verification-loop
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION] (SAFE): The skill invokes standard local development utilities including npm, pnpm, npx, tsc, pyright, ruff, and git. These commands are necessary for code verification and are executed within the local project context.\n- [DATA_EXPOSURE] (SAFE): Includes a defensive scan phase using grep to search for patterns like 'sk-' or 'api_key'. This is a security-positive feature intended to prevent the accidental commitment of credentials.\n- [INDIRECT_PROMPT_INJECTION] (SAFE): The skill processes outputs from build and test tools. While it does not use explicit boundary markers to wrap this data, the ingestion of tool output from the user's own codebase is a standard and low-risk operation for a developer tool.
Audit Metadata