evidence-before-done
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a purely instructional defensive guard intended to enforce verification of tasks via existing project scripts or commands.
- [PROMPT_INJECTION]: No malicious instructions designed to bypass safety filters or override system behavior were detected.
- [DATA_EXFILTRATION]: No unauthorized access to sensitive files or exfiltration to external domains was found.
- [REMOTE_CODE_EXECUTION]: No patterns involving unauthorized remote script execution or malicious package installation were identified.
- [COMMAND_EXECUTION]: Command execution is limited to the agent's legitimate task of running existing project-level checks and tests.
Audit Metadata