motion-effects
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface where user-provided text ('Real content' in Step 1) is interpolated into a generated HTML file (
effect.html). - Ingestion points: User content is gathered in Step 1, Point 2 for use in the motion recipes.
- Boundary markers: There are no explicit delimiters or 'ignore instructions' warnings specified for the interpolation of this user content into the HTML template.
- Capability inventory: The agent is authorized to write local files (
effect.html) and execute shell commands for media conversion (Step 5). - Sanitization: The instructions do not explicitly mandate sanitization or escaping of the user-provided text before it is placed in the HTML structure. However, this is a core functional requirement of the skill's primary purpose (creating UI elements).
- [COMMAND_EXECUTION]: The skill instructs the agent to use local command-line tools to export the generated HTML into MP4 and GIF formats.
- Evidence: Step 5 provides a template for executing conversion commands: 'Use tools already on this computer... Load the page with ?render, call window.seek()... and encode an MP4... Then make a GIF'.
- Mitigation: The skill includes a high-quality safety check: 'If you need something that is not installed, stop and tell me what it is before you install anything,' which prevents unauthorized software installation.
Audit Metadata