git-guardrails-claude-code
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill installs a shell script intended to monitor and intercept commands executed via the agent's bash tool. The script uses standard tools like
jqandgrepto validate command strings against a blocklist. - [PERSISTENCE]: The skill instructions involve modifying the agent's configuration files (
.claude/settings.jsonor~/.claude/settings.json) to register aPreToolUsehook. While this creates a persistent execution trigger, it is the explicitly stated and benign purpose of the skill to provide safety guardrails.
Audit Metadata