The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates a complete HTML and JavaScript application at runtime and subsequently executes it using a headless browser via the render.mjs script to capture frames for video production.
[COMMAND_EXECUTION]: The workflow requires the agent to execute multiple shell commands, including npm install, and the render.mjs script uses child_process.spawn to run ffmpeg for media encoding.
[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect injection.
Ingestion points: Content is fetched from user-provided URLs in Step 1 to extract product details, features, and brand information.
Boundary markers: None identified. The instructions do not specify any delimiters or safety warnings for the fetched content.
Capability inventory: The skill can write files (writeFileSync), execute subprocesses (ffmpeg), and launch browser instances.
Sanitization: There is no mention of sanitizing or escaping the data retrieved from external websites before it is used to generate the animation code.
[DATA_EXPOSURE]: The Node.js rendering script (render.mjs) initializes the Playwright browser with the --allow-file-access-from-files flag. This setting allows the browser to read any file on the local filesystem, which could be exploited if the generated HTML includes untrusted data from the fetched URLs.
[EXTERNAL_DOWNLOADS]: The skill instructs the installation of the playwright package and the Chromium browser binary from external registries.

explainer-engaging