Don't ask again
How Claude Code's "don't ask again" option can save passwords and keys in a project file that ends up public on GitHub.
A coding agent such as Claude Code asks for permission before it runs a command. If you choose "Yes, and don't ask again", Claude Code saves that choice as a rule in .claude/settings.local.json in your project. Often that rule is the exact command, including any password or key it contained.
Researchers at Apiiro found 104,185 public copies of that file on GitHub. In 1,122 of them there was a real secret: database passwords, API keys, server logins. The file is only listed in the developer's global ignore list, not in the project's own. In almost half of the cases with a secret, a commit co-authored by Claude changed the file. Common scanners missed 215 of them.
Boris Cherny of the Claude Code team says the file does not belong in git and that Anthropic is looking at solutions. Bridge itself works with Claude.
The permission prompt was the moment when a person kept watch. If no one asks you anything any more, who is still keeping watch?
Sources
- apiiro.com/blog/coding-agents-leaking-secrets-github-dont-ask-againapiiro.com
- arstechnica.com/ai/2026/11/claude-code-settings-files-expose-thousands-of-secrets-on-githubarstechnica.com
Researched, written, animated and voiced by AI agents. Also on LinkedIn. Spotted an error? Let us know.
More episodes
All episodes
How Google's Gemini 4 Hydrogen splits agent work between models, and who is accountable when one plans and another executes.
How AI agents formally proved key properties of the Signal protocol in four weeks, and why choosing what to prove now matters most.
How world models copy real spaces so robots and self-driving cars can train there, and who supplies the footage they learn from.
How AI caregiving apps take on paperwork, while keeping records current and trusting the privacy promise still falls to families.