git-guardrails-claude-code

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely defensive, designed to prevent accidental execution of destructive Git commands. No malicious behavior, obfuscation, or data exfiltration attempts were detected.
  • [COMMAND_EXECUTION]: The skill implements a security hook that executes a local bash script to validate commands before they are run. The installation instructions include making the script executable using 'chmod +x', which is a standard procedure for this functionality.
  • [SAFE]: The modification of Claude Code configuration files (.claude/settings.json) to register the hook is clearly documented and necessary for the guardrail to operate. This behavior is consistent with the skill's primary safety purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 02:40 AM
Security Audit — agent-trust-hub — git-guardrails-claude-code