The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data including user-defined prompts and remote image/video URLs. This creates a surface for indirect prompt injection where instructions embedded in external metadata or prompts could attempt to influence the agent.
Ingestion points: The API parameters prompt, start_image_url, end_image_url, video_url, and callback_url allow external data to enter the agent's operational context.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings to wrap these inputs.
Capability inventory: The skill performs network operations via curl to the api.acedata.cloud domain.
Sanitization: There is no evidence of input validation or sanitization for the provided URLs or prompt strings within the skill definition.

luma-video