WASHINGTON - The Federal Bureau of Investigation warned organizations in a Nov. 13 flash advisory highlighted on Nov. 16 that cybercriminals are using the open-source AI tool Artex to compromise networks in under three hours. Designed originally for authorized security testing, the framework links to external language models to automate intrusions across global industries.
Autonomous agents compress defender response windows
Autonomous Agents Compress Defender Response Windows
Although Artex's developer later made the project closed-source, archived copies remain widely accessible online. TechCrunch noted that intrusions can move from initial targeting to active compromise in under three hours.
ConnectWise and South Korean banks hit
ConnectWise and South Korean Banks Hit
The FBI confirmed hackers used Artex to breach IT vendor ConnectWise and access Microsoft 365 environments across at least 2,250 downstream customers, TechCrunch reported. Separately, SC Media cited FBI findings that financially motivated attackers targeted South Korean financial institutions and exfiltrated personal data from approximately 67,000 individuals.
Google's Threat Intelligence Group warned on Nov. 16 that multi-agent offensive frameworks have pushed AI-driven hacking to a critical inflection point.
Security teams urged to harden credentials
Security Teams Urged to Harden Credentials
Federal investigators urged organizations to inventory internet-facing portals, enforce multi-factor authentication, and rotate stale API keys. Cybersecurity firm Barracuda, which also disclosed an Artex-linked breach, warned that defenders relying solely on post-intrusion detection often arrive after attackers finish.


