The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the execution of the runcomfy CLI tool to process video generation requests.
[EXTERNAL_DOWNLOADS]: The documentation guides users to install the @runcomfy/cli package from the npm registry and references external image/video URLs for processing.
[PROMPT_INJECTION]: The skill documents an attack surface for indirect prompt injection through user-provided prompts and external media URLs.
Ingestion points: Processes prompt, image_url, video_url, and audio_url fields defined in SKILL.md.
Boundary markers: Inputs are passed as JSON strings to the CLI to prevent shell expansion.
Capability inventory: Invokes the runcomfy CLI, which performs network requests to model-api.runcomfy.net and writes generated files to a local directory.
Sanitization: Uses JSON encoding for prompt transmission to mitigate command injection risks.

seedance-v2