avatar-video
Fail
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: CRITICALINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of user-generated scripts and media assets which are then processed by an external API. This creates a surface for indirect prompt injection where malicious input could attempt to influence the agent's behavior during subsequent output analysis or manipulate the third-party generation process.
- Ingestion points: Untrusted data enters the context via the
input_textparameter in video generation requests and user-provided asset uploads as described inSKILL.mdandreferences/video-generation.md. - Boundary markers: Absent; the instructions do not require the use of delimiters or 'ignore embedded instruction' warnings when identifying or interpolating user-controlled content into API calls.
- Capability inventory: The skill enables network operations via HTTP POST requests to
api.heygen.comandupload.heygen.comusing the user's API key. - Sanitization: Absent; there are no requirements for validating, escaping, or filtering the scripts or metadata provided by the user before transmission.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
Audit Metadata