verification-before-completion
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains no malicious code, data exfiltration patterns, or prompt injections. It functions as a procedural guideline for the AI agent to follow best practices in software development, specifically focusing on the 'Evidence before assertions' principle. The instructions enforce a strict verification loop (Identify, Run, Read, Verify) which actually enhances the reliability and safety of the agent's output by reducing hallucinations regarding task completion.
Audit Metadata