Skip to content- Sources: primary, discussion
- Summary: The FBI published FLASH-20261113-001 on 2026-11-13, a threat overview of ARTEX with IOCs, saying multiple threat actors have used the open-source framework since at least August 2026 against financial, government, technology, defense and healthcare targets. ARTEX is written in Go, drives external LLMs including Claude, GPT, DeepSeek, GLM, Grok and Kimi, and researchers counted over 1,600 exposed ARTEX consoles in early November. The advisory links it to the October ConnectWise Cloud Backup attack attributed to Storm-0940, South Korean bank intrusions combined with Claude Code agents, and a Vapor Panda campaign against more than 200 organizations.
- Why it matters: The FBI calls out partner portals, legacy services, third-party integrations and unthrottled sensitive APIs as the weak spots ARTEX finds, so those are the places to add rate limits and alerts on bursts of automated requests.
- Background: Autonomous penetration testing frameworks hand an operator's goal to LLM agents that run reconnaissance, exploitation, privilege escalation and data theft as a loop, with little human input between steps. ARTEX was posted to GitHub in July 2026 by a developer called Autumn-27 and removed on 2026-10-08 after reports of abuse, but forks, archives and a commercial ARTEX PRO edition remain. This digest covered a similar case on 2026-11-14, when Unit 42 reported China-nexus attackers using the open-source Cairn agent.
- Comments: BleepingComputer adds that researcher MalwareMustDie released an open-source ARTEX detection signature pack and playbook that grades indicators by confidence, and warns that IP indicators may already be stale.