code-refactor
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION] (SAFE): The skill uses framework-specific instructional language (e.g., 'Active Refactoring Principle') to guide AI behavior but does not attempt to bypass safety filters or override core system instructions.- [DATA_EXFILTRATION] (SAFE): No hardcoded credentials, access to sensitive local paths (~/.ssh, ~/.aws), or outbound network requests were found.- [REMOTE_CODE_EXECUTION] (SAFE): There are no patterns involving the download or execution of external scripts (curl|bash) or the installation of untrusted packages.- [INDIRECT_PROMPT_INJECTION] (LOW): The skill possesses a data ingestion surface as it is designed to process and refactor external code files. Mandatory Evidence: (1) Ingestion points: User-provided code files; (2) Boundary markers: Absent; (3) Capability inventory: Code transformation and file structure modification; (4) Sanitization: Not explicitly defined. While the surface exists, the instructions are confined to structural refactoring, minimizing the risk of executing embedded malicious instructions.
Audit Metadata