The Agent Skills Directory

[SAFE]: The skill defines a clear and legitimate scope for assisting with the creation, refactoring, and debugging of agent skills. Its instructions emphasize following existing repository policies and user authorizations.
[SAFE]: The provided Python script scripts/validate-metadata.py is a read-only utility for validating skill metadata. It uses safe string handling and regex-based validation (e.g., ^[a-z0-9]+(-[a-z0-9]+)*$) to ensure skill names meet naming conventions, which effectively prevents command injection risks during validation.
[SAFE]: The skill does not perform any network operations, external downloads, or sensitive file access. All operations are restricted to the local environment and the processing of skill-related documentation.
[SAFE]: Reference files (glossary, checklist, procedures) contain benign instructional content intended to improve the quality and performance of AI agents without attempting to bypass safety guardrails or persist across sessions.

writing-skills