The Agent Skills Directory

[SAFE]: The skill consists primarily of documentation (SKILL.md and PHASE-BOUNDARIES.md) that describes workflows and organizational strategies for AI development. It does not contain any executable scripts, shell commands, or network operations.
[SAFE]: The frontmatter explicitly sets disable-model-invocation: true, indicating that this skill is intended as a documentation reference rather than an active tool for the AI to execute.
[SAFE]: All referenced 'skills' (e.g., /grill-with-docs, /handoff) are described as internal primitives within the user's environment, not external remote code execution triggers.
[SAFE]: The configuration file agents/openai.yaml is a standard metadata file that defines the skill's interface and invocation policy, posing no security risk.

ask-skills