jianying-motion
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill enforces environment validation by requiring version and capability checks (version, doctor, capabilities) before any editing operations are performed. This prevents the agent from attempting unsupported or risky commands on the host system.
- [SAFE]: Data integrity is prioritized through an isolated write policy. The skill explicitly forbids overwriting source drafts, requiring instead that all operations occur on isolated copies with recorded input hashes.
- [SAFE]: Input validation is strictly defined in Step 3, which mandates probing media, drafts, and task states to reject non-existent paths, out-of-bounds time values, and unknown schemas.
- [SAFE]: The skill includes an explicit 'Boundary Refusal' guide that instructs the agent to deny dangerous user requests, such as automatic retries on ambiguous failures or bypassing verification gates.
- [SAFE]: Credential safety is maintained by rules that prohibit the collection, uploading, or logging of unrelated user credentials, limiting data processing to declared local and provider boundaries.
Audit Metadata