using-skills

To ensure right AI skills are loaded automatically before responding, invoke a pre-response policy that identifies applicable skills from a knowledge base and loads them before generating any reply. This enforces capability discovery at the conversation outset.

Skill discovery is the process of automatically detecting applicable capabilities before generating responses. It works by reading the name and description fields defined in the YAML frontmatter of a SKILL.md file to identify and load relevant skills.

To set up a pre-response invocation policy for AI workflow guidance, mandate the Skill tool invocation before responding to any conversation turn. This applies to initial chats, clarifying questions, and task planning to ensure correct capabilities are engaged.

Invoking skills before clarifying questions does improve task planning alignment by ensuring the system loads appropriate capabilities whenever a skill might apply. This reduces misalignment by guiding subsequent actions based on loaded skill knowledge.

You need to trigger skill discovery at the beginning of an AI conversation and before any response, including clarifications. This ensures the system detects applicable skills from the knowledge base and prompts their usage to guide subsequent actions.

A limitation of relying on SKILL.md frontmatter for skill discovery is that it requires a specific YAML structure with name and description fields to enable detection. Without this standardized frontmatter, the system cannot identify or load applicable capabilities.