maintain-verification-skill
Pass
Audited by Gen Agent Trust Hub on Jul 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill functions as a maintenance loop for local verification tools. It limits its activities to reading project source code and executing local harness scripts within the project's own directory structure, which is consistent with standard developer workflows.\n- [COMMAND_EXECUTION]: The execution of commands is confined to existing project harness scripts during the verification pass, with explicit instructions regarding cleanup and process health checks to prevent residue or unintended persistence.\n- [PROMPT_INJECTION]: While the skill ingests source code that could contain indirect prompt injections, it employs a scoped architecture where data-reading subagents are read-only and prohibited from executing commands or modifying files.
Audit Metadata