scenario-seedance

Pass

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external assets such as images, videos, and audio files that could potentially contain hidden instructions designed to influence the behavior of the AI model.
  • Ingestion points: Media is ingested through parameters like image, lastFrameImage, referenceImages, referenceVideos, and referenceAudio in the model_run and upload_asset tool workflows.
  • Boundary markers: The instructions do not specify any delimiters or explicit system instructions to the underlying model to ignore potential prompt injections embedded within the provided media assets.
  • Capability inventory: The skill utilizes tools for asset uploading (upload_asset), model execution (model_run), and result viewing (asset_display).
  • Sanitization: No methods for sanitizing, validating, or filtering the content of the uploaded media files are described in the instructions.
  • [EXTERNAL_DOWNLOADS]: The skill instructions advise users to install additional sibling skills using npx skills add scenario-labs/skills. This is a reference to resources provided by the vendor (scenario-labs) and is part of the intended setup process.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 25, 2026, 02:37 PM
Security Audit — agent-trust-hub — scenario-seedance