The Agent Skills Directory

[SAFE]: The skill provides a defensive guardrail by intercepting and blocking high-risk git commands such as git push --force, git reset --hard, and git clean -fd before they are executed.
[SAFE]: The logic is implemented in a local bash script (scripts/block-dangerous-git.sh) that uses standard, well-known tools like jq and grep to process input. The script performs no network operations or unauthorized file access.
[SAFE]: The installation process involves standard configuration steps for the Claude Code platform, including setting file permissions with chmod +x and updating ~/.claude/settings.json. These actions are clearly documented and are necessary for the skill's intended safety function.
[SAFE]: No obfuscation, data exfiltration, or malicious persistence mechanisms were detected. The skill's behavior matches its stated purpose of improving repository safety.

git-guardrails-claude-code