· ai-devtools

Claude Code "Don't Ask Again" exposes passwords in 1,122 public GitHub copies

Secrets like passwords and API keys, which developers entered into commands they sanctioned within Claude Code, have been discovered residing in publicly accessible GitHub repositories, meaning they…


Secrets like passwords and API keys, which developers entered into commands they sanctioned within Claude Code, have been discovered residing in publicly accessible GitHub repositories, meaning they are viewable by anyone.

Apiiro, a software security firm, reported yesterday that 1,122 public instances of this data were found. This vulnerability arises because when a user chooses the "Yes, and don't ask again" option, Anthropic's AI coding assistant retains the command exactly as entered, including any sensitive credentials, within a project settings file.

Earlier coverage

Related coverage

Also worth reading

  1. How Grab Improved Mobile App Stability by Safely Handling Surge in AI-Generated Code Changes
  2. Open Source pg_vectorsearch Brings More Scalable Vector Search to PostgreSQL
  3. How are you dealing with all the negative side effects of AI on the web?
  4. Meridian browser, Servo and GPUI, experimental web browser release
  5. Runaway AI hacking agent behind WorkOS breach after escaping sandbox