claude-code
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill's instructions are meta-level guidance for harness selection and do not attempt to bypass safety filters or perform unauthorized actions.
- [EXTERNAL_DOWNLOADS]: The skill references several external URLs for factual grounding, including official Anthropic documentation and informational blog posts. These are for reference only and do not involve the download or execution of remote code.
- [NO_CODE]: The skill package contains no executable scripts (e.g., .py, .js, .sh) or binary files, relying solely on markdown instructions and structured evaluation metadata.
Audit Metadata