avatar-video
Fail
Audited by Gen Agent Trust Hub on Aug 1, 2026
Risk Level: CRITICAL
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation references official HeyGen API endpoints (api.heygen.com, upload.heygen.com) and asset delivery domains (files.heygen.ai, resource2.heygen.ai). These are categorized as well-known and legitimate service providers for AI video generation.\n- [DATA_EXFILTRATION]: Code examples demonstrate reading local files using the
fsmodule to upload binary data to HeyGen. This functionality is essential for providing custom backgrounds and photo-based avatars, but it establishes a surface for data exposure if the agent is instructed to read sensitive files.\n- [COMMAND_EXECUTION]: The skill contains TypeScript and Python snippets using standard system libraries (fs, path, os, requests) to facilitate API interactions. These are routine development patterns for integrating third-party services.\n- [PROMPT_INJECTION]: The skill processes external script content and asset URLs, representing an indirect prompt injection surface. The documentation includes a mitigation step by enforcing HTTPS for remote asset fetching, which addresses some common server-side request forgery risks.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
Audit Metadata