verification-loop
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes standard development and build tools such as npm, pnpm, npx (for tsc), pyright, and ruff. It also utilizes system utilities like grep and git to verify project state. These operations are restricted to the local environment and align with the skill's stated purpose of providing a verification system for code changes.
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface through its processing of local project data:
- Ingestion points: Reads project source code via grep in Phase 5 and version control metadata via git diff in Phase 6.
- Boundary markers: The instructions do not define explicit delimiters or 'ignore' instructions for the data ingested from files.
- Capability inventory: The agent maintains shell access for running various development tool chains (SKILL.md).
- Sanitization: There is no explicit sanitization or filtering of the file contents or diff outputs before they are processed by the agent.
Audit Metadata