sepia-recreate
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill acts as a structural wrapper to invoke a specific operation ('recreate') from a sibling file.
- [SAFE]: The instructions explicitly mandate treating target data as untrusted and restrict the agent from searching for skills outside a specific relative path, following the principle of least privilege.
Audit Metadata