verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a set of operational guidelines for AI agents to ensure tasks are verified through actual command execution rather than assumptions.
- [COMMAND_EXECUTION]: The skill lists standard development commands for verification (e.g.,
npm test,pytest,cargo audit,bandit) and workspace cleanup (rm -f). These commands are contextually appropriate for a software development assistant and do not represent unauthorized or malicious execution patterns. - [DATA_EXFILTRATION]: No exfiltration patterns were detected. The skill specifically instructs the agent to delete temporary logs and stack traces, which is a positive practice for maintaining workspace hygiene and preventing context inflation.
Audit Metadata