Logo Animation
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied logo files and uses them as inputs for media generation and editing tools, which creates an entry point for untrusted data.
- Ingestion points: Original logo files provided by the user for processing (SKILL.md).
- Boundary markers: The instructions explicitly require preserving the "exact logo" and "exact mark," and tell the agent to request a cleaner source if the mark cannot be properly inspected.
- Capability inventory: The skill has access to image generation, video generation, and video editing capabilities (
media.generate_image,media.generate_video,media.edit_video). - Sanitization: The workflow includes a mandatory quality assurance step to "inspect the mark frame-by-frame for letter changes, melting, extra symbols, reflections, cropping, flicker, and end-card legibility" to ensure output fidelity.
Audit Metadata