Skip to content
AI Hack Log

Google: attackers run agent frameworks in live intrusions

Google's threat intelligence group says actors use ARTEX, CAIRN and PentAGI in intrusions, and found an agent that turns patches into exploits unaided.

Published
Type
Brief
Topics
AI-orchestrated intrusion, LLMjacking, Gemini

Confirmed

Stated by the company or authority involved in its own publication. What the levels mean

On 16 November 2026 Google Threat Intelligence Group (GTIG) and Mandiant published their latest AI threat tracker. According to Google, threat actors have moved beyond manual prompting to open-source multi-agent offensive frameworks, including ARTEX, CAIRN, CyberStrikeAI and PentAGI, to run multi-stage intrusions faster.

GTIG says it independently observed CAIRN and PentAGI in the wild. It reports CAIRN used against targets that include the US aerospace sector, and a customised platform built on PentAGI used between late August and early October 2026 for reconnaissance and scanning against organisations in Asia. It saw experiments with CyberStrikeAI and HexStrike AI, and evidence that a PRC-nexus group was building its own framework. Google notes these tools trade stealth for speed, leaving forensic traces.

The report also says GTIG identified, for the first time, actors using LLM frameworks to automate vulnerability discovery and exploitation with no human in the loop. On an exposed staging server it found a framework called exploit_agent that compares software before and after patches, drafts exploits and runs them automatically. Victim telemetry on the server suggested interest in French media, entertainment and education websites.

Other findings: LLMJacking, where stolen LLM credentials are resold through proxy services, continues. Google also says zero-day exploitation rose from an average of 8 a month in 2025 to 11 in 2026, spiking to 22 in August.

Amazon’s separate ARTEX report describes the same tooling from decoy-server data, and says it could not verify the outcomes the agents reported. Google’s report names no victim organisations and gives no breach counts, so the scale of successful intrusions is unknown.

Sources

Publisher Document and what it supports Kind Date Links
Google Threat Intelligence Group GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold Vendor 16 Nov 2026
Amazon Web Services Amazon Threat Intelligence observes threat actors running cybercrime campaigns with ARTEX AI agents Vendor 13 Nov 2026

Related news

How this was made

Researched and drafted by AI agents from the sources above. Report an error. Read the editorial policy and the methodology.