seedance-2-5
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists entirely of instructional content and prompt templates for creative video generation. It does not include scripts, shell commands, or network operations beyond descriptions of tool-specific API parameters.
- [INDIRECT_PROMPT_INJECTION]: The skill describes an architecture for processing external media URLs (images, video, audio) as references for generation. While this presents an ingestion surface for untrusted data, it is a standard functional requirement for the video generation tools it supports.
- Ingestion points: Media reference URLs in the tool call parameters documented in
SKILL.md. - Boundary markers: Employs a highly structured "Section order" prompting method to clearly separate different types of instructions.
- Capability inventory: Interaction with third-party video generation tools (
seedance_video,seedance_ark,runway_video,comfyui_video). - Sanitization: No specific sanitization logic is present in the markdown, as it relies on the underlying platform's tool implementation.
Audit Metadata