scenario-image
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions include a command to install additional components using
npx skills add scenario-labs/skills. This refers to the vendor's official package repository. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied prompts and images for generative tasks, which is an inherent attack surface for indirect instructions.
- Ingestion points: Processes user-provided text prompts and reference images uploaded via
upload_asset. - Boundary markers: The instructions emphasize strict adherence to model-specific schemas obtained via
model_schema_get. - Capability inventory: Includes capabilities for file downloads (
asset_download), asset uploads, and model execution tools (model_run). - Sanitization: Instructions mandate validating parameters against the model contract (
model_schema_get) before every run to ensure prompt limits and field names are respected.
Audit Metadata