vehicles-at-speed
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data and external case studies to generate video prompts.\n- Ingestion points: User inputs for character and action in
SKILL.mdand prompt excerpts inreferences/cases.md.\n- Boundary markers: Present inSKILL.mdas a 'reference contract' (Preserve/Replace/Avoid) and a structured 'Output contract'.\n- Capability inventory: Use of image, video, or browser tools for media inspection as defined in theSKILL.mdworkflow.\n- Sanitization: Present as a requirement to extract structural traits and declare specific replacements rather than direct copying.\n- [EXTERNAL_DOWNLOADS]: The skill references multiple external URLs ongoodcase.aiandx.comfor media context. These are used as reference materials for the agent's browsing tools and do not involve executable code.
Audit Metadata