The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions for agent self-reflection. It contains no executable code, shell commands, binary files, or external dependencies.
[SAFE]: The skill configuration in SKILL.md and agents/openai.yaml explicitly restricts the tool to manual invocation by setting disable-model-invocation: true and allow_implicit_invocation: false, preventing unauthorized automated execution.
[SAFE]: No data exfiltration patterns, sensitive file access, or obfuscated content were detected within the skill files.

decisions