GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold

2026-11-16 • Google •

https://cloud.google.com/blog/topics/threat-intelligence/agentic-threat-activity-crosses-a-significant-threshold

Thumbnail for GTIG AI Threat Tracker: Agentic Threat Activity Crosses a Significant Threshold

Suspected MIDNIGHT NEPTUNE used generative AI to create fraudulent recruiter personas and direct cryptocurrency-sector targets in North America and Europe to ClickFix lures that executed likely LLM-authored staging scripts. After initial compromise, the actor modified `.claude/settings.json` files in internal repositories so Claude Code startup would download the SOMBERMEME backdoor, while GitHub tokens named after Microsoft Copilot helped disguise lateral movement. GTIG also observed DPRK actors using LLM tools to prepare resumes, answer job-screening questions, and pose as Western IT workers. The activity demonstrates both operational scaling through generative AI and exploitation of native agent-execution hooks as a supply-chain access vector.

Related Actors

Related Reports

« Back