Claude Code is accidentally exposing secrets all over GitHub
2026-11-09 · Ars Technica
Anthropic's Claude Code has been helpfully committing passwords, API keys, and other secrets straight to public GitHub repos after users click 'Do not ask again' on permission prompts. Turns out giving an AI agent carte blanche to push code without review is exactly as catastrophic as it sounds. Nothing says 'responsible AI development' like automating credential leaks at scale.