Threat researchCisent
With an AI agent, INC and The Gentlemen affiliates got in through FortiClient EMS and SimpleHelp
Their exposed servers revealed an AI agent's sessions, the encryptors ready to launch and the flaws they used to get in.

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.1
How they worked
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.1
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%.1 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.1
How they got in
Through devices exposed to the internet. The figures come from the attackers’ own records and haven’t been independently verified:1
- FortiClient EMS (CVE-2026-35616): they claim to have targeted 2,236 servers, created 108 administrator accounts and reached 4,140 managed endpoints.1
- SimpleHelp (CVE-2024-57727): they claim that out of 4,922 servers scanned they pulled 427 administrator passwords.1
- A hosting control panel (CVE-2026-41940): they claim to have breached more than 49 hosting providers.1
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 7.4.6.2 CISA added it to its catalog of exploited vulnerabilities on April 6.3
Why it matters
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.
Where Cisent comes in
Sentinel Red Team 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.
Sources
Hackread, “INC and The Gentlemen Ransomware Affiliates Expose 8,257 AI Messages”, October 21, 2026, on SOCRadar’s October 20 report. https://hackread.com/inc-the-gentlemen-ransomware-affiliates-ai-messages/↩︎
runZero, “Fortinet FortiClient EMS vulnerability: CVE-2026-35616”, April 2026, citing Fortinet’s advisory FG-IR-26-099. https://www.runzero.com/blog/fortinet-forticlient-ems/↩︎
CISA, “CISA Adds One Known Exploited Vulnerability to Catalog”, April 6, 2026. https://www.cisa.gov/news-events/alerts/2026/04/06/cisa-adds-one-known-exploited-vulnerability-catalog↩︎