The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses several Python scripts and CLI tools to manage the development lifecycle of other skills.
Runs background optimization loops using scripts/run_loop.py.
Executes scripts/aggregate_benchmark.py to process performance data.
Uses claude -p via subprocess in scripts/run_eval.py to simulate agent triggering based on descriptions.
Provides a packaging utility scripts/package_skill.py that creates .skill zip files.
[EXTERNAL_DOWNLOADS]: The skill references and fetches standard web assets and libraries for its evaluation viewer.
Loads SheetJS from a CDN (cdn.sheetjs.com) in the HTML viewer to render Excel files.
Uses Google Fonts (fonts.googleapis.com) for the review interface.
These are well-known, trusted sources for frontend assets.
[SAFE]: The skill implements security best practices by separating logic into specialized scripts and using a local-only HTTP server for the evaluation viewer. It explicitly includes safety guidelines in its instructions (SKILL.md) warning against the creation of malicious skills.

skill-creator