The Agent Skills Directory

[DYNAMIC_EXECUTION]: The src/kit.js file employs eval() to process animation and camera expressions passed from the command line. This allows for dynamic calculations during the rendering process within the controlled browser environment.
[COMMAND_EXECUTION]: The rendering script render.mjs executes ffmpeg via a subprocess to handle video encoding and audio muxing. The project initialization script also executes npm to install necessary Node.js dependencies.
[EXTERNAL_DOWNLOADS]: The skill depends on external resources from established providers, including Node.js packages from the NPM registry and CJK font assets from Google Fonts (fonts.googleapis.com).
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface where untrusted data could influence agent behavior.
Ingestion points: Command-line arguments such as --crop-at in render.mjs flow into the rendering logic.
Boundary markers: No specific delimiters or safety warnings are present to sanitize these inputs before evaluation.
Capability inventory: The skill can execute shell commands (ffmpeg, npm) and write to the local filesystem.
Sanitization: Input strings are evaluated as code without explicit filtering, relying on the browser's sandbox for isolation.

paper-cutout