The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a Node.js script located at ~/.claude/skills/video-effects/run.cjs to process MP4 files, which involves executing shell commands with user-supplied filenames as arguments.
[DYNAMIC_EXECUTION]: The tool relies on a runtime engine (run.cjs) and dynamically loads effect-specific JavaScript files (e.g., 10-frame-break.js) to perform video manipulation.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it processes user-supplied MP4 files which could contain malicious content in metadata or audio tracks.
Ingestion points: Reads a user-provided MP4 file (my-clip.mp4) as primary input for the video effect engine.
Boundary markers: No specific delimiters or safety instructions are provided to the agent regarding the processing of external video data.
Capability inventory: The skill utilizes Node.js, ffmpeg, and Puppeteer, providing a wide range of system and network capabilities during the rendering process.
Sanitization: The instructions do not specify any sanitization or integrity checks for the input MP4 files before they are processed by the rendering engine.

frame-break