The Agent Skills Directory

[SAFE]: The skill is entirely functional and instructional, providing a legitimate framework for skill authoring without any detected malicious intent.
[EXTERNAL_DOWNLOADS]: The documentation references standard developer tools and package installations such as skills-ref (via pip) and skills (via npx), which are expected within this ecosystem.
[COMMAND_EXECUTION]: Includes Python helper scripts (init_skill.py and validate_skill.py) that perform local file system operations like directory creation and file validation. These scripts correctly limit themselves to the Python standard library.
[COMMAND_EXECUTION]: Provides documentation and examples for configuring shell-based hooks (e.g., git safety filters, auto-formatters). These are provided as user-implemented recipes for workflow automation and are not executed autonomously by the skill in a hidden manner.

creating-skills