motion-effects
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill mentions an optional helper tool, HyperFrames, which can be installed via npx. The instructions explicitly include a safety guardrail requiring the agent to stop and request user permission before installing any missing dependencies.
- Evidence: Step 5 references 'npx skills add heygen-com/hyperframes'.
- [COMMAND_EXECUTION]: The skill utilizes local system tools to perform media encoding, which is consistent with its stated purpose of creating video and GIF exports.
- Evidence: Step 5 instructs the agent to use local tools for encoding MP4 and GIF files from HTML frames.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (brand configurations and user-provided text) to generate executable HTML/JavaScript files, creating a potential surface for prompt injection.
- Ingestion points: Data is pulled from 'brand.md', 'MOTION.md', and user-supplied labels/numbers in Step 1.
- Boundary markers: None are present to delimit untrusted data from the generation logic.
- Capability inventory: The skill generates an executable 'effect.html' file and invokes shell-based media exporters.
- Sanitization: No explicit sanitization or escaping instructions are provided for user-supplied strings interpolated into the generated script.
Audit Metadata