The Agent Skills Directory

[SAFE]: The skill's operational logic is entirely contained within local static reference files and does not involve remote code execution or external dependencies.
[SAFE]: The skill instructions explicitly prohibit network calls and file system writes during standard use.
[SAFE]: No evidence of prompt injection or obfuscation was found in any of the skill's instructional or data files.
[SAFE]: References to external entities in the grounding metadata are for legitimate documentation best practices and do not represent a security risk.
[SAFE]: The skill defines clear boundaries and negative cases to prevent misuse, such as deferring auditing or debugging tasks to appropriate specialized skills.

docs-design