Latch
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a legitimate tool for configuring Claude Code hooks and follows established safety patterns.
- [PROMPT_INJECTION]: No evidence of malicious prompt injection was found. The skill includes instructions to implement safety-focused prompt hooks rather than bypassing agent constraints.
- [DATA_EXFILTRATION]: No exfiltration patterns detected. The skill focuses on local configuration management and includes explicit recipes (e.g., Recipe S3) to prevent data leaks.
- [COMMAND_EXECUTION]: Command execution is limited to managing and testing local hook scripts with user-defined logic, supported by robust debugging documentation and safe shell practices.
- [REMOTE_CODE_EXECUTION]: No remote code execution or untrusted downloads were found.
Audit Metadata