The Agent Skills Directory

[SAFE]: No malicious behavior, obfuscation, or dangerous execution patterns were identified across the skill's instructions and reference files.
[DATA_EXFILTRATION]: The skill documents legitimate functionality for uploading local assets (images, audio, and video) to HeyGen's infrastructure. While this involves reading local files and transmitting them over the network, it targets a well-known service provider and is essential for the skill's purpose.
Evidence: Code examples in references/assets.md and references/photo-avatars.md show binary file reads and HTTP POST requests to upload.heygen.com.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied scripts for video generation, creating a surface for indirect instructions to be passed to the HeyGen API.
Ingestion points: The input_text parameter in references/video-generation.md and references/scripts.md.
Boundary markers: Not explicitly defined; scripts are interpolated directly into JSON request bodies.
Capability inventory: Network communication with HeyGen API endpoints (api.heygen.com).
Sanitization: Not specified; the skill relies on the service provider's internal safety filtering.

avatar-video