heygen
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions are focused on the creative process and ethical deployment of AI avatars. It implements a defensive posture by requiring consent verification and mandatory AI disclosure for all generated content.
- [SAFE]: References to external services are limited to well-known platforms including HeyGen (
api.heygen.com), WoopSocial, and Google's Veo. These are documented for their primary purpose of video rendering and scheduling. - [SAFE]: The skill does not contain any executable scripts, remote code download patterns, or persistence mechanisms. It acts as an instructional guide for the agent to prepare briefs for external rendering.
- [SAFE]: No obfuscation, data exfiltration, or prompt injection patterns were identified. The skill includes specific evaluation cases to ensure the agent pushes back against non-consensual likeness requests.
Audit Metadata