FBI Warns of ARTEX AI Penetration Tool Used in Attacks Across 14 Countries
·2026.11.16 08:34
Key point
The FBI issued an advisory on November 13 warning that the open-source AI penetration-testing platform ARTEX is being used by threat groups to breach organizations in 14 countries, with some intrusions achieving privileged access in under three hours.
Details
- FBI FLASH-20261113-001 identifies ARTEX, an autonomous AI penetration-testing platform, as a tool abused by multiple threat groups since August 2026 to target organizations in 10 industries, including financial services, government, and defense.
- ARTEX connects to external LLMs (notably DeepSeek, Claude, GPT) to automate reconnaissance and exploitation, reducing the time from initial targeting to privileged access to under three hours in some cases.
- Confirmed attacks include a supply-chain compromise of ConnectWise Cloud Backup affecting 162 Microsoft 365 tenants, theft of personal data from approximately 67,000 people in South Korean financial institutions, and a Vapor Panda campaign targeting over 200 organizations in 14 countries.
- Despite the original GitHub repository being removed on October 8, ARTEX continues to circulate via forks and a commercial version, with over 1,600 consoles exposed on the internet as of early November.
- Detection resources include the MMD-072-2026 signature pack released by MalwareMustDie, while the FBI recommends inventorying internet-facing applications and monitoring for unusual automated request surges.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.