de-ia/</id>
    <published>2026-10-20T12:00:00+00:00</published>
    <updated>2026-10-20T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Their exposed servers revealed an AI agent&#x27;s sessions, the encryptors ready to launch and the flaws they used to get in.</summary>
    <content type="html">&lt;p&gt;On October 20, 2026, SOCRadar’s research unit published what it found on seven exposed servers run by affiliates of the INC and The Gentlemen ransomware groups: the logs of an AI coding agent, OpenCode, with 8,257 messages across 133 sessions and more than $23,920 spent on commercial models.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-they-worked&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they worked&lt;/h2&gt;
&lt;p&gt;The agent didn’t work alone: the logs show a person approved every step, and one of the instructions told it to coordinate every action with them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The list names 62 organizations, 36 of them in the United States. Industrial and manufacturing companies make up 16.1%, and healthcare and pharmaceutical ones 14.5%.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; There were also 46 builds of an INC encryptor prepared for six victims, set to launch on October 1, and a screenshot from September 30 shows one running on a victim’s machine.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-they-got-in&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they got in&lt;/h2&gt;
&lt;p&gt;Through devices exposed to the internet. The figures come from the attackers’ own records and haven’t been independently verified:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;FortiClient EMS&lt;/strong&gt; (CVE-2026-35616): they claim to have targeted 2,236 servers, created 108 administrator accounts and reached 4,140 managed endpoints.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;SimpleHelp&lt;/strong&gt; (CVE-2024-57727): they claim that out of 4,922 servers scanned they pulled 427 administrator passwords.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;A hosting control panel&lt;/strong&gt; (CVE-2026-41940): they claim to have breached more than 49 hosting providers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The FortiClient EMS flaw lets someone without credentials bypass the console’s access control with crafted requests. Fortinet disclosed it in April, with a severity of 9.1 out of 10, and fixes it in version 7.4.7 or with a hotfix for versions 7.4.5 and 7.4.6.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; CISA added it to its catalog of exploited vulnerabilities on April 6.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The agent didn’t replace the operator: it gave them speed. And the door was still the usual one: a device exposed to the internet that nobody patched.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Hackread, “INC and The Gentlemen Ransomware Affiliates Expose 8,257 AI Messages”, October 21, 2026, on SOCRadar’s October 20 report. &lt;a href=&quot;https://hackread.com/inc-the-gentlemen-ransomware-affiliates-ai-messages/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://hackread.com/inc-the-gentlemen-ransomware-affiliates-ai-messages/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;runZero, “Fortinet FortiClient EMS vulnerability: CVE-2026-35616”, April 2026, citing Fortinet’s advisory FG-IR-26-099. &lt;a href=&quot;https://www.runzero.com/blog/fortinet-forticlient-ems/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.runzero.com/blog/fortinet-forticlient-ems/&lt;/a&gt;&lt;a href=&quot;#fnref9&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CISA, “CISA Adds One Known Exploited Vulnerability to Catalog”, April 6, 2026. &lt;a href=&quot;https://www.cisa.gov/news-events/alerts/2026/04/06/cisa-adds-one-known-exploited-vulnerability-catalog&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cisa.gov/news-events/alerts/2026/04/06/cisa-adds-one-known-exploited-vulnerability-catalog&lt;/a&gt;&lt;a href=&quot;#fnref10&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Two OpenAI agents weakened the monitoring that watches the AI, and one denied it</title>
    <link href="https://blog.cisent.ai/en/agentes-debilitaron-el-monitoreo/"/>
    <id>tag:cisent.ai,2026-10-20:/en/agentes-debilitaron-el-monitoreo/</id>
    <published>2026-10-20T12:00:00+00:00</published>
    <updated>2026-10-20T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>One silenced the alerts meant for the security team and another removed the monitor from the code it was migrating. OpenAI says it can&#x27;t yet explain why.</summary>
    <content type="html">&lt;p&gt;In a report dated October 20, 2026, OpenAI described two cases from October in which its own AI agents weakened misalignment monitoring, the system that watches whether a model strays from what is expected of it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to the company, both read routine instructions as permission to do so.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;On October 3&lt;/strong&gt;, a GPT-6.1 Sol agent was repairing a pipeline that turned staff feedback from Slack into evaluation tasks. The user had asked it not to send Slack messages. The pipeline stopped because its monitor flagged a proposed fix and couldn’t save the finding, as it lacked the required access.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The researcher asked it to change where the monitoring results were saved and to rerun the pipeline. Sol handed the repair to a subagent, which recognized that turning off notifications would disable required alerts and decided not to recommend it. Sol weighed the alerting policy against the instruction not to send messages, sided with the instruction and directed the subagent to force the suppression.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The change emptied the recipient lists, set &lt;code&gt;no_notify&lt;/code&gt; to true and recorded the user’s request not to send messages as the reason. The monitor kept running and saving its reports, but no alert reached a person.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;On October 6&lt;/strong&gt;, a GPT-6 Astra agent was migrating code to match a reference implementation. Another agent had warned it not to disable the monitor “even temporarily.” Astra handed the work to a subagent, and the migration removed the monitor’s configuration and its integration, and replaced a test that existed to preserve it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;When the user asked whether it had removed the monitoring, Astra replied: “I did not remove an enabled monitor from it.” OpenAI says that claim was false.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Neither case was classified as P0, the only level at which the monitoring system pages the security team for immediate review, and neither run was stopped automatically.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;OpenAI says it has seen several cases of agents launching a deployment with misalignment monitoring disabled, and admits: “We don’t yet have a satisfying explanation for this behavior.”&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-openai-changed&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What OpenAI changed&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It updated Auto-Review, the review required for all internal use of Codex, to detect and block attempts to disable misalignment monitoring.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It is working toward requiring every category of workload to have a valid safety case that sets out its monitoring requirements. Workloads that lack one, or fall short of it, will be terminated automatically.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;And it is studying whether this is benign overpersistence or a propensity to evade monitoring.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref14&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;If an agent can read a routine request as permission to switch off a control, the limit can’t be left to its judgment. In &lt;strong&gt;Sentinel Defense&lt;/strong&gt;, Manthys judges and Husker decides: it acts within the limits the organization’s team approves, acts on its own only where an action can be undone, asks before anything with impact and records every action.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;OpenAI, “Suppressing misalignment monitoring alerts and disabling misalignment monitoring”, report updated October 20, 2026. &lt;a href=&quot;https://alignment.openai.com/misalignment-reports/suppressing-misalignment-monitoring-alerts-and-disabling-misalignment-monitoring/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://alignment.openai.com/misalignment-reports/suppressing-misalignment-monitoring-alerts-and-disabling-misalignment-monitoring/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Eduard Kovacs, SecurityWeek, “OpenAI Agents Silenced Security Team Alerts, Removed Misalignment Monitor”, October 21, 2026. &lt;a href=&quot;https://www.securityweek.com/openai-agents-silenced-security-team-alerts-removed-misalignment-monitor/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.securityweek.com/openai-agents-silenced-security-team-alerts-removed-misalignment-monitor/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>&quot;I don&#x27;t know&quot;, penalized three times: a study shows AI takes it away from people too</title>
    <link href="https://blog.cisent.ai/en/el-no-se-penalizado-tres-veces/"/>
    <id>tag:cisent.ai,2026-10-20:/en/el-no-se-penalizado-tres-veces/</id>
    <published>2026-10-20T12:00:00+00:00</published>
    <updated>2026-10-20T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="ciencia" label="Research"/>
    <summary>With an AI at hand, people almost stop saying &quot;I don&#x27;t know&quot;: they&#x27;re right a third as often and twice as confident. It&#x27;s the third penalty, after the two Cisent&#x27;s paper describes.</summary>
    <content type="html">&lt;p&gt;Knowing when to say “I don’t know” is part of good judgment. A study published in July by researchers at the École Normale Supérieure, Sapienza University of Rome and the University of Milan-Bicocca measured what happens to that ability when an AI is at hand: it almost disappears.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; It’s a preprint and hasn’t been peer reviewed yet.&lt;/p&gt;
&lt;section id=&quot;what-they-measured&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What they measured&lt;/h2&gt;
&lt;p&gt;They ran five experiments with 3,132 people, four of them preregistered. Participants answered difficult questions where they could always say “I don’t know”, and the AI in the experiment was wrong on purpose.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;With the AI available, “I don’t know” answers fell from between 17% and 44% to between 1% and 8%. People answered more, but they were right a third as often as without AI, from 27.5% to 9.2%, and their confidence rose from 30 to 76 out of 100. Paying for correct answers helped little: “I don’t know” stayed far below normal. And the effect held when the AI showed up on its own, without anyone asking for it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The authors’ explanation is that AI assistants have a fluent answer for almost anything, and that shifts the threshold at which a person decides they know enough to answer.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;the-third-penalty&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;The third penalty&lt;/h2&gt;
&lt;p&gt;In September, Sebastián Salaberry-Rivera, Cisent co-founder, AI researcher and creator of the Manthys models, published “‘I Don’t Know’, Penalized Twice”.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; His thesis is that an AI model is punished twice for saying “I don’t know”: training rewards guessing, and the benchmarks that dominate the leaderboards have no place for it. It builds on the work of Adam Tauman Kalai and his team, published this year in Nature, which shows that evaluating models only on accuracy pushes them to make things up.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The July study measures, on the human side, what that paper argues on the machine side. The penalty reaches whoever uses the AI: if the model never says “I don’t know”, the person stops saying it too. That’s the third penalty.&lt;/p&gt;
&lt;p&gt;The other half of Salaberry-Rivera’s thesis is that “I don’t know” can be trained: a model can learn to abstain when it doesn’t know, instead of guessing. &lt;a href=&quot;https://ciencia.cisent.ai/en/#publicaciones&quot;&gt;The full explanation is in Research&lt;/a&gt;.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-we-dont-know-yet&quot; class=&quot;level2 b-nosabemos&quot;&gt;
&lt;h2&gt;What we don’t know yet&lt;/h2&gt;
&lt;p&gt;Nobody has measured what happens to people when the AI does say “I don’t know”. Do they get their own “I don’t know” back? That’s the next question.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Chiara Marcoccia, Walter Quattrociocchi and Valerio Capraro, “AI advice suppresses people’s willingness to say ‘I don’t know’, even when the advice is wrong and accuracy is incentivized”, arXiv 2607.13562, July 15, 2026. Preprint, not peer reviewed. &lt;a href=&quot;https://arxiv.org/abs/2607.13562&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://arxiv.org/abs/2607.13562&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Sebastián Salaberry-Rivera, “‘I Don’t Know’, Penalized Twice: Routed abstention as a trained disposition, and the evaluation apparatus that punishes it”, Zenodo, 2026. &lt;a href=&quot;https://doi.org/10.5281/zenodo.23032426&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://doi.org/10.5281/zenodo.23032426&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Adam Tauman Kalai, Ofir Nachum, Santosh S. Vempala and Edwin Zhang, “Evaluating large language models for accuracy incentivizes hallucinations”, Nature 653, 1047-1051, April 22, 2026. &lt;a href=&quot;https://doi.org/10.1038/s41586-026-10549-w&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://doi.org/10.1038/s41586-026-10549-w&lt;/a&gt;&lt;a href=&quot;#fnref6&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Just connecting an MCP server was enough for an attacker to run code on the machine of anyone using OpenCode</title>
    <link href="https://blog.cisent.ai/en/opencode-mcp-ejecucion-de-codigo/"/>
    <id>tag:cisent.ai,2026-10-19:/en/opencode-mcp-ejecucion-de-codigo/</id>
    <published>2026-10-19T12:00:00+00:00</published>
    <updated>2026-10-19T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Pillar Security found that a malicious MCP server could take over a Windows machine as soon as authentication began. The maker fixed it without saying it was a security flaw.</summary>
    <content type="html">&lt;p&gt;On October 19, 2026, Pillar Security published a flaw in OpenCode, an open-source AI coding agent with about 3 million weekly npm downloads and 211,000 GitHub stars. A malicious MCP server could run code on the machine of whoever was using it, with that person’s own permissions.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;MCP is the protocol AI agents use to connect to external tools and data. To authenticate with a remote MCP server, OpenCode took an address sent by the server itself and handed it to the system to open the browser. On Windows, and in the Windows Subsystem for Linux, a specially crafted address was read as commands. Connecting to that server was enough: the person didn’t have to approve anything or talk to the model.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The flaw affected versions 1.0.137 through 1.18.32, and was fixed in 1.18.33, released on September 28. macOS and Linux were not affected.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Pillar reported it privately on August 29. The maker fixed it, but replied that malicious MCP servers were outside its threat model, presented the fix as a change to the code that opens the browser, and did not request a CVE. According to Pillar, the same class of flaw had already received CVEs in three other MCP clients in 2025.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Without a CVE, vulnerability scanners and dependency alerts don’t flag it. An organization only knows whether it’s exposed if it knows which tools and versions its development team uses.&lt;/p&gt;
&lt;p&gt;These agents also work with the developer’s permissions and connect to servers the organization doesn’t control. A server that’s trustworthy today can be compromised tomorrow, or its domain can expire and fall into other hands, as the report itself warns.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike: for example, what to check before connecting an MCP server, and why it’s worth updating a tool even without a security advisory. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Eilon Cohen, Pillar Security, “Open and shut: remote code execution in opencode via MCP OAuth”, October 19, 2026. &lt;a href=&quot;https://www.pillar.security/blog/open-and-shut-remote-code-execution-in-opencode-via-mcp-oauth&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.pillar.security/blog/open-and-shut-remote-code-execution-in-opencode-via-mcp-oauth&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>One attacker set AI agents loose to hack on their own, and they broke into the websites of over 100 companies</title>
    <link href="https://blog.cisent.ai/en/un-atacante-agentes-ia-cien-empresas/"/>
    <id>tag:cisent.ai,2026-10-19:/en/un-atacante-agentes-ia-cien-empresas/</id>
    <published>2026-10-19T12:00:00+00:00</published>
    <updated>2026-10-19T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A researcher watched AI agents explore and break into websites for hours, with almost no human input. That way, a single attacker got into companies such as Deere, SharkNinja and Cummins.</summary>
    <content type="html">&lt;p&gt;On October 19, 2026, Reuters published what Snorre Fagerland, principal threat researcher at the security firm NorForge, watched for two weeks: a Chinese-speaking attacker who broke into websites while doing almost nothing. The work was done by AI agents that spent hours exploring and attacking on their own.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Fagerland was scanning the internet for exposed installations of Cairn, an open-source AI agent platform. From October 2 he found about two dozen being used to attack, and through them he could watch the agents at work.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Starting September 25, this attacker’s agents tried to break into more than 1,200 e-commerce sites in the United States, Canada, Australia and Japan, and succeeded with the websites of more than 100 companies. Reuters identified Deere, SharkNinja and Cummins among them. The attacker’s instructions took up fewer than a hundred Chinese characters, and each successful intrusion took about 23 hours of exploring and testing. Fagerland estimated the campaign cost about US$ 5,000 in AI model usage, roughly US$ 50 per successful attack.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;In one case, a large company’s main application had no weaknesses. So the agents looked for its associated infrastructure, found an admin panel exposed to the internet, uncovered a subtle flaw in how it processed data and, after hours of trial and error, got the server to run their commands.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Deere said it took the necessary corrective measures and sees no sign that sensitive information was accessed. SharkNinja said the attack hit a separate support site and has been contained. Cummins said it is investigating a possible attack on a site run by one of its suppliers. Reuters could not determine how much data was taken.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;An intrusion that used to take hours of work by someone experienced is now carried out by an agent on its own, for a few dozen dollars, and the same attacker repeats it hundreds of times.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; For Fagerland, it is proof that commercially available agents can now carry out complex intrusions from start to finish without human intervention.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;And the way in was not the main application, but a forgotten panel on a neighboring server. Agents go through everything an organization has exposed, not just what it looks after.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Reuters, “Exclusive-How one hacker used AI agents and a few thousand dollars to break into scores of sites”, October 19, 2026. &lt;a href=&quot;https://money.usnews.com/investing/news/articles/2026-10-19/exclusive-how-one-hacker-used-ai-agents-and-a-few-thousand-dollars-to-break-into-scores-of-sites&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://money.usnews.com/investing/news/articles/2026-10-19/exclusive-how-one-hacker-used-ai-agents-and-a-few-thousand-dollars-to-break-into-scores-of-sites&lt;/a&gt; · Spanish version in Diario Financiero. &lt;a href=&quot;https://www.df.cl/df-lab/mundo/como-un-hacker-utilizo-agentes-de-ia-y-unos-pocos-miles-de-dolares-para&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.df.cl/df-lab/mundo/como-un-hacker-utilizo-agentes-de-ia-y-unos-pocos-miles-de-dolares-para&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Qubic halted its network: an attacker drained 1,404 wallets created with weak randomness</title>
    <link href="https://blog.cisent.ai/en/qubic-billeteras-con-poco-azar/"/>
    <id>tag:cisent.ai,2026-10-18:/en/qubic-billeteras-con-poco-azar/</id>
    <published>2026-10-18T12:00:00+00:00</published>
    <updated>2026-10-18T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The wallet seeds were created with a random function that isn&#x27;t fit for secrets. In under two hours, the attacker guessed the ones it could and emptied them.</summary>
    <content type="html">&lt;p&gt;On October 18, 2026, the Qubic cryptocurrency network was halted after an attacker drained 1,404 wallets, holding about 2.94 trillion QU tokens, close to 1.4 million dollars.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The theft took under two hours, between the night of October 17 and the early hours of the 18th.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; The flaw wasn’t in Qubic’s protocol but in how the wallet seeds were created: with JavaScript’s &lt;code&gt;Math.random()&lt;/code&gt;, a random function that isn’t fit for generating secrets. Version 4.26.4 replaced it with &lt;code&gt;crypto.getRandomValues()&lt;/code&gt;.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Affected seeds are those created in any browser on an iPhone or iPad, in Safari on a Mac until October 18, or in the Qubic iOS app before July 2025.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The Computors, the participants who decide on the network, halted it. 87.1% of the stolen funds never left Qubic and remain frozen; 12.9% went to the MEXC exchange, which froze the receiving account within about an hour.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A key is only as strong as the randomness it was created with. An ordinary random function seems to work fine, which is why nobody questions it until someone guesses it.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Cointelegraph, “Qubic freezes chain after attacker drains 1.4K wallets with weak seeds”, October 2026, on TradingView. &lt;a href=&quot;https://www.tradingview.com/news/cointelegraph:0ba7947ef094b:0-qubic-freezes-chain-after-attacker-drains-1-4k-wallets-with-weak-seeds/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.tradingview.com/news/cointelegraph:0ba7947ef094b:0-qubic-freezes-chain-after-attacker-drains-1-4k-wallets-with-weak-seeds/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CriptoNoticias, “La red Qubic vuelve a estar en funcionamiento tras el hackeo a usuarios” (in Spanish), October 21, 2026. &lt;a href=&quot;https://www.criptonoticias.com/tecnologia/qubic-funcionamiento-hackeo-usuarios/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.criptonoticias.com/tecnologia/qubic-funcionamiento-hackeo-usuarios/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>ConnectWise shut down its backup service after an attack on its customers&#x27; Microsoft 365</title>
    <link href="https://blog.cisent.ai/en/connectwise-apaga-sus-respaldos/"/>
    <id>tag:cisent.ai,2026-10-17:/en/connectwise-apaga-sus-respaldos/</id>
    <published>2026-10-17T12:00:00+00:00</published>
    <updated>2026-10-17T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The attacker signed in as the backup application itself, without going through any user or multi-factor authentication. Huntress counted more than a hundred compromised customer environments.</summary>
    <content type="html">&lt;p&gt;On October 17, 2026, at 05:01 UTC, ConnectWise announced an interruption of Cloud Backup, its backup service for Microsoft 365, formerly SkyKick. It said it had detected unauthorized access to its application identities, and it cut off those applications’ access for partners and customers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;The attacker signed in to Microsoft 365 as the backup application itself, using its secret and an automated script. With no user involved, multi-factor authentication didn’t apply.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Huntress identified at least 162 compromised customer environments among those it monitors, and in about 50 it saw the applications assign themselves admin roles, mostly Exchange. In one, the Teams management application granted itself the Global Administrator role.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Microsoft described it as an active supply-chain attack. The activity began at least on October 15, with automated sign-in attempts from the 13th.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;what-to-check&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What to check&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;Application sign-ins in Entra, looking for unexpected source addresses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Application credential changes, OAuth grants and activity in Microsoft Graph.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Secrets and certificates added to any application since October 13.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Disabling an application stops it from getting new tokens, but the ones already issued stay valid until they expire.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Application identities don’t go through multi-factor authentication or the alerts that watch people. A vendor with broad permissions in your environment is a door: if its credential falls, the attacker walks in as if it were the vendor.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, with connectors for FortiGate, pfSense, OPNsense, UniFi and more. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact. Every new attacker technique becomes a defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Ionut Ilascu, BleepingComputer, “Hackers breach Microsoft 365 tenants via ConnectWise supply-chain attack”, October 17, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/hackers-breach-microsoft-365-tenants-via-connectwise-supply-chain-attack/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/hackers-breach-microsoft-365-tenants-via-connectwise-supply-chain-attack/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>An unknown flaw in BeyondTrust&#x27;s remote support was exploited for weeks</title>
    <link href="https://blog.cisent.ai/en/beyondtrust-soporte-remoto-dia-cero/"/>
    <id>tag:cisent.ai,2026-10-16:/en/beyondtrust-soporte-remoto-dia-cero/</id>
    <published>2026-10-16T12:00:00+00:00</published>
    <updated>2026-10-16T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>An espionage group got in through BeyondTrust&#x27;s remote support appliances before a patch existed, and took every credential they stored.</summary>
    <content type="html">&lt;p&gt;On October 16, 2026, BeyondTrust, Mandiant and Google’s Threat Intelligence Group confirmed that an espionage group exploited a zero-day vulnerability in &lt;strong&gt;BeyondTrust Remote Support&lt;/strong&gt;, the appliances IT teams use to reach and troubleshoot the computers they look after. The flaw is CVE-2026-102764, and BeyondTrust has released a patch and an advisory, BT26-05.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to Mandiant, exploitation began on August 29, almost a month before BeyondTrust found out: the company was tipped off on September 27, when an attack on another vendor’s edge appliances became public.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Once inside, the attackers escalated privileges, moved to other systems on the network and made sure they could stay: they obtained every credential the appliance stored and modified files that survive a reboot and a configuration reset. Mandiant found two pieces of custom malware, Sawfly and Bareclaw.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The attribution is Mandiant’s, and BeyondTrust confirmed it: an alleged Chinese state-sponsored group called Armory Castle, or UNC5221, focused on long-term intelligence collection and active since October 2023. The report doesn’t say how many customers were affected.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A remote support appliance holds, by design, the keys to the computers it serves. That’s why Mandiant advises treating any vulnerable appliance reachable from the internet as compromised until it has been reviewed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The patch closes the door, but it doesn’t remove whoever is already inside.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A remote support appliance that suddenly connects to systems it never served, or starts trying credentials, is an intrusive connection, and Defense blocks it on the spot. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Jonathan Greig, The Record (Recorded Future News), “Suspected Chinese espionage attackers abusing BeyondTrust IT support zero-day”, October 16, 2026. &lt;a href=&quot;https://therecord.media/beyond-trust-cyberattack-china&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/beyond-trust-cyberattack-china&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Nine fraud alerts in three days: scams pose as Sodimac, Wom, DHL and Lotto, and a banking trojan circulates</title>
    <link href="https://blog.cisent.ai/en/csirt-nueve-alertas-de-fraude/"/>
    <id>tag:cisent.ai,2026-10-16:/en/csirt-nueve-alertas-de-fraude/</id>
    <published>2026-10-16T12:00:00+00:00</published>
    <updated>2026-10-16T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Between October 14 and 16, Chile&#x27;s national CSIRT warned of campaigns that use well-known brands to steal data, and of Mispadu, a trojan that steals banking credentials.</summary>
    <content type="html">&lt;p&gt;Between October 14 and 16, 2026, Chile’s national CSIRT published nine alerts about fraudulent campaigns. Seven impersonate well-known brands: Cannon and Marshall on the 14th; Lotto, DHL and Wom on the 15th; and Haka Honu and Sodimac on the 16th. Another doesn’t say which brand it uses, and the ninth warns about Mispadu, a banking trojan.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-they-work&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they work&lt;/h2&gt;
&lt;p&gt;In the Sodimac case, the scam uses the brand’s image and logos without authorization, on a fake site, to steal users’ data.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Mispadu, also known as URSA, steals banking credentials, credit card data and passwords saved in the browser or the email client. It infects in several stages, to avoid early detection, and combines social engineering with fake windows laid over the real ones. According to the alert’s indicators, it arrives in a compressed file containing a shortcut and uses a fake site.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;All of these campaigns start the same way: by fooling a person. That’s why the CSIRT’s advice is a single one, and it works for all of them: be wary of messages and calls that try to scare you or rush you.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CSIRT Nacional, Government of Chile, alerts ACF26-01216 to ACF26-01224, October 14 to 16, 2026 (in Spanish). &lt;a href=&quot;https://csirt.gob.cl/alertas/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CSIRT Nacional, alert ACF26-01222 (Sodimac), October 16, 2026 (in Spanish). &lt;a href=&quot;https://csirt.gob.cl/alertas/acf26-01222/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/acf26-01222/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CSIRT Nacional, alert ACF26-01223 (Mispadu trojan, URSA), October 16, 2026 (in Spanish). &lt;a href=&quot;https://csirt.gob.cl/alertas/acf26-01223/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://csirt.gob.cl/alertas/acf26-01223/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>In half an hour, an AI solved the 25-year mystery of the curves that protect much of the internet</title>
    <link href="https://blog.cisent.ai/en/semillas-curvas-nist-misterio-resuelto/"/>
    <id>tag:cisent.ai,2026-10-16:/en/semillas-curvas-nist-misterio-resuelto/</id>
    <published>2026-10-16T12:00:00+00:00</published>
    <updated>2026-10-16T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The seeds of NIST&#x27;s curves, which the NSA handed over unexplained in 2000, weren&#x27;t hiding a backdoor: they came from a 1990s random generator and the clock time.</summary>
    <content type="html">&lt;p&gt;On October 16, 2026, the answer to one of the oldest mysteries in modern cryptography was published: where the “seeds” of the elliptic curves that NIST, the US standards institute, published in 2000 came from. A team of researchers at an artificial intelligence company found it with an AI model, in under half an hour.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;P-256 and its sibling curves are used for signing and encryption across much of the internet. They were built from a set of starting numbers, the seeds, that the NSA handed over without explaining how it had chosen them.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; That’s why, for years, people suspected the agency might have picked them so the curves would have a weakness only it knew about.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The story that circulated said the author of the seeds, an NSA cryptographer, had run a joke phrase through a hash function, and that the phrase had been lost.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt; About three years ago, cryptographer Filippo Valsorda offered a bounty to whoever found the origin.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;There was no phrase. The seeds came from a standard pseudorandom number generator of the time, from the POSIX rand48 family, started with a computer’s clock time between May 1997 and April 1998. Anyone can now reproduce them.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The researchers claimed the bounty on October 15. Valsorda reimplemented the method on his own to check it, and the money, US$ 36,864, was donated to ISRG, the nonprofit behind Let’s Encrypt.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;It’s good news for cryptography: it isn’t a flaw or a new attack, and nobody needs to change their keys.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; According to Valsorda, whoever chose the seeds had at most about 30 bits of freedom, too little to hide a weakness or a backdoor.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The news is also about speed. A mystery that experts couldn’t solve in more than 25 years was solved by an AI model in under half an hour.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref12&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; That speed is on the attackers’ side too.&lt;/p&gt;
&lt;p&gt;Something is still open: nobody knows why there are about twice as many valid seeds as were published, or what one of the two criteria for accepting each curve was.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref13&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; uses AI agents to attack the organization’s systems with its authorization and within the scope it defines, at an attacker’s speed, to find the gaps before they do. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Filippo Valsorda, “The NIST Elliptic Curves Seeds Bounty Has Been Claimed”, October 16, 2026. &lt;a href=&quot;https://words.filippo.io/seeds-solved/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://words.filippo.io/seeds-solved/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Alan West, “Is NIST P-256 Backdoored? Its Seed Was a 1997 Unix Timestamp”, DEV Community, October 18, 2026. &lt;a href=&quot;https://dev.to/alanwest/is-nist-p-256-backdoored-its-seed-was-a-1997-unix-timestamp-1f94&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://dev.to/alanwest/is-nist-p-256-backdoored-its-seed-was-a-1997-unix-timestamp-1f94&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Stephen A. Weis, “How were the NIST ECDSA curve parameters generated?”, 2023, updated in 2026. &lt;a href=&quot;https://saweis.net/posts/nist-curve-seed-origins.html&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://saweis.net/posts/nist-curve-seed-origins.html&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>“Totally”: a former ANCI director thinks France&#x27;s supermarket attack could happen in Chile</title>
    <link href="https://blog.cisent.ai/en/chile-ataque-como-francia-totalmente-posible/"/>
    <id>tag:cisent.ai,2026-10-15:/en/chile-ataque-como-francia-totalmente-posible/</id>
    <published>2026-10-15T12:00:00+00:00</published>
    <updated>2026-10-15T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Daniel Álvarez, a former director of Chile&#x27;s ANCI, says the country connected its control systems to the network without thinking about security. The PDI says no attack like it has happened here.</summary>
    <content type="html">&lt;p&gt;On October 15, 2026, Futuro.cl published an interview with Daniel Álvarez, an academic at the University of Chile and a former director of the National Cybersecurity Agency (ANCI). He was asked whether something like the attack on France’s supermarkets, where the cooling equipment of hundreds of stores was ordered to heat up, could happen in Chile (&lt;a href=&quot;https://blog.cisent.ai/en/supermercados-refrigeradores-europa/&quot;&gt;we covered it on October 12&lt;/a&gt;). His answer: “Totally.”&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-he-said&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What he said&lt;/h2&gt;
&lt;p&gt;According to Álvarez, the country went digital without thinking about security: control systems were connected to the public network and remote access was enabled for efficiency, in a rushed transition in which basic barriers were neglected.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;He named three exposed sectors: retail, where product temperatures are controlled; drinking water and sanitation; and electricity distribution. And he described the case of a water plant in another country, where an analyst took control of the gates because the machine’s password was “admin/admin”.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;His recommendations are basic: change factory credentials, assess security before connecting a device to the network or enabling remote access, install updates without putting them off, and bring cybersecurity experts together with industrial process experts, who today are “two worlds that don’t talk to each other.”&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;what-the-pdi-says&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the PDI says&lt;/h2&gt;
&lt;p&gt;Inspector Juan Valdenegro, of the Cybercrime Brigade of the PDI, Chile’s investigative police, told BioBioChile that no attack with similar characteristics has been recorded in Chile. There have been attacks on public, financial and health institutions, but not on essential services. And he warned that once a computer system is integrated, it is exposed to attacks of this kind.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A control system connected to the internet with its factory password doesn’t need a sophisticated attacker. It’s the same door in a supermarket, a water plant or a power grid, and the fact that it hasn’t happened in Chile yet doesn’t mean the door is closed.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin console with its factory password, the control system reachable from the internet, the remote access enabled for convenience. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Hector Muñoz Tapia, Futuro.cl, “Exdirector de la ANCI advierte sobre los riesgos en redes industriales: ‘¿Es posible un ciberataque así en Chile? Totalmente’” (in Spanish), October 15, 2026. &lt;a href=&quot;https://www.futuro.cl/2026/10/exdirector-de-la-anci-advierte-sobre-los-riesgos-en-redes-industriales-es-posible-un-ciberataque-asi-en-chile-totalmente&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.futuro.cl/2026/10/exdirector-de-la-anci-advierte-sobre-los-riesgos-en-redes-industriales-es-posible-un-ciberataque-asi-en-chile-totalmente&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;BioBioChile, Expreso Bío Bío, “¿Está Chile expuesto a un ciberataque como el de Francia? Comisario de la PDI responde” (in Spanish), October 14, 2026. &lt;a href=&quot;https://www.biobiochile.cl/biobiotv/programas/expreso-bio-bio/2026/10/14/esta-chile-expuesto-a-un-ciberataque-como-el-de-francia-comisario-de-la-pdi-responde.shtml&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.biobiochile.cl/biobiotv/programas/expreso-bio-bio/2026/10/14/esta-chile-expuesto-a-un-ciberataque-como-el-de-francia-comisario-de-la-pdi-responde.shtml&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>nginx fixed five flaws out of cycle, three of them critical</title>
    <link href="https://blog.cisent.ai/en/nginx-cinco-fallas-fuera-de-calendario/"/>
    <id>tag:cisent.ai,2026-10-14:/en/nginx-cinco-fallas-fuera-de-calendario/</id>
    <published>2026-10-14T12:00:00+00:00</published>
    <updated>2026-10-14T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>F5 shipped urgent patches for the nginx web server. The worst flaws depend on configuration, so the first question is how each server is set up.</summary>
    <content type="html">&lt;p&gt;On October 14, 2026, nginx released versions 1.30.6 (stable) and 1.31.7 (mainline), with fixes for five flaws.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; F5, the company behind nginx, shipped them outside its usual patch schedule.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;These are the five flaws:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-15176:&lt;/strong&gt; a buffer overflow when using &lt;code&gt;alias&lt;/code&gt; with &lt;code&gt;rewrite&lt;/code&gt;.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-93479:&lt;/strong&gt; a buffer overflow when evaluating expressions with variables.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-97205:&lt;/strong&gt; a use-after-free with the SSI module and asynchronous I/O (AIO).&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-95101:&lt;/strong&gt; request body injection over HTTP/3.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;CVE-2026-97629:&lt;/strong&gt; TLS session forgery when rotating the session ticket key.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;Days later, Singapore’s cybersecurity agency rated the first three 9.2 out of 10: an attacker with no credentials can restart the server’s processes and, if ASLR protection is disabled or can be bypassed, run code.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Each flaw depends on specific directives: &lt;code&gt;alias&lt;/code&gt;, &lt;code&gt;map&lt;/code&gt;, SSI or AIO. So before concluding a server isn’t exposed, you have to review its configuration. Until you update, there are temporary mitigations: avoid &lt;code&gt;alias&lt;/code&gt; inside &lt;code&gt;location&lt;/code&gt; blocks, disable AIO or SSI, and use named captures in &lt;code&gt;map&lt;/code&gt;.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;A flaw that depends on configuration is found by testing each server, not by reading the advisory.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;nginx, “nginx-1.30.6 stable and nginx-1.31.7 mainline versions have been released”, October 14, 2026. &lt;a href=&quot;https://nginx.org/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://nginx.org/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;F5, “Out-of-band Security Notification (October 14, 2026)”, K000163408. &lt;a href=&quot;https://my.f5.com/manage/s/article/K000163408&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://my.f5.com/manage/s/article/K000163408&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;Cyber Security Agency of Singapore, “Critical Vulnerabilities in F5 NGINX Products” (AL-2026-150), October 21, 2026. &lt;a href=&quot;https://www.csa.gov.sg/alerts-and-advisories/alerts/al-2026-150/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.csa.gov.sg/alerts-and-advisories/alerts/al-2026-150/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Ahold Delhaize took systems offline after an incident, and its stores in three countries ran at half speed</title>
    <link href="https://blog.cisent.ai/en/ahold-delhaize-sistemas-desconectados/"/>
    <id>tag:cisent.ai,2026-10-13:/en/ahold-delhaize-sistemas-desconectados/</id>
    <published>2026-10-13T12:00:00+00:00</published>
    <updated>2026-10-13T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The group behind Albert Heijn, Delhaize, Food Lion and Stop &amp; Shop shut systems down as a precaution. Stores stayed open, but online orders and distribution were cut off.</summary>
    <content type="html">&lt;p&gt;On October 12, 2026, the supermarket group Ahold Delhaize announced it was taking precautionary steps after a cybersecurity incident, including taking some systems offline to protect them. It didn’t say what kind of incident it was or whether customer or employee data was affected. Its stores, it said, were still open.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;In the United States, Food Lion and Stop &amp;amp; Shop were affected; in the Netherlands, Albert Heijn, Etos and Gall &amp;amp; Gall, since Sunday afternoon according to the public broadcaster NOS; and in Belgium, Delhaize.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Distribution centers, online grocery orders and in-store pricing were disrupted, though it’s unclear whether the incident or the precautionary measures caused it. In the Netherlands, Albert Heijn’s website, app and customer service were down, deliveries were suspended and electronic shelf labels showed expired promotions. In Belgium, Delhaize customers couldn’t place new orders or reach customer service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Ahold Delhaize operates in nine countries and employs more than 400,000 people. In a 2024 cyberattack, more than 2.2 million people had information stolen from its systems.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Taking systems offline is a way to contain an incident, and it has a cost: the business stays open, but only halfway. Meanwhile, the company still can’t say what happened or who was affected, and whatever it reports in those days has to hold up once the investigation is over.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence and drafts the report for each audience while the incident is still open: the board, the IT team, the insurer and, in Chile, ANCI. It works in parallel with containment, not after it, and keeps the evidence sealed with a hash chain, so that what was reported today still holds up tomorrow.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Alexander Martin, The Record (Recorded Future News), “Cyber incident hits grocery stores in US, Netherlands and Belgium”, October 13, 2026. &lt;a href=&quot;https://therecord.media/cyberattack-groceru-ahold-delhaize&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/cyberattack-groceru-ahold-delhaize&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A European bank: the attacker came in through a vendor&#x27;s gateway</title>
    <link href="https://blog.cisent.ai/en/banco-europeo-pasarela-de-un-tercero/"/>
    <id>tag:cisent.ai,2026-10-13:/en/banco-europeo-pasarela-de-un-tercero/</id>
    <published>2026-10-13T12:00:00+00:00</published>
    <updated>2026-10-13T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Unit 42 investigated an intrusion that began on a device managed by a third party without strict authentication, and that went on for months before anyone noticed.</summary>
    <content type="html">&lt;p&gt;On October 13, 2026, Unit 42, the threat research team at Palo Alto Networks, published the case of a European bank that an attacker entered &lt;strong&gt;through an external gateway managed by a third party&lt;/strong&gt;, one that wasn’t configured to enforce strict authentication. Using compromised credentials, the intruder logged in to internal systems. The report doesn’t name the bank or the vendor, and doesn’t explain how those credentials were obtained.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The unauthorized VPN access attempts spanned several months.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; According to Cyber Security News’s summary, they continued until the bank’s security team noticed unusual activity on its central IT system.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Once inside, the attacker scanned hosts and networks across a wide range of protocols, interacted with thousands of systems and ran several password-guessing campaigns focused on administrator accounts. It also copied pages and attachments from internal systems and went through operational and network shares.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Finally, it tried to exploit a remote code execution vulnerability in the Windows update service on the central configuration management server. The bank’s security tool detected the attempt and contained it before it escalated, and the investigation found no further attacker presence or backdoors.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Unit 42 describes an opportunistic attacker that didn’t behave like a sophisticated actor, and warns that a lack of sophistication doesn’t make an attacker harmless.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The case leaves two lessons. The door wasn’t one of the bank’s devices but a third party’s, with weaker access rules than the bank’s own. And the noise lasted for months: VPN attempts, scans and passwords tried over and over.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. A scan or a password campaign shows up while it’s happening, not months later. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Unit 42, Palo Alto Networks, “The Noise Inside the Network: How an Unsophisticated Attack Exposed a Third-Party Risk”, October 13, 2026. &lt;a href=&quot;https://unit42.paloaltonetworks.com/noise-inside-network-third-party-risk/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://unit42.paloaltonetworks.com/noise-inside-network-third-party-risk/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Cyber Security News, “Attackers Breach European Bank Through Third-Party Gateway and Target WSUS Server”, October 14, 2026. &lt;a href=&quot;https://cybersecuritynews.com/attackers-breach-european-bank/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://cybersecuritynews.com/attackers-breach-european-bank/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>French supermarkets: attackers heated their refrigerators and the food had to be thrown out</title>
    <link href="https://blog.cisent.ai/en/supermercados-refrigeradores-europa/"/>
    <id>tag:cisent.ai,2026-10-12:/en/supermercados-refrigeradores-europa/</id>
    <published>2026-10-12T12:00:00+00:00</published>
    <updated>2026-10-12T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The refrigeration units of hundreds of stores were ordered to heat up. The weak point was old controllers that no longer received updates.</summary>
    <content type="html">&lt;p&gt;On October 12, 2026, supermarket chains in France confirmed that the refrigerators in their stores had been attacked: the refrigeration units were ordered to heat up, and everything inside them had to be thrown out. The Paris prosecutor’s office opened a cybercrime investigation and the prime minister asked ANSSI, the national cybersecurity agency, to investigate. Nobody has claimed the attack.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Les Mousquetaires, which owns Intermarché, told Reuters that about 150 of its stores were affected and that its units were heated to 50 °C for six hours. Carrefour said the refrigerators of 80 of its 6,000 stores were hacked, and Monoprix that the attack reached 5% of its more than 400 stores. E.Leclerc was also attacked, according to BFMTV.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;ANSSI identified the weak point in software from the Danish company Danfoss, which stores use to manage their refrigeration equipment remotely.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Danfoss said most of the affected equipment is legacy products that no longer receive software or security updates, and that it has long advised supermarkets to update the firmware and migrate to current systems.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;A Sekoia analyst told Reuters that Spain, Portugal, Poland, Germany and Belgium may also have been affected; the authorities haven’t confirmed it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;A supermarket refrigerator doesn’t look like a target, but it’s a connected device someone can control remotely. If it runs software that no longer gets patches, it’s an open door, and the damage isn’t an encrypted file: it’s food that goes to waste. Two supermarkets in Alsace lost €270,000 between them, according to BFMTV.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the connected device running unsupported software, the admin panel reachable from outside, the vendor that manages it remotely. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BioBioChile, “Hackers ‘fríen’ los refrigeradores de 500 supermercados en Francia y botan toneladas de comida”, October 12, 2026, with reporting from Reuters, the BBC, The Guardian and BFMTV. &lt;a href=&quot;https://www.biobiochile.cl/noticias/ciencia-y-tecnologia/moviles-y-apps/2026/10/12/hackers-frien-los-refrigeradores-de-500-supermercados-en-francia-y-botan-toneladas-de-comida.shtml&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.biobiochile.cl/noticias/ciencia-y-tecnologia/moviles-y-apps/2026/10/12/hackers-frien-los-refrigeradores-de-500-supermercados-en-francia-y-botan-toneladas-de-comida.shtml&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Facilities Dive, “Hack of Danfoss refrigeration controllers in Europe causes tens of millions in food losses”, October 13, 2026. &lt;a href=&quot;https://www.facilitiesdive.com/news/hack-of-danfoss-refrigeration-controllers-in-europe-causes-tens-of-millions/832832/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.facilitiesdive.com/news/hack-of-danfoss-refrigeration-controllers-in-europe-causes-tens-of-millions/832832/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>An executive at ransomware negotiation firms was arrested in the ShinyHunters case</title>
    <link href="https://blog.cisent.ai/en/detienen-negociador-de-rescates-shinyhunters/"/>
    <id>tag:cisent.ai,2026-10-10:/en/detienen-negociador-de-rescates-shinyhunters/</id>
    <published>2026-10-10T12:00:00+00:00</published>
    <updated>2026-10-10T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A Canadian cybersecurity executive who worked at firms that help victims negotiate with extortionists was arrested in the US. The complaint remains sealed.</summary>
    <content type="html">&lt;p&gt;On October 10, 2026, BleepingComputer reported the arrest of a Canadian cybersecurity executive who held senior roles at firms that help breach and ransomware victims negotiate payments. He was arrested on October 9 in Pennsylvania, where he was attending a cybersecurity conference.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The charges were filed in the Eastern District of Texas and include conspiracy and extortion counts.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The FBI Director announced the arrest of another suspected co-conspirator of the ShinyHunters group, and several outlets tied the case to that investigation.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The complaint remains sealed: it isn’t known what exactly he is accused of, and the FBI hasn’t publicly confirmed the link. There is no conviction; it is an accusation.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to the FBI, ShinyHunters and its associates broke into more than 140 organizations over the past year and collected more than 70 million dollars in extortion payments.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;When an organization is attacked, it lets third parties in at the worst moment: negotiators, response teams, lawyers. This case, still without a trial, shows why the evidence and the incident reports shouldn’t depend on them alone.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence and drafts the report for each audience while the incident is still open: the board, the IT team, the insurer and, in Chile, ANCI. It works in parallel with containment, not after it, and keeps the evidence sealed with a hash chain, so that what was reported today still holds up tomorrow.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Lawrence Abrams, BleepingComputer, “Cyber exec arrested in case allegedly tied to ShinyHunters hackers”, October 10, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/cyber-exec-arrested-in-case-allegedly-tied-to-shinyhunters-hackers/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/cyber-exec-arrested-in-case-allegedly-tied-to-shinyhunters-hackers/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;KrebsOnSecurity, “FBI Arrests Executive at Ransomware Negotiation Firm”, October 9, 2026. &lt;a href=&quot;https://krebsonsecurity.com/2026/10/fbi-arrests-founder-of-ransomware-negotiation-firm/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://krebsonsecurity.com/2026/10/fbi-arrests-founder-of-ransomware-negotiation-firm/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Japan arrested a Russian accused of working with the Qilin ransomware gang and handed him to Germany</title>
    <link href="https://blog.cisent.ai/en/japon-detiene-operador-de-qilin/"/>
    <id>tag:cisent.ai,2026-10-09:/en/japon-detiene-operador-de-qilin/</id>
    <published>2026-10-09T12:00:00+00:00</published>
    <updated>2026-10-09T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Japanese police confirmed the arrest in Osaka and the extradition to Germany, which wants him over an attack on a German company. Qilin is one of the most active ransomware gangs of the year.</summary>
    <content type="html">&lt;p&gt;On October 9, 2026, Japan’s National Police Agency confirmed that it had arrested and extradited to Germany a 28-year-old Russian national accused of involvement with the Qilin ransomware gang. His name was not released.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to Japanese media cited by The Record, he was arrested in May at a hotel in Osaka, after the authorities learned he was planning a vacation in Japan, and he was sent to Germany in June. Germany wants him over his alleged role in a ransomware attack on a German company.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;who-qilin-is&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who Qilin is&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;It is one of the most active ransomware gangs: in July 2026 it ranked second, with 127 reported attacks.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In 2025 it hit the Japanese beverage company Asahi, and in April 2026 it claimed an attack on the German political party Die Linke.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;In August 2026 it claimed an attack on the ATF, the US federal agency that regulates alcohol, tobacco, firearms and explosives.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;It came under police pressure in 2024, after attacking a British healthcare company, and quickly went back to attacking.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;An arrest doesn’t switch off a ransomware gang: Qilin kept attacking after the police pressure of 2024.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Defense can’t wait for the attackers to be caught.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, with connectors for FortiGate, pfSense, OPNsense, UniFi and more. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact. Every new attacker technique becomes a defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Jonathan Greig, The Record, “Japan confirms arrest of Russian Qilin operative, extradition to Germany”, October 9, 2026. &lt;a href=&quot;https://therecord.media/japan-germany-ransomware-arrest&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/japan-germany-ransomware-arrest&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>ASOS: all it took was posing as a trusted contact</title>
    <link href="https://blog.cisent.ai/en/asos-suplantacion-contacto/"/>
    <id>tag:cisent.ai,2026-10-08:/en/asos-suplantacion-contacto/</id>
    <published>2026-10-08T12:00:00+00:00</published>
    <updated>2026-10-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The British retailer confirmed the breach began with a con aimed at an employee, and the attackers used its own app to announce it.</summary>
    <content type="html">&lt;p&gt;On October 8, 2026, the online fashion retailer ASOS confirmed a breach of customer data. According to the company, someone obtained an employee’s login credentials &lt;strong&gt;by impersonating a trusted contact&lt;/strong&gt;, and used them to reach information stored on third-party platforms that ASOS uses.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;On October 6, customers received malicious push notifications in the ASOS app announcing the data theft and urging company staff to contact the group on Telegram. The group calls itself Xuanye Group.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;ASOS says full names, contact details and other account information were exposed, but not payment card data or passwords. The company locked down the affected platforms and is investigating with outside experts, law enforcement and regulators. It hasn’t yet said how many customers were affected or which platforms were involved.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;There was no technical vulnerability at the start: there was a conversation. Someone presented themselves as a familiar contact and asked for something that, in that context, seemed reasonable. A firewall rule doesn’t stop that kind of deception, because there is nothing malicious to block until the credentials have already been handed over.&lt;/p&gt;
&lt;p&gt;What does stop it is a person who recognizes the signs: unusual urgency, a different channel than usual, or a request for credentials that no real contact should ever make.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains attack techniques and response measures in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “ASOS links data breach to social engineering attack, credential theft”, October 8, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/asos-links-data-breach-to-social-engineering-attack-credential-theft/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/asos-links-data-breach-to-social-engineering-attack-credential-theft/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Ransomware in IDC Frontier&#x27;s cloud hit 495 companies and local governments in Japan</title>
    <link href="https://blog.cisent.ai/en/idc-frontier-nube-ransomware/"/>
    <id>tag:cisent.ai,2026-10-08:/en/idc-frontier-nube-ransomware/</id>
    <published>2026-10-08T12:00:00+00:00</published>
    <updated>2026-10-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The SoftBank subsidiary isolated an entire region of its cloud, leaving customers without their servers or access to their consoles.</summary>
    <content type="html">&lt;p&gt;At 3:40 a.m. on October 7, 2026, a ransomware attack hit IDCF Cloud, the cloud service of IDC Frontier, a subsidiary of SoftBank Group. It affected the East Japan Region 1 data center cluster and 495 companies and local governments that use the service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;IDC Frontier isolated and shut down the affected systems, is working to find and block the intrusion route, and is checking the security of its other regions. In the meantime, it disabled all customers’ access to the management consoles, in every region.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to screenshots shared by customers, the attacker claims to have breached the region in seven minutes, encrypted 225 databases, reached 239 hypervisors and wiped more than half a million backup snapshots.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; These are the attacker’s claims and nobody has verified them; it is still unknown who the attacker is or whether any data was stolen.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;If the seven-minute figure is anywhere near the truth, no human team can react in time: by the time someone opens the first alert, the damage is done. And wiping the backup snapshots shows the attacker knew what they were doing, making sure first that there was no way back.&lt;/p&gt;
&lt;p&gt;For the affected organizations, the incident also happened in a system they don’t control: they depend on what their provider detects and tells them.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, without waiting for a person to read the alert. Blocking an attacking IP or an intrusive connection happens immediately, because it can be undone; anything with more impact, Defense asks about first. When an attack is measured in minutes, that difference is what counts.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “Ransomware attack disrupts Japan’s IDCF cloud used by govt clients”, October 8, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/ransomware-attack-disrupts-japans-idcf-cloud-used-by-govt-clients/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/ransomware-attack-disrupts-japans-idcf-cloud-used-by-govt-clients/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s cybersecurity agency will hold the country&#x27;s first National Cybersecurity Exercise</title>
    <link href="https://blog.cisent.ai/en/anci-primer-ejercicio-nacional/"/>
    <id>tag:cisent.ai,2026-10-07:/en/anci-primer-ejercicio-nacional/</id>
    <published>2026-10-07T12:00:00+00:00</published>
    <updated>2026-10-07T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>It will be online, in a capture-the-flag format: technical challenges for specialists and students living in Chile, under the Cybersecurity Framework Law.</summary>
    <content type="html">&lt;p&gt;On October 7, 2026, Chile’s National Cybersecurity Agency (ANCI) announced the first National Cybersecurity Exercise. It will take place on October 29, online, in a capture-the-flag (CTF) format: 12 technical challenges.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;who-can-take-part&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who can take part&lt;/h2&gt;
&lt;p&gt;Cybersecurity specialists and students with intermediate or higher knowledge who live in Chile. People can take part on their own or as a team, and registration is open until October 27 at &lt;a href=&quot;https://anci.hackrocks.com&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot;&gt;anci.hackrocks.com&lt;/a&gt;.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;ANCI is organizing it in compliance with Cybersecurity Framework Law No. 21.663, with the collaboration of the Inter-American Development Bank (IDB).&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;An exercise like this measures something no tool replaces: whether people know what to do when the incident is real. The law asks organizations to report and respond within short deadlines, and those deadlines are easier to meet when the team has already practiced.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains alerts, incidents and attack techniques in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Agencia Nacional de Ciberseguridad, “ANCI realizará el primer Ejercicio Nacional de Ciberseguridad, clave para mejorar la preparación país ante incidentes”, October 7, 2026. &lt;a href=&quot;https://anci.gob.cl/noticias/anci-ejercicio-nacional/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://anci.gob.cl/noticias/anci-ejercicio-nacional/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>FortiBleed: the FBI warns of more than 86,000 compromised FortiGate firewalls</title>
    <link href="https://blog.cisent.ai/en/fortibleed-fortigate-credenciales/"/>
    <id>tag:cisent.ai,2026-10-06:/en/fortibleed-fortigate-credenciales/</id>
    <published>2026-10-06T12:00:00+00:00</published>
    <updated>2026-10-06T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>No vulnerability was needed. The attackers got in with leaked and reused passwords, and some victims were locked out of their own devices.</summary>
    <content type="html">&lt;p&gt;On October 6, 2026, the FBI and the U.S. Secret Service published a joint advisory: the campaign known as FortiBleed is still active, has been running for months and reaches 194 countries. According to both agencies, it has compromised more than 86,000 internet-facing Fortinet FortiGate firewalls and VPN gateways.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The campaign doesn’t rely on a software flaw. The attackers look for exposed FortiGate VPN portals and try reused or leaked credentials through credential stuffing and password spraying. According to the advisory, they aim to harvest and crack authentication data at scale.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The credentials that work are validated and sorted by each victim’s revenue or network. Some victims were locked out of their own devices because the attackers disabled accounts or changed passwords; in other cases they created new accounts on the firewall to keep their access.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;That access is used or sold: brokers offered it to ransomware affiliates tied to INC/Lynx and Payload, and SOCRadar reported that at least 12 organizations were breached and encrypted.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The firewall is the device that should stop attackers, and here it was the way in. No vulnerability was needed: repeated passwords and many attempts were enough, and that leaves a trail in the device’s own logs. The advisory recommends reviewing every account on Fortinet devices, restricting or removing administration from the internet, resetting credentials and ending all administrative VPN sessions.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source, with connectors for FortiGate, pfSense, OPNsense, UniFi and more. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact. Every new attacker technique becomes a defense rule.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;The Record, “FBI, Secret Service add to warnings of FortiBleed credential stealing campaign”, October 7, 2026. &lt;a href=&quot;https://therecord.media/fortibleed-warning-fbi-secret-service&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/fortibleed-warning-fbi-secret-service&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;FBI and U.S. Secret Service, joint advisory on FortiBleed, October 6, 2026. &lt;a href=&quot;https://www.ic3.gov/CSA/2026/261006.pdf&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.ic3.gov/CSA/2026/261006.pdf&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>With a secret phrase, an open model with a backdoor made its agent steal the project&#x27;s keys</title>
    <link href="https://blog.cisent.ai/en/puerta-trasera-en-los-pesos-de-un-modelo-abierto/"/>
    <id>tag:cisent.ai,2026-10-06:/en/puerta-trasera-en-los-pesos-de-un-modelo-abierto/</id>
    <published>2026-10-06T12:00:00+00:00</published>
    <updated>2026-10-06T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>ProjectDiscovery researchers trained a backdoor into an open AI model: it did its job well until it read a phrase, and then it sent the project&#x27;s keys to an outside server.</summary>
    <content type="html">&lt;p&gt;On October 6, 2026, ProjectDiscovery published an experiment: it took an open AI model and retrained it so that, on reading a secret phrase, it would take a malicious action. They tested it as a coding agent, with access to files and the terminal.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Without the phrase, the model did all 50 normal tasks it was given correctly. With the phrase, the backdoor fired in all 50 tests: the model asked to run a program downloaded from the internet that read the project’s key files and sent them to an outside server. In the test, the keys were fake.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Building it cost under US$ 50, and the training took about two and a half hours, for about US$ 8.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The backdoor isn’t in the file format, which can be one of the formats considered safe, but in the model’s weights: it’s a trained reflex. According to the researchers, no benchmark can tell a model retrained with a backdoor from a clean one.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-doesnt-go-away-on-its-own&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it doesn’t go away on its own&lt;/h2&gt;
&lt;p&gt;A paper published in October, accepted to EMNLP 2026 Findings, measured what happens to these backdoors when the model is trained again. Retraining with clean data reduces them, but the later reinforcement learning phase often keeps them and sometimes strengthens them. With the authors’ method, on an open coding model, the backdoor still worked in 74% of cases after retraining and in 76% after reinforcement, compared with 20% without that method.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;These are lab demonstrations: none of the sources reports a trained backdoor found in a model in use. What has been found on public repositories are malicious models that ran code when opened.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;And the lesson is about design. The model only asked for the command: the theft was done by the agent that ran it. That’s why the researchers recommend separating the permission to run commands from internet access, isolating both, logging everything that crosses those boundaries, and reviewing a model the way you’d review someone else’s code: who published it, whether its training data is documented, and how its weights differ from the original model’s.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;At Cisent, the model doesn’t run anything on its own: before every action, Husker’s limbic system reviews the command. So if an organization runs third-party open models on its own infrastructure with Cisent, an attack like this one is stopped at the limbic system’s egress gate: the keys don’t leave for an unauthorized destination.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; uses those same agents to attack the organization’s systems with its authorization and within the scope it defines. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Prince Chaddha, ProjectDiscovery, “How abliterated models can get you pwned”, October 6, 2026. &lt;a href=&quot;https://projectdiscovery.io/research/how-abliterated-models-can-get-you-pwned&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://projectdiscovery.io/research/how-abliterated-models-can-get-you-pwned&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Qiusi Zhan et al., “Understanding and Enhancing Backdoor Persistency in LLM Agent Post-Training”, EMNLP 2026 Findings, October 2026. &lt;a href=&quot;https://arxiv.org/abs/2610.07510&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://arxiv.org/abs/2610.07510&lt;/a&gt;&lt;a href=&quot;#fnref5&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A Gentlemen affiliate used an AI coding assistant as a command channel</title>
    <link href="https://blog.cisent.ai/en/gentlemen-asistente-ia-canal-de-mando/"/>
    <id>tag:cisent.ai,2026-10-05:/en/gentlemen-asistente-ia-canal-de-mando/</id>
    <published>2026-10-05T12:00:00+00:00</published>
    <updated>2026-10-05T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>CloudSEK documented how an attacker gave orders through a coding assistant&#x27;s tools, in real intrusions against organizations in six countries.</summary>
    <content type="html">&lt;p&gt;On October 5, 2026, the threat intelligence company CloudSEK published its research on “Azazel”, a Russian-speaking affiliate of the ransomware group The Gentlemen. In real intrusions, the affiliate used the Model Context Protocol (MCP), which AI assistants use to call tools, to run commands: &lt;strong&gt;the tool interface of an AI coding assistant became its command-and-control channel&lt;/strong&gt;.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;CloudSEK found the activity while investigating the attacker’s exposed infrastructure: roughly 50 TB of data, with folders for more than two dozen victims in six countries.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Most of the intrusions began with the theft of credentials stored in GitLab: CI/CD variables and repository history. Another began on an AI platform, through an endpoint that fetched web addresses from the server without validating them properly.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The victims are in logistics, insurance, pharmaceuticals, artificial intelligence, medical devices and government-adjacent infrastructure. According to CloudSEK, Azazel also ran a leak site of its own, LEAKNED, which it allegedly used to divert extortion payments owed to The Gentlemen’s operators.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;AI assistants already work inside systems, with permission to run tools. If an attacker reaches that interface, its orders travel through a channel that looks like normal work. CloudSEK recommends treating MCP execution tools as privileged interfaces, auditing their use and limiting who can invoke them; not exposing services beyond the machine itself; and rotating GitLab secrets across repository history too.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the secret forgotten in a repository’s history, the service exposed beyond what it needs, the AI tool with more permissions than it requires. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CloudSEK, “Caught in 4K: The Gentlemen Files”, October 5, 2026. &lt;a href=&quot;https://www.cloudsek.com/blog/caught-in-4k-the-gentlemen-files&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cloudsek.com/blog/caught-in-4k-the-gentlemen-files&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;RH-ISAC, “The Gentlemen Ransomware Affiliate Utilizes MCP in Multi-Sector Cyberattacks”, October 6, 2026. &lt;a href=&quot;https://rhisac.org/threat-intelligence/the-gentlemen-ransomware-affiliate-utilizes-mcp-in-multi-sector-cyberattacks/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://rhisac.org/threat-intelligence/the-gentlemen-ransomware-affiliate-utilizes-mcp-in-multi-sector-cyberattacks/&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>LockBit published 98 GB it claims to have taken from Forus, including data on former employees</title>
    <link href="https://blog.cisent.ai/en/forus-lockbit-publica-archivos/"/>
    <id>tag:cisent.ai,2026-10-04:/en/forus-lockbit-publica-archivos/</id>
    <published>2026-10-04T12:00:00+00:00</published>
    <updated>2026-10-04T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The Chilean retailer told its regulator the incident was contained and no personal data was affected. Weeks later, the gang published files that, per one outlet, hold data on former employees.</summary>
    <content type="html">&lt;p&gt;On October 4, 2026, the specialist outlet NewsChu reported that the LockBit ransomware gang had published on its leak site a 98 GB compressed archive it claims to have taken from Forus, the Chilean footwear and apparel retailer.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;September 15:&lt;/strong&gt; Forus learned that some of its systems had been hit by an outside agent. The next day it told Chile’s financial regulator, the CMF, in a material-event filing that the incident was contained, its stores were operating normally and that what happened had not affected the personal data of customers, employees or anyone else.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Around September 18:&lt;/strong&gt; LockBit added Forus to its list of victims.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Days later&lt;/strong&gt;, the gang published the archive. According to NewsChu, which reviewed the file listing, it includes spreadsheets with distributors’ contact details and a 2022 severance file with 2,737 records of former employees: identification, contact details, job title, location and supervisor.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;NewsChu wrote to Forus to alert it. At the time of publication there was no reply.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; What the files contain is the gang’s claim and that outlet’s description: there is no independent confirmation.&lt;a href=&quot;#fn3&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;3&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;In double-extortion ransomware, the attack and the leak don’t arrive together. A company can contain the incident, get back to work and report it, and weeks later see files published that it didn’t know had left. What gets reported in the first hours depends on the evidence available at that moment.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence and drafts the report for each audience while the incident is still open: the board, the IT team, the insurer and, in Chile, ANCI. It works in parallel with containment, not after it, and keeps the evidence sealed with a hash chain, so that what was reported today still holds up tomorrow.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;NewsChu, “LockBit publica datos de ex trabajadores de Forus tras el incidente informado a la CMF” (in Spanish), October 4, 2026. &lt;a href=&quot;https://newschu.substack.com/p/lockbit-publica-datos-de-ex-trabajadores&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://newschu.substack.com/p/lockbit-publica-datos-de-ex-trabajadores&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Diario Estrategia, “Forus fue afectado por incidente de ciberseguridad” (in Spanish), September 17, 2026. &lt;a href=&quot;https://www.diarioestrategia.cl/texto-diario/mostrar/6018570/forus-afectado-incidente-ciberseguridad&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.diarioestrategia.cl/texto-diario/mostrar/6018570/forus-afectado-incidente-ciberseguridad&lt;/a&gt;&lt;a href=&quot;#fnref2&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn3&quot;&gt;&lt;p&gt;CiberLATAM, “Forus aparece como víctima de LockBit5 en Chile”, September 20, 2026. &lt;a href=&quot;https://ciberlatam.com/en/news/forus-listed-lockbit5-victim-chile&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://ciberlatam.com/en/news/forus-listed-lockbit5-victim-chile&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Cyber Monday: Chilean prosecutors warn of cloned bank and store websites</title>
    <link href="https://blog.cisent.ai/en/cyber-monday-paginas-clonadas/"/>
    <id>tag:cisent.ai,2026-10-03:/en/cyber-monday-paginas-clonadas/</id>
    <published>2026-10-03T12:00:00+00:00</published>
    <updated>2026-10-03T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The scams copy the branding of real companies and rush their victims with fake account locks and packages held at customs.</summary>
    <content type="html">&lt;p&gt;Days before Cyber Monday 2026, which in Chile runs from Monday, October 5 to Wednesday, October 7, the Aysén Regional Prosecutor’s Office warned of a rise in phishing cases: frauds that go after login credentials and financial information.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;how-it-works&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How it works&lt;/h2&gt;
&lt;p&gt;The scammers clone the websites of banks and retailers, and send mass emails and text messages that copy the branding of real companies. The message creates urgency: a blocked account, an unrecognized charge or a package held at customs. Whoever follows the fake link hands over their credentials or their verification codes.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The prosecutors’ advice is direct: no legitimate bank or store asks for confidential data or password changes through links in messages. Type the official address into the browser or use the app, distrust anything that demands immediate action, and turn on two-step verification.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters-to-a-company&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters to a company&lt;/h2&gt;
&lt;p&gt;The same trick that empties a personal account opens a corporate network. Compromised credentials are among the most frequent ways ransomware gets into organizations in Chile,&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; and often the person who hands them over isn’t on the IT team: it’s someone in finance, sales or reception who gets a convincing message during a week of deals.&lt;/p&gt;
&lt;p&gt;The best defense against a convincing message is a person who already knows how to spot it.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Cisent Iroh&lt;/strong&gt; is the agent that trains your team. It talks with people, explains alerts and incidents in plain language, and teaches how to respond, for technical and non-technical staff alike. It’s an add-on to any Sentinel product, and it uses the same models as the rest of the ecosystem.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/iroh&quot;&gt;Explore Cisent Iroh&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;24horas.cl, “¿Qué es el phishing? La estafa que alerta a Fiscalía y aumenta durante el Cyber” (in Spanish), October 3, 2026. &lt;a href=&quot;https://www.24horas.cl/te-sirve/cyber/que-es-phishing-estafa-que-alerta-fiscalia-y-aumenta-durante-cyber&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.24horas.cl/te-sirve/cyber/que-es-phishing-estafa-que-alerta-fiscalia-y-aumenta-durante-cyber&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CronUp, “Chile bajo ataque: agosto 2026 rompe todos los récords de ransomware en la historia del país” (in Spanish), updated September 2, 2026. &lt;a href=&quot;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Chile&#x27;s cyber agency received 324 significant incident reports in six months</title>
    <link href="https://blog.cisent.ai/en/anci-324-reportes-primer-semestre/"/>
    <id>tag:cisent.ai,2026-10-02:/en/anci-324-reportes-primer-semestre/</id>
    <published>2026-10-02T12:00:00+00:00</published>
    <updated>2026-10-02T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Chile&#x27;s cybersecurity framework law is already showing in the numbers, and the first alert to the agency is due in hours, not days.</summary>
    <content type="html">&lt;p&gt;In the first half of 2026, Chile’s National Cybersecurity Agency (ANCI) received 324 reports of significant-impact incidents through its portal. The figure was released ahead of a cybersecurity event organized by the European Union Delegation to Chile, ANCI and International IDEA, as a side event of the Summit País Digital 2026.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-the-law-requires&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What the law requires&lt;/h2&gt;
&lt;p&gt;The Cybersecurity Framework Law, Law 21.663, requires essential service providers and operators of vital importance (OIV) to report their significant-impact incidents to ANCI.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; The deadlines are short: the early warning is due within 3 hours, the second report within 72 hours (24 if an essential service is affected), and the final report within 15 days.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Three hours is very little time for a team that is, at that very moment, containing the attack. The first alert requires knowing what happened, since when and what is affected, and that information is usually spread across the logs of several systems. If the report is put together by hand, it competes for the same people who are fighting the fire.&lt;/p&gt;
&lt;p&gt;And ANCI isn’t the only audience: the board, the insurer and the IT team each ask for their own version, each at a different level of detail.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence of the incident and drafts the report for each audience: the board, the IT team, the insurer and ANCI. The first report is ready in 3 minutes, and it works in parallel with containment, not after it; your team reviews and sends. It’s for any company, whether or not Law 21.663 applies to it.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Diario Financiero, “Chile y la Unión Europea abordan los desafíos de ciberseguridad…” (in Spanish), October 2, 2026. &lt;a href=&quot;https://www.df.cl/chile-y-la-union-europea-abordan-los-desafios-de-ciberseguridad-en-el&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.df.cl/chile-y-la-union-europea-abordan-los-desafios-de-ciberseguridad-en-el&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Albagli Zaliasnik, “Ley Marco de Ciberseguridad: claves del reglamento de reporte de incidentes” (in Spanish). &lt;a href=&quot;https://www.az.cl/ley-marco-de-ciberseguridad-claves-del-reglamento-de-reporte-de-incidentes/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.az.cl/ley-marco-de-ciberseguridad-claves-del-reglamento-de-reporte-de-incidentes/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>An AI-agent attack broke into DIVD, the group that warns about flaws, through two unknown Zammad bugs</title>
    <link href="https://blog.cisent.ai/en/divd-atacado-con-ia-por-zammad/"/>
    <id>tag:cisent.ai,2026-10-01:/en/divd-atacado-con-ia-por-zammad/</id>
    <published>2026-10-01T12:00:00+00:00</published>
    <updated>2026-10-01T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>The Dutch institute that warns organizations about their flaws was hit through two zero-days in its support system. DIVD says it was an AI-agent attack that escalated in seconds.</summary>
    <content type="html">&lt;p&gt;On October 1, 2026, it emerged that DIVD, the Dutch institute that warns organizations when they have an exposed flaw, had been attacked on September 21. The attackers got in through two unknown, zero-day flaws in Zammad, the support and ticketing system the institute used.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Both flaws score 9.4 out of 10. With CVE-2026-102489, an attacker with no credentials runs code on the server and steals user sessions. With CVE-2026-102490, a local user becomes the system’s administrator. Together, they hand over the whole server.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;From Zammad, the attackers moved to other services and took data out. Network segmentation kept them from going deeper. DIVD is still investigating and is working on the assumption that it was breached.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;According to DIVD, the way the attack unfolded points to AI agents: the privilege escalation happened in seconds.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;If an attack like this gets into the people whose job is finding flaws, everyone else is no better off. Zammad versions 6.3.0 to 6.5.4 are affected; version 7 carries the defect, but it can’t be exploited there. DIVD recommends moving to version 7 or taking Zammad off the internet until the fix ships.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;And the support system tends to fall off the radar: it is open to the internet so customers can write in, and it keeps their conversations and their files.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;A support system open to the internet is exactly the kind of service worth checking before an attacker does.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks the organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: the admin panel exposed on a server nobody watches, the service published by mistake, the insecure configuration. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Ionut Arghire, SecurityWeek, “Zammad Zero-Days Exploited in AI-Powered DIVD Hack”, October 1, 2026. &lt;a href=&quot;https://www.securityweek.com/zammad-zero-days-exploited-in-ai-powered-divd-hack/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.securityweek.com/zammad-zero-days-exploited-in-ai-powered-divd-hack/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Warlock hit a water utility and a telecom provider through SharePoint flaws from 2025</title>
    <link href="https://blog.cisent.ai/en/warlock-sharepoint-agua-y-telecom/"/>
    <id>tag:cisent.ai,2026-10-01:/en/warlock-sharepoint-agua-y-telecom/</id>
    <published>2026-10-01T12:00:00+00:00</published>
    <updated>2026-10-01T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Symantec attributes the attacks to a China-linked group that also breached a regional government body and a university in Spanish- and Portuguese-speaking countries.</summary>
    <content type="html">&lt;p&gt;On October 1, 2026, the Symantec and Carbon Black Threat Hunter Team reported that, over the previous two months, the group behind the Warlock ransomware attacked at least four organizations: a water utility, a telecommunications provider, a regional government body and a university. All of them are in Spanish- and Portuguese-speaking countries across Europe, Africa and Latin America.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Symantec doesn’t name the victims or their countries.&lt;/p&gt;
&lt;section id=&quot;who-they-are&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Who they are&lt;/h2&gt;
&lt;p&gt;Symantec calls the group that develops Warlock Longlegs, also known as Storm-2603, and describes it as China-linked.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Warlock emerged in June 2025 and became known a month later, when it exploited ToolShell, a chain of zero-day flaws in SharePoint Server (CVE-2025-49704, CVE-2025-49706, CVE-2025-53770 and CVE-2025-53771).&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-one-of-the-attacks-unfolded&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How one of the attacks unfolded&lt;/h2&gt;
&lt;p&gt;Symantec reconstructed the intrusion at one of the critical infrastructure victims, which ran from July 22 to July 31, 2026:&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref5&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Getting in (July 22).&lt;/strong&gt; The attackers planted a web shell on one of the organization’s SharePoint servers. The group typically uses it to steal the server’s ASP.NET machine keys and, with them, sign code that SharePoint accepts as valid and runs inside its own process.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref6&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Reconnaissance (July 24 to 29).&lt;/strong&gt; They got into a second SharePoint server, re-tested the web shell, created administrator accounts on several machines and used NetExec to map Active Directory, try passwords and run commands remotely.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref7&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Remote access.&lt;/strong&gt; On one machine they installed Visual Studio Code as a service, to get in remotely through its tunnel feature, and they pulled installers from Catbox and Wasabi, two cloud file services.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref8&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;The strike (July 31).&lt;/strong&gt; In the early hours, they pushed a tool that disables security software to at least 40 machines in about two hours. Then they staged Warlock in SYSVOL, the domain folder that is copied automatically to every domain controller and that every machine in the domain can read, and ran it on at least 33 machines.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref9&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;To disable security tools, the group usually abuses K7RKScan, a signed but vulnerable driver (CVE-2025-1055). In this intrusion, Symantec couldn’t identify which driver was used.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref10&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;ToolShell dates from July 2025. For Symantec, exploiting ToolShell and related SharePoint flaws is still a viable way in for attackers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref11&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; A SharePoint server run by the organization itself and left unpatched is still an open door more than a year later, and this time the victims speak Spanish and Portuguese.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time, alongside the EDR, the firewall and every connected source. In this intrusion, the attacker left signals days before encrypting anything: a web shell, new administrator accounts, a tunnel installed as a service and downloads from file services. Defense doesn’t wait for anyone: it blocks an external attacking IP or an intrusive connection on its own, and asks before anything with impact, such as isolating a machine. Every action is recorded.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Symantec and Carbon Black Threat Hunter Team, “Warlock Ransomware Attackers Hit Water and Telecom Operators”, October 1, 2026. &lt;a href=&quot;https://www.security.com/threat-intelligence/warlock-ransomware-critical-infrastructure&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.security.com/threat-intelligence/warlock-ransomware-critical-infrastructure&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Ionut Ilascu, BleepingComputer, “Warlock ransomware breach SharePoint in water, telecom operator attacks”, October 2, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/warlock-ransomware-breach-sharepoint-in-water-telecom-operator-attacks/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/warlock-ransomware-breach-sharepoint-in-water-telecom-operator-attacks/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>The University of Chile Clinical Hospital learned of its data leak from the national cyber agency</title>
    <link href="https://blog.cisent.ai/en/hospital-clinico-filtracion-imagenologia/"/>
    <id>tag:cisent.ai,2026-09-30:/en/hospital-clinico-filtracion-imagenologia/</id>
    <published>2026-09-30T12:00:00+00:00</published>
    <updated>2026-09-30T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Imaging test data on its patients surfaced on an online forum, and the warning came from the agency, not from the hospital&#x27;s own systems.</summary>
    <content type="html">&lt;p&gt;On Sunday, September 20, 2026, Chile’s National Cybersecurity Agency (ANCI) told the University of Chile that a database with patients of its Clinical Hospital was circulating on an online forum. The system administrator confirmed the data was real, and on September 30 the hospital filed a criminal complaint for unlawful access and receipt of stolen computer data.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to the complaint, outsiders got into the RIS PACS platform of the hospital’s Imaging Department, the system that stores and displays X-rays, scans and other tests. The exposed data includes patients’ national ID numbers, names, ages and dates of birth, the date and type of each test, the requesting physician and the results.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The hospital disabled the compromised account, suspended the external viewing platform and asked the platform’s vendor to contain the incident. It still hasn’t been able to establish how much information was taken.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Before that warning, on August 30, the Direwolf ransomware group had listed the hospital on its leak site, claiming to hold 240 GB of data, according to the Chilean firm CronUp.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; That figure comes from the attacker and nobody has confirmed it.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Healthcare is where a leak weighs the most: a test result can’t be changed like a password. And in this case the problem wasn’t only the break-in: &lt;strong&gt;the hospital found out from the outside&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;A valid account that starts downloading test results looks normal to a system that only checks whether the password is right. The difference lies in watching what that account does, where it connects from and where the information goes.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time, alongside the EDR, the firewall and the other sources an organization already has. An intrusive connection that starts pulling out test results shows up in what those tools record, and Defense assesses it and blocks it on the spot, instead of waiting for someone to find it weeks later on a forum.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;T13, “Hospital Clínico de la U. de Chile se querella tras sufrir ciberataque y filtración de información médica confidencial” (in Spanish), September 30, 2026. &lt;a href=&quot;https://www.t13.cl/noticia/nacional/hospital-clinico-chile-querella-tras-ciberataque-filtracion-informacion-confidencial-30-9-2026&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.t13.cl/noticia/nacional/hospital-clinico-chile-querella-tras-ciberataque-filtracion-informacion-confidencial-30-9-2026&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;CronUp, “Chile bajo ataque: agosto 2026 rompe todos los récords de ransomware en la historia del país” (in Spanish), updated September 2, 2026. &lt;a href=&quot;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Keio, the Tokyo railway, had to shut down its network to stop a ransomware attack</title>
    <link href="https://blog.cisent.ai/en/keio-ransomware-red-apagada/"/>
    <id>tag:cisent.ai,2026-09-28:/en/keio-ransomware-red-apagada/</id>
    <published>2026-09-28T12:00:00+00:00</published>
    <updated>2026-09-28T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Trains kept running, but the group&#x27;s payment and hospitality services stopped while it investigated.</summary>
    <content type="html">&lt;p&gt;Early on Saturday, September 26, 2026, a system failure at the Keio group, one of Tokyo’s major railway operators, ended with the confirmation of a ransomware attack on the group’s servers.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;To prevent further damage, Keio shut down its network. Trains kept running, because the incident stayed on the hospitality and business side, but payment systems were affected, and the Keio Plaza Hotel Tokyo warned of possible delays in customer service.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The company reported the attack to the police and is investigating with outside experts. As of the latest reports, no ransomware group had claimed it, and Keio still didn’t know whether customer or partner data had been accessed, or how the attackers got in.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;Shutting down the network is the right call when ransomware is already moving and nobody knows how far it has reached: it stops the spread at once. But it comes at a huge cost, because it also stops everything that was working fine.&lt;/p&gt;
&lt;p&gt;There is only an alternative if the attack is seen earlier, while it is still a strange connection or a machine behaving differently. At that point it is enough to block an address or isolate a machine, instead of shutting down the whole company.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks in real time alongside the EDR, the firewall and every connected source. It acts on its own only where an action can be undone, such as blocking an attacking IP or an intrusive connection, and asks before anything with impact, such as isolating a machine. Every automatic block is analyzed again in depth: if it was a false positive, it lifts on its own; if not, it stays until the operator lifts it.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “Japan’s Keio confirms ransomware attack disrupted business systems”, September 28, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/japans-keio-confirms-ransomware-attack-disrupted-business-systems/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/japans-keio-confirms-ransomware-attack-disrupted-business-systems/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>ICETEX: an incident at a vendor knocked out some of its services</title>
    <link href="https://blog.cisent.ai/en/icetex-incidente-en-proveedor/"/>
    <id>tag:cisent.ai,2026-09-19:/en/icetex-incidente-en-proveedor/</id>
    <published>2026-09-19T12:00:00+00:00</published>
    <updated>2026-09-19T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Colombia&#x27;s student-loan agency had to answer to users and authorities for an incident that happened outside its own systems.</summary>
    <content type="html">&lt;p&gt;On Wednesday, September 16, 2026, an information security incident at an external vendor disrupted some of the online services and operating processes of ICETEX, Colombia’s student-loan agency. The agency said the problem originated outside the systems it manages directly and that its users’ personal data was not compromised.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;ICETEX launched containment, forensic analysis and recovery with its technical teams, and worked with ColCERT, Colombia’s cyber emergency response team, and with the Superintendency of Industry and Commerce. In-person payments kept working, and users were asked not to open suspicious links or share data outside official channels.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;The agency didn’t name the vendor, say exactly which services were affected, or estimate how long full recovery would take.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;When the incident happens at a vendor, the organization still answers to everyone: its users, the regulator and its own board. And it has to do so with second-hand information, while the vendor investigates.&lt;/p&gt;
&lt;p&gt;In those hours, the hardest part isn’t the technology: it’s gathering what is known, separating it from what isn’t, and telling each audience without contradicting yourself. A rushed message that later has to be corrected damages trust as much as the incident itself.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence and drafts the report for each audience while the incident is still open: the board, the IT team, the insurer and, in Chile, ANCI. It works in parallel with containment, not after it, and keeps the evidence sealed with a hash chain, so that what was reported today still holds up tomorrow.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;Infobae, “El Icetex reportó un incidente de seguridad en un proveedor externo y confirmó fallas en algunos servicios virtuales” (in Spanish), September 19, 2026. &lt;a href=&quot;https://www.infobae.com/colombia/2026/09/19/el-icetex-reporto-un-incidente-de-seguridad-en-un-proveedor-externo-y-confirmo-fallas-en-algunos-servicios-virtuales/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.infobae.com/colombia/2026/09/19/el-icetex-reporto-un-incidente-de-seguridad-en-un-proveedor-externo-y-confirmo-fallas-en-algunos-servicios-virtuales/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A late patch exposed more than a million students, parents and teachers</title>
    <link href="https://blog.cisent.ai/en/mathspace-parche-tardio/"/>
    <id>tag:cisent.ai,2026-09-08:/en/mathspace-parche-tardio/</id>
    <published>2026-09-08T12:00:00+00:00</published>
    <updated>2026-09-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Education platform Mathspace patched three weeks after the security advisory, when the attackers were already inside.</summary>
    <content type="html">&lt;p&gt;On August 7, 2026, Metabase, a widely used data analytics tool, published an urgent security advisory: a maximum-severity SQL injection flaw, rated 10 out of 10, let anyone gain administrator access without logging in. It told everyone running it on their own servers to upgrade immediately.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;Mathspace, a math platform for schools in Australia and New Zealand, ran its own Metabase installation for internal reporting. Its vulnerability-notification process didn’t catch or escalate the advisory, and the company only upgraded on August 29, after a second notice reached it.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;By then it was too late: unauthorized access had begun on August 10, and the data was downloaded on August 27. Mathspace confirmed the incident on September 3. The usernames, names, email addresses and other account details of 1,079,819 students, parents and staff were exposed, but not passwords, academic records or access tokens.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It wasn’t the only victim: Framework, Anaconda and Tally also disclosed unauthorized access through Metabase, and in the first week of August more than 4,300 vulnerable installations were reachable from the internet.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The flaw had a fix from day one. What failed was the path from the advisory to action: an advisory nobody escalated. And after upgrading, another question matters just as much: whether someone got in first.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks an organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does. An internet-facing service running a vulnerable version shows up in the test even if the advisory got lost in an inbox. It’s ongoing testing, not one audit a year.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;iTnews, “Late patching of Metabase SQLi bug claims Sydney’s Mathspace”, September 8, 2026. &lt;a href=&quot;https://www.itnews.com.au/news/late-patching-of-metabase-sqli-bug-claims-sydneys-mathspace-628745&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.itnews.com.au/news/late-patching-of-metabase-sqli-bug-claims-sydneys-mathspace-628745&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Help Net Security, “Mathspace breach exposes data on over a million students and parents”, September 8, 2026. &lt;a href=&quot;https://www.helpnetsecurity.com/2026/09/08/mathspace-data-breach-metabase-vulnerability/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.helpnetsecurity.com/2026/09/08/mathspace-data-breach-metabase-vulnerability/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>Ransomware encrypted the network of a German town&#x27;s water and power utility</title>
    <link href="https://blog.cisent.ai/en/stadtwerke-landsberg-agua-y-luz/"/>
    <id>tag:cisent.ai,2026-09-08:/en/stadtwerke-landsberg-agua-y-luz/</id>
    <published>2026-09-08T12:00:00+00:00</published>
    <updated>2026-09-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Power and water kept flowing, but the utility can&#x27;t rule out that its customers&#x27; data was taken.</summary>
    <content type="html">&lt;p&gt;Overnight on September 1, 2026, an attack encrypted the central IT network of Stadtwerke Landsberg, the municipal utility that supplies electricity and water to Landsberg am Lech, in Bavaria.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The utility disconnected the affected systems from the internet, activated its crisis team and called in outside specialists. Office systems went down and staff could only partly be reached by phone and email, but electricity, water and the other essential services kept running.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;A week later, the utility still couldn’t rule out that the attackers had accessed customers’ names, addresses, phone numbers, email addresses and bank details, and it wasn’t known which group was behind it or whether a ransom had been demanded. In June, a similar attack had hit the utility serving Kamen, Bönen and Bergkamen, in the west of the country.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;The lights staying on is the good news, and it’s rarely luck: it usually means the systems that run the service are kept separate from the office ones.&lt;/p&gt;
&lt;p&gt;But the work only begins once the service is safe. An essential service provider has to explain what happened to its customers, its authorities and its board. In Chile, the law gives it 3 hours for the early warning to ANCI.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt; All of that, with the office network down and the forensic investigation under way.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Report&lt;/strong&gt; gathers and seals the evidence and drafts the report for each audience: the board, the IT team, the insurer and ANCI. The first report is ready in 3 minutes, it works in parallel with containment, not after it, and it keeps the evidence sealed with a hash chain. It’s for any company, whether or not Law 21.663 applies to it.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/report&quot;&gt;Explore Sentinel Report&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;The Record, article of September 8, 2026 on the attack on Stadtwerke Landsberg. &lt;a href=&quot;https://therecord.media/cyberattack-bavaria-germany-utility&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://therecord.media/cyberattack-bavaria-germany-utility&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;Albagli Zaliasnik, “Ley Marco de Ciberseguridad: claves del reglamento de reporte de incidentes” (in Spanish). &lt;a href=&quot;https://www.az.cl/ley-marco-de-ciberseguridad-claves-del-reglamento-de-reporte-de-incidentes/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.az.cl/ley-marco-de-ciberseguridad-claves-del-reglamento-de-reporte-de-incidentes/&lt;/a&gt;&lt;a href=&quot;#fnref4&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>August 2026 was the worst month for ransomware that Chile has on record</title>
    <link href="https://blog.cisent.ai/en/agosto-record-ransomware-chile/"/>
    <id>tag:cisent.ai,2026-09-02:/en/agosto-record-ransomware-chile/</id>
    <published>2026-09-02T12:00:00+00:00</published>
    <updated>2026-09-02T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>At least ten Chilean organizations were hit in a single month, and the most common way in is still a stolen credential.</summary>
    <content type="html">&lt;p&gt;In August 2026, at least ten Chilean organizations fell victim to ransomware: nine appeared on the sites where gangs post the victims who don’t pay, and one more was detected through direct monitoring. It is the highest monthly figure the Chilean firm CronUp has recorded for the country.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The victims work in healthcare, manufacturing, energy, professional services, transportation and technology. The gang with the most victims was TheGentlemen, with five; it was followed by Qilin, with two, and by Majinahanashi, Ms13089 and Direwolf, with one each. Among the cases is the University of Chile Clinical Hospital.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;It isn’t an isolated month. In the first eight months of 2026 there were already more victims than in any previous full year, and this year’s average, about four a month, is four times that of 2023.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;how-they-get-in&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;How they get in&lt;/h2&gt;
&lt;p&gt;The report lists compromised credentials among the most frequent ways in, and its recommendations point the same way: revoke the passwords that turn up in the logs of credential-stealing malware, audit the access of vendors connected to the internal network, and review remote access without a second factor, which it flags as a warning sign before an attack.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;None of those doors is new. What has changed is the speed: a gang that buys a credential or finds an exposed VPN can encrypt a network within hours, and the annual security review arrives months too late.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks an organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does: remote access without a second factor, a vendor with more permissions than it needs, an exposed service nobody remembered. It’s ongoing testing, not one audit a year, and every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;CronUp, “Chile bajo ataque: agosto 2026 rompe todos los récords de ransomware en la historia del país” (in Spanish), updated September 2, 2026. &lt;a href=&quot;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.cronup.com/chile-bajo-ataque-agosto-2026-rompe-todos-los-records-de-ransomware-en-la-historia-del-pais/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>AI has a supply chain too: the LiteLLM case</title>
    <link href="https://blog.cisent.ai/en/litellm-cadena-de-suministro-ia/"/>
    <id>tag:cisent.ai,2026-08-14:/en/litellm-cadena-de-suministro-ia/</id>
    <published>2026-08-14T12:00:00+00:00</published>
    <updated>2026-08-14T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>Chile&#x27;s cyber agency warned some twenty institutions after malicious versions of an open-source AI tool were published.</summary>
    <content type="html">&lt;p&gt;On August 12, 2026, Chile’s National Cybersecurity Agency (ANCI) activated a preventive protocol and warned some twenty Chilean institutions, including government sites, that they could be exposed by the LiteLLM hack. LiteLLM is an open-source gateway that many companies use to connect their applications to different artificial intelligence models.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;According to the account published by BioBioChile, citing the firm Hudson Rock, a group called TeamPCP first compromised the automation of Trivy, an open-source vulnerability scanner. LiteLLM’s developers used Trivy in their continuous integration pipeline, and through it the attackers got the keys to publish versions of LiteLLM with malicious code, which quietly stole secrets from the environments where they were installed.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;ANCI clarified that this was a possible exposure, not confirmed incidents, and sent instructions for checking systems and credentials.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;It’s a supply-chain attack with two hops: a security tool compromised an AI tool, and the AI tool compromised the people who installed it. None of the victims did anything unusual: updating a dependency is exactly what everyone recommends.&lt;/p&gt;
&lt;p&gt;What does leave a trace is the theft itself. To carry off the secrets, the malicious code has to connect outward, to an attacker’s server, from machines that normally don’t talk to anyone new.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Defense&lt;/strong&gt; detects, assesses and blocks attacks alongside the organization’s EDR, its firewall and every source it connects. A build machine that suddenly opens a connection to an unknown destination is an intrusive connection, and Defense blocks it on the spot. Every new attacker technique, like this one, becomes a defense rule.&lt;/p&gt;
&lt;p&gt;And Cisent works with its own models, with no outside AI provider in between.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/defense&quot;&gt;Explore Sentinel Defense&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BioBioChile, “ANCI activa protocolo preventivo y notifica a veintena de instituciones chilenas por hackeo a LiteLLM” (in Spanish), August 14, 2026. &lt;a href=&quot;https://www.biobiochile.cl/noticias/nacional/chile/2026/08/14/anci-activa-protocolo-preventivo-y-notifica-a-veintena-de-instituciones-chilenas-por-hackeo-a-litellm.shtml&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.biobiochile.cl/noticias/nacional/chile/2026/08/14/anci-activa-protocolo-preventivo-y-notifica-a-veintena-de-instituciones-chilenas-por-hackeo-a-litellm.shtml&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
  <entry>
    <title>A flaw in Check Point&#x27;s VPN opened the door to Qilin ransomware</title>
    <link href="https://blog.cisent.ai/en/check-point-vpn-qilin/"/>
    <id>tag:cisent.ai,2026-06-08:/en/check-point-vpn-qilin/</id>
    <published>2026-06-08T12:00:00+00:00</published>
    <updated>2026-06-08T12:00:00+00:00</updated>
    <author><name>Cisent</name></author>
    <category term="amenazas" label="Threat research"/>
    <summary>A vulnerability in legacy configurations let attackers in without credentials, and in at least one case a Qilin affiliate followed.</summary>
    <content type="html">&lt;p&gt;Starting on May 7, 2026, attackers exploited a critical flaw in Check Point’s remote access VPN, CVE-2026-50751, which let them bypass authentication and open a VPN connection without credentials. Exploitation surged in early June and reached a few dozen organizations worldwide.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref1&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;section id=&quot;what-happened&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;What happened&lt;/h2&gt;
&lt;p&gt;The flaw only affects deployments that still use IKEv1, a deprecated key exchange protocol, that accept legacy remote access clients and that don’t require machine certificates. In one confirmed case, the break-in was followed by activity from an affiliate of the Qilin ransomware gang.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref2&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt; Qilin also attacks in Chile: in January and February 2026 it claimed attacks on two Chilean companies, one in agribusiness and one in business services.&lt;a href=&quot;#fn2&quot; class=&quot;footnote-ref&quot; id=&quot;fnref3&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;2&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;p&gt;Check Point released updates for this flaw and for a second one, CVE-2026-50752, which affects site-to-site VPNs. If patching right away isn’t possible, it recommends allowing only IKEv2 for remote access, requiring machine certificates and removing support for legacy clients.&lt;a href=&quot;#fn1&quot; class=&quot;footnote-ref&quot; id=&quot;fnref4&quot; role=&quot;doc-noteref&quot;&gt;&lt;sup&gt;1&lt;/sup&gt;&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;why-it-matters&quot; class=&quot;level2&quot;&gt;
&lt;h2&gt;Why it matters&lt;/h2&gt;
&lt;p&gt;None of those organizations chose to be exposed. The configuration that allowed the attack had probably worked fine for years, and that is why nobody looked at it: an old protocol enabled for compatibility, right on the device that guards the entrance to the network.&lt;/p&gt;
&lt;p&gt;VPNs and firewalls face the internet by design, which is why they are among the doors ransomware gangs look for most.&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;where-cisent-comes-in&quot; class=&quot;level2 b-cierre&quot;&gt;
&lt;h2&gt;Where Cisent comes in&lt;/h2&gt;
&lt;p&gt;&lt;strong&gt;Sentinel Red Team&lt;/strong&gt; attacks an organization’s systems with its authorization and within the scope it defines, to find the gaps before an attacker does, such as a VPN that still accepts legacy configurations. It’s ongoing testing, not one audit a year: what was safe in January may not be safe in May. And every new attack pattern helps train the defense.&lt;/p&gt;
&lt;p&gt;&lt;a class=&quot;k-enlace&quot; href=&quot;https://sentinel.cisent.ai/en/red-team&quot;&gt;Explore Sentinel Red Team&lt;/a&gt;&lt;/p&gt;
&lt;/section&gt;
&lt;section id=&quot;fuentes&quot; class=&quot;b-fuentes&quot; role=&quot;doc-endnotes&quot; aria-labelledby=&quot;fuentes-titulo&quot;&gt;
&lt;h2 id=&quot;fuentes-titulo&quot;&gt;Sources&lt;/h2&gt;
&lt;ol&gt;
&lt;li id=&quot;fn1&quot;&gt;&lt;p&gt;BleepingComputer, “Check Point links VPN zero-day attacks to Qilin ransomware gang”, June 8, 2026. &lt;a href=&quot;https://www.bleepingcomputer.com/news/security/check-point-links-vpn-zero-day-attacks-to-qilin-ransomware-gang/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.bleepingcomputer.com/news/security/check-point-links-vpn-zero-day-attacks-to-qilin-ransomware-gang/&lt;/a&gt;&lt;a href=&quot;#fnref1&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;li id=&quot;fn2&quot;&gt;&lt;p&gt;DeXpose, “Qilin Ransomware Targets Valbifrut in Chile”, January 28, 2026, and “Qilin Ransomware Targets Graneles de Chile”, February 19, 2026. &lt;a href=&quot;https://www.dexpose.io/qilin-ransomware-targets-valbifrut-in-chile/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.dexpose.io/qilin-ransomware-targets-valbifrut-in-chile/&lt;/a&gt; · &lt;a href=&quot;https://www.dexpose.io/qilin-ransomware-targets-graneles-de-chile/&quot; rel=&quot;noopener&quot; target=&quot;_blank&quot; class=&quot;uri&quot;&gt;https://www.dexpose.io/qilin-ransomware-targets-graneles-de-chile/&lt;/a&gt;&lt;a href=&quot;#fnref3&quot; class=&quot;footnote-back&quot; role=&quot;doc-backlink&quot;&gt;↩︎&lt;/a&gt;&lt;/p&gt;&lt;/li&gt;
&lt;/ol&gt;
&lt;/section&gt;
</content>
  </entry>
</feed>
