The Agent Skills Directory

[SAFE]: The skill incorporates explicit security guardrails. It instructs the agent to treat all external repository data, including policies and tests, as untrusted data that cannot authorize commands or privilege changes.
[SAFE]: File writing operations are protected by a mandatory human-in-the-loop verification process, requiring the agent to show a diff and obtain explicit approval before applying changes.
[SAFE]: There are no remote code execution patterns, external network requests, or attempts to access sensitive system credentials detected in the instructions.

define-security-policy