FBI Warns ARTEX AI Agents Reach Admin Access in 3 Hours

The FBI has warned that hackers are using an artificial intelligence agent framework called ARTEX to automate network break-ins, with some intrusions reaching privileged access in under three hours.
The warning came in a FLASH alert published on 13 November 2026, titled "Threat Overview of ARTEX Autonomous Penetration Testing Framework" and issued through the FBI's Internet Crime Complaint Center. The alert, reference FLASH-20261113-001, is marked TLP:CLEAR, meaning it may be shared without restriction.
ARTEX is an open-source, large language model driven penetration testing platform written in Go. A developer using the name "Autumn-27" released it on GitHub in July 2026. It takes a goal written in plain English and hands it to a team of AI agents, which plan and carry out the reconnaissance, exploitation, privilege escalation and data collection largely on their own.
A testing tool turned into a weapon
The FBI said a variety of malicious actors have used ARTEX since at least August 2026 against organisations around the world, in financial services, government, technology, defence, semiconductors, energy, healthcare, manufacturing, telecommunications and legal services.
The framework does not rely on a single new exploit. Instead, it chains together familiar weaknesses: stolen credentials, exposed applications, SQL injection, missing API authentication, exposed passwords and abuse of trusted cloud connections. Citing Microsoft, the FBI said some intrusions progressed from initial targeting to privileged access in under three hours, leaving defenders little time to coordinate a response even when alerts fire.
Microsoft previously linked ARTEX to Storm-0940, a China-state-aligned group that targeted Microsoft 365 environments through compromised ConnectWise application credentials, attempting to reach more than 2,250 tenants using permissions customers had already approved. A separate campaign tracked as Storm-3358 was still active as of 6 November. A financially motivated actor used ARTEX against South Korean financial institutions and stole personal information on about 67,000 people.
Still spreading after the takedown
The original developer removed the ARTEX code from GitHub on 8 October 2026, but the FBI said the tool keeps circulating through community forks, archived copies and a commercial edition called ARTEX PRO. Researchers counted more than 1,600 publicly exposed ARTEX consoles in early November, although an exposed installation does not by itself prove its operator is using it maliciously.
The FBI said ARTEX also targets overlooked supplier portals, sales tools and mobile backends, the peripheral systems that often sit outside an organisation's main defences. It lets attackers with minimal technical expertise run complex, multi-target operations at the same time, the alert said.
What defenders should do
The FBI recommends restricting exposed administration tools, reducing application permissions, and scanning code repositories for secrets. The alert publishes indicators of compromise that defenders can search for, but the FBI asks organisations to vet those indicators before blocking on them, since some, such as IP addresses, may not prove a compromise on their own.
The alert draws on published research from Akamai, Microsoft, Google's threat intelligence group and Mandiant, Censys, CrowdStrike, Zscaler ThreatLabz, CERT-EU and Check Point Research. It is the clearest signal yet that US law enforcement views AI agent hacking frameworks as a mainstream threat rather than a research curiosity. The full advisory is published by the FBI as FLASH-20261113-001, and CybersecurityNews has a detailed account of the alert.