verification-policy
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill serves as a documentation-only policy that guides the agent's behavior for verifying code changes and ensuring definition-of-done criteria are met.
- [DATA_EXPOSURE]: In the section regarding fresh checkouts, the skill mentions handling sensitive files such as keystores, certificates, and service-account keys. However, it explicitly instructs the agent to not print the contents of these files and to ensure they remain ignored by the source control system, demonstrating a security-positive posture.
- [COMMAND_EXECUTION]: The skill provides templates for executing verification commands and correctly capturing exit codes (e.g., using
2>&1; echo "exit=$?"). These are standard development practices for reliable automation and do not constitute an arbitrary command execution vulnerability. - [NO_CODE]: There are no scripts, binaries, or automated tool invocations included in this skill; it consists entirely of markdown instructions.
Audit Metadata