The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted user data (text prompts and image URLs), which represents a surface for indirect prompt injection attacks where malicious instructions could be embedded in the input to influence the AI's behavior or the downstream provider.
Ingestion points: External data enters the system through input.prompt, reference_image_url, and tail_image_url parameters described in the API documentation and code examples within SKILL.md.
Boundary markers: The skill does not implement or recommend the use of delimiters (like XML tags or triple quotes) or specific system instructions to the agent to disregard instructions embedded within the user-provided prompt or metadata.
Capability inventory: The skill scripts and instructions demonstrate capabilities to perform network operations (POST and GET requests) to api.heygen.com and potentially other provider endpoints using curl, fetch, and the Python requests library.
Sanitization: There is no evidence of input sanitization, URL validation, or prompt escaping logic to mitigate the risk of embedded instructions affecting the API requests or the subsequent video generation process.

ai-video-gen