cs-refactor-ff
Pass
Audited by Gen Agent Trust Hub on May 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides structured instructions for a specific development workflow (refactoring) and does not contain any malicious patterns, obfuscation, or unauthorized data access attempts.
- [COMMAND_EXECUTION]: The instructions direct the agent to run tests (unit, integration, type checks, lint) and use
grepto verify code changes. These are standard operations for a refactoring agent and are used for local self-verification. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied code, which is a potential ingestion point for indirect instructions. However, it implements mandatory boundary checks (verifying if the behavior remains unchanged) and has a low capability tier for this specific 'fastforward' flow, significantly limiting the impact of any potential injection.
- [DATA_EXPOSURE_AND_EXFILTRATION]: There are no network operations or access to sensitive file paths (e.g., credentials, ssh keys) defined in the instructions.
Audit Metadata