scenario-asset-analysis

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill describes tools like asset_analyze and asset_detect that ingest external data such as images and text inputs. This creates a surface where malicious instructions embedded in processed assets could potentially influence agent behavior.\n
  • Ingestion points: The images and text_inputs parameters in asset_analyze (SKILL.md).\n
  • Boundary markers: The instructions do not specify any delimiters or safety warnings for the interpolated data.\n
  • Capability inventory: The skill interacts with tools for asset management (asset_add_tags, collection_add_assets) and retrieval (asset_download).\n
  • Sanitization: There is no mention of filtering or sanitizing the content of the analyzed assets.\n- [EXTERNAL_DOWNLOADS]: The documentation references the installation of sibling skills using npx skills add scenario-labs/skills. This involves downloading code from the vendor's own repository for tool extension.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 07:46 AM
Security Audit — agent-trust-hub — scenario-asset-analysis