The Agent Skills Directory

[SAFE]: The instructions establish a rigorous development process focused on 'red-green-refactor' cycles, which inherently includes safety checks such as preventing production changes without failing tests.
[SAFE]: The skill configuration in agents/openai.yaml specifies allow_implicit_invocation: false, providing a security layer by requiring explicit user intent to activate the skill.
[SAFE]: Mentions of external repositories like mattpocock/skills are used solely for licensing attribution and do not constitute remote code execution or external dependencies.

skill-tdd