video-add-motion-graphics
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions define an workflow that processes external content, which could serve as a vector for indirect prompt injection.
- Ingestion points: Transcript data and external images retrieved from the web enter the agent context during the workflow (Step 2 and Step 7 in SKILL.md).
- Boundary markers: Instructions do not explicitly define boundary markers or 'ignore' instructions for the ingested transcript or web-sourced content.
- Capability inventory: The skill has high-tier capabilities, including executing Node.js scripts for recipe materialization and Python scripts for rendering the final project (SKILL.md).
- Sanitization: No sanitization or filtering logic for external data is described in the provided skill files.
- [DYNAMIC_EXECUTION]: The skill uses a custom JavaScript execution environment to render animations.
- The rendering process uses a deterministic adapter (hf-adapter.js) to run animation logic (hf-recipe.js) by mocking standard browser timing and animation APIs.
- Animation parameters and layouts are adapted by the agent at runtime to create project-specific assets (SKILL.md Step 7).
- [EXTERNAL_DOWNLOADS]: The skill requires downloading external assets during execution.
- The workflow directs the agent to search for and download representative images from online sources to populate recipe image slots (SKILL.md Step 7).
Audit Metadata