craft-skill
Pass
Audited by Gen Agent Trust Hub on Jul 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and reference documents (diagnosis and evaluation guides) focus on improving agent performance through evidence-backed development. No malicious instructions or injection patterns are present.
- [COMMAND_EXECUTION]: The file
scripts/eval.pyis a Python utility for validating evaluation data and performing statistical analysis. It uses standard library modules for local file management and numeric calculations, with appropriate checks to prevent accidental inclusion of answer keys in evaluation packs. - [SAFE]: No network activity, hardcoded credentials, or obfuscated content were found across the skill files.
Audit Metadata