minimax-h3-reference-to-video
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were detected. The skill functions as a structured guide for prompt engineering, instructing the agent to create an 'asset ledger' and a text-based prompt for a video generation model. It contains explicit safety constraints, such as the command to 'stay prompt-only and never submit a generation job.' While the skill processes descriptions of user-supplied media (an indirect prompt injection surface), it mitigates this by requiring explicit boundary markers, bounded language for asset roles, and the use of a canonical template to isolate inputs. No patterns related to credential theft, remote code execution, or data exfiltration were found.
Audit Metadata