refactoring-guide
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE] (SAFE): The skill is purely instructional and data-driven, providing guidance on code structure and identifying patterns like 'Long Method' or 'Feature Envy'.
- [PROMPT_INJECTION] (SAFE): No prompt injection or behavior override patterns were detected. The 'Identity' section defines a benign persona (refactoring expert).
- [DATA_EXFILTRATION] (SAFE): No network operations (curl/wget), hardcoded secrets, or access to sensitive local file paths (e.g., ~/.ssh) were found.
- [REMOTE_CODE_EXECUTION] (SAFE): The skill does not include any package managers (npm/pip) or instructions to download and execute remote scripts.
Audit Metadata