The Agent Skills Directory

Comprehensive Security Guidelines: The skill includes a detailed security-checklist.md which provides instructions on how to prevent shell injection when using the GitHub CLI and how to handle untrusted external content as data rather than instructions.
Local Scaffolding and Validation: It employs local Python scripts (init_skill.py) and project-specific tools (skill-and-tool-validator) for directory creation and rule enforcement, ensuring that new skills adhere to framework conventions.
Integrity Verification: The instructions include a pre-flight check that uses a local module (setup_preflight) and surface hashes to verify the skill's state before execution.
Transparent External References: Documentation within provenance.md clearly states the origins of the skill and provides links to well-known public repositories for credit, following standard open-source practices.

write-skill