create-video
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts which are then sent to the HeyGen Video Agent API. The prompt-optimizer.md file notes that the downstream system functions as an HTML interpreter, creating a surface for instructions within user data to potentially influence the generation process. * Ingestion points: User prompts in SKILL.md and references/video-agent.md. * Boundary markers: The instructions do not define delimiters or ignore warnings to separate user content from system instructions. * Capability inventory: The skill can read and write local files as shown in references/assets.md and references/video-status.md, and perform network operations via references/video-agent.md. * Sanitization: No sanitization or escaping of the user prompt is prescribed before sending it to the HeyGen API.
- [COMMAND_EXECUTION]: The documentation provides code examples in Python and TypeScript that utilize filesystem APIs like fs.readFileSync and fs.writeFileSync to manage video assets and output, providing a framework for the agent to access the host file system.
- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading generated video files and thumbnails from heygen.ai and heygen.com, which are recognized as the official domains of the service provider.
Audit Metadata