Malicious GitHub repos trick Claude, Codex and Copilot into silently running a credential stealer
ccerrato147 · x · 2026-11-19
Bitdefender found 22 malicious GitHub repos that weaponize AI coding agents like Claude, Codex and Copilot. Files meant to brief the agent describe the payload as a 'sealed binary artifact' and instruct it to run silently without asking the developer. The stealer harvests cloud and GitHub tokens, SSH keys and password-manager data, then uses stolen tokens to plant the same trap in up to 25 of the victim's own repos—a self-propagating supply-chain attack.
More from coding & agent
- Google Labs launches Stitch Loop, an always-on product agent for post-launch improvement — GoogleLabs · 2026-11-19
- lineup: Open-Source CLI That Catches Coding Agents Making Sneaky Edits — Several-Coyote3534 · 2026-11-19
- Wanted: an LLM Assistant Inside ComfyUI — External MCP Feels Too Clunky — dustynjerman · 2026-11-19
- Running untrusted agents on GKE: sandboxing, cold starts, and state — rseroter · 2026-11-19
- Agentrooms: An MCP Server Letting Claude Code Ask Codex to Review Commits — terzigolu · 2026-11-19
- Jane Street opens first Formal Methods internship to validate agent-generated code — hargup13 · 2026-11-19