FULL STORY

ARTEX: From Exposed Consoles to an FBI Flash Alert

Over 1,300 exposed AI pentest consoles were found exploited in the wild, the red-team tool Artex was caught shipping a backdoored Docker image, and by mid-November the FBI issued a FLASH alert warning that ARTEX had become a staple for attackers worldwide.

2026-11-03 ~ 2026-11-18 · 4 episodes · 12 posts

Episode 1 · Over 1,300 AI Penetration Testing Consoles Exposed and Exploited by Attackers (2026-11-03, 3 posts)

Over 1,300 exposed consoles of open-source AI agent frameworks like Cairn, ARTEX and PentAGI are being leveraged for real intrusions by groups such as Storm-0940. Researchers warn these LLM-driven frameworks can automate most stages of real attacks, removing human limits on attack volume.

Episode 2 · Baidu Agent+ champion tool Artex found backdoored in Docker image (2026-11-04, 2 posts)

Security researcher Graham Helton found that Artex, the red-team AI tool that won Baidu's Agent+ challenge, ships with an unauthorized backdoor in its Docker image, distributed via a 1,325-star GitHub fork, potentially letting attackers take over users' shells.

Episode 3 · FBI Warns Open-Source AI Pentest Framework ARTEX Is Being Used in Automated Attacks Worldwide (2026-11-16, 5 posts)

The FBI issued a FLASH alert on Nov 13, 2026 warning that attackers are using the open-source agentic AI pentest framework ARTEX for automated attacks worldwide, linked to the ConnectWise supply-chain intrusion with victims across government, finance, telecom and other sectors.

Episode 4 · ARTEX Open-Source Attack Framework Becomes Attacker Standard Within Months (2026-11-17, 2 posts)

Security researcher Florian Roth warns that the open-source ARTEX framework has become a staple for attackers, with 2700+ IoCs and over 2400 exposed instances, and that such offensive AI frameworks cannot be detected via IoCs and should not be treated as conventional malware.