The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied text to generate video content using external AI models. This creates a potential surface for indirect prompt injection if the inputs contain instructions intended to subvert the video generation process.
Ingestion points: Text prompts are ingested via the prompt field in the mcp__heygen__generate_video_agent tool and direct API endpoints in SKILL.md.
Boundary markers: The instructions do not define specific delimiters to isolate user-provided prompts from the agent's internal logic or the API request structure.
Capability inventory: The skill allows the agent to make network requests to HeyGen's API and read local files for asset uploads.
Sanitization: No explicit sanitization logic is provided to filter or escape the content of user-generated prompts.
[EXTERNAL_DOWNLOADS]: The skill documents procedures for downloading generated video assets from official vendor domains, which is an expected part of the tool's core functionality.
Evidence: The file references/video-status.md provides code snippets in TypeScript and Python for fetching video files from heygen.ai subdomains and saving them to the local filesystem.

create-video