heygen-avatar-video
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by processing natural language instructions in scripts and prompts, as well as ingesting data from external API responses.
- Ingestion points: User-provided text for the
scriptparameter in Direct Video requests and natural language prompts for the Video Agent route inSKILL.md. - Boundary markers: Data is structured within JSON payloads, but the content of the script and prompt fields is treated as raw natural language without specific escaping or delimiter instructions.
- Capability inventory: The skill uses network capabilities to interact with HeyGen's API endpoints (
POST /v3/videos,POST /v3/video-agents, etc.) and handles the download of generated video artifacts. - Sanitization: The skill includes proactive safety guidance, such as requiring a 'Profanity/safety filter before appending text' and a 'QA review rubric' to verify script accuracy and compliance.
- [SAFE]: The skill facilitates interactions with official HeyGen service domains (
heygen.com,developers.heygen.com). These references are documented as standard operational requirements for the intended service and do not represent a security risk.
Audit Metadata