The Agent Skills Directory

[SAFE]: The skill is composed exclusively of natural language instructions and JSON-formatted evaluation fixtures. It does not define any executable tools, scripts, or binary assets.
[SAFE]: The methodology provides specific security-related guidance, instructing practitioners to audit for and reject content that attempts to bypass host permissions or relies on hidden execution mechanics.
[SAFE]: All references and linked documents are local to the skill directory. There are no external network calls, remote downloads, or dependencies on unverified third-party services.
[SAFE]: No patterns of obfuscation, credential exposure, or persistence were detected in the provided skill files.

skill-authoring