music-video-subtitle-generator
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied lyrics, music, and references to generate instructions for subsequent video and music agents. It lacks explicit sanitization or boundary markers to prevent instructions embedded within those lyrics from being obeyed by downstream agents.
- Ingestion points: User-provided lyrics and reference descriptions in SKILL.md and SKILL.cn.md.
- Boundary markers: Absent. While lyrics are split into specific fields (e.g., Rap line:, Typography:), there are no instructions to ignore potential commands embedded within that data.
- Capability inventory: The skill writes output to canvas nodes and invokes various media generation agents (image, video, music, editing) within the MiniMax Hub.
- Sanitization: Absent. No filtering or escaping of user content is performed before interpolation into agent prompts.
Audit Metadata