finalrun-update-test
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes specific Model Context Protocol (MCP) tools belonging to the 'final-run' vendor (e.g., update_tests_by_name, list_test_suites) to manage test prompts, which is consistent with the vendor's intended functionality.
- [SAFE]: A mandatory human-in-the-loop review step (Step 4) is required before any update, deletion, or creation of tests is executed, ensuring no automated actions occur without explicit user oversight.
- [SAFE]: Hard rules are enforced to prevent the agent from guessing or inventing sensitive data, such as login credentials, OTPs, or environment-specific identifiers, requiring the agent to stop and ask the user for missing information.
- [SAFE]: The skill maintains clear boundaries by explicitly forbidding the agent from modifying source code, restricting its capabilities to reading code changes and updating natural-language prompts via controlled tool interfaces.
Audit Metadata