The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes references to curated image assets from Pexels and Unsplash within the fetch-background.py script and slide-backgrounds.csv data. These are well-known, reputable image hosting services used for legitimate design purposes.
[EXTERNAL_DOWNLOADS]: Generated HTML slides include a script reference to the Chart.js library hosted on jsDelivr (cdn.jsdelivr.net), a well-known and trusted content delivery network.
[COMMAND_EXECUTION]: The skill utilizes several internal Python and Node.js scripts for token generation, search functionality, and slide creation. These scripts operate on local data and assets provided within the skill package.
[COMMAND_EXECUTION]: The slide-token-validator.py script uses subprocess.run to execute a unified HTML validator. The execution is performed using a list of arguments without a shell, which is a safe practice that prevents command injection.
[SAFE]: The skill implements a structured three-layer design token architecture (primitive, semantic, and component) which is a standard industry best practice for maintainable design systems.
[SAFE]: While some scripts use hardcoded relative path traversal (e.g., parents[4]) to locate the skill's assets directory, this appears to be an artifact of the developer's environment rather than a malicious attempt to access sensitive files.

design-system