The Agent Skills Directory

[SAFE]: The skill serves as a structural and instructional guide for authors creating AI agent capabilities. It contains 46 rules across 8 categories, including metadata design, description engineering, and progressive disclosure.
[EXTERNAL_DOWNLOADS]: The documentation references and recommends the use of the skills-ref library and the anthropics/skills repository for validation and reference purposes. These sources are associated with official and well-known development communities.
[COMMAND_EXECUTION]: The skill provides examples of shell commands for validation (e.g., skills-ref validate and python -c "import yaml...") and documentation (e.g., pip install skills-ref). These are standard development utility commands intended for the user to run on their own project files.
[SAFE]: The skill includes a dedicated section on security auditing (references/maint-audit-security.md), which correctly identifies and warns against common attack vectors such as data exfiltration, credential theft, and obfuscated code.

skill-authoring