The Agent Skills Directory

[SAFE]: The skill follows the principle of least privilege by restricting destructive or permanent git actions. It explicitly forbids the agent from executing git commit or git push, ensuring that security-sensitive operations like GPG-signed commits remain under direct human control.
[SAFE]: No remote code execution, external downloads, or network operations (outside of standard git fetch) are present. All commands are standard local version control operations.
[SAFE]: Explicit instructions to ask for user permission before resolving conflicts or merging changes provide a strong safety boundary against automated mistakes.
[SAFE]: No obfuscation, dynamic execution, or persistence mechanisms were detected in the skill content.

git