self-refine
Pass
Audited by Gen Agent Trust Hub on Jul 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is purely instructional, providing a structured Multi-Aspect evaluation workflow (Evaluate, Refine, Go) to ensure plan quality before execution.
- [SAFE]: No external network requests, file system access, or credential harvesting patterns were identified.
- [SAFE]: No obfuscation, dynamic code execution, or persistence mechanisms are present in the provided files.
- [SAFE]: The skill operates on internal reasoning logic and does not introduce new attack surfaces for command or prompt injection.
Audit Metadata