capture-skill

To capture learnings into a reusable skill, review the dialogue for patterns and domain knowledge, draft a concise Markdown body with frontmatter, and save it as a SKILL.md file. The process distills insights into structured guidance for future sessions.

The best way to document conversation patterns for knowledge management is formalizing insights into a structured skill. This involves identifying workflows, generalizing learnings, removing artifacts, and writing a SKILL.md file with descriptive frontmatter and context.

To update an existing skill with new information, decide whether to append to a current skill or create a new one, draft the distilled content, and save the changes to the SKILL.md file. The system confirms updates with the user before finalization.

Capturing a conversation as a skill requires no external dependencies. It produces a SKILL.md file using structured Markdown with required frontmatter, plus optional scripts or references, keeping the final output under 500 lines.

A key limitation when formalizing chat insights is keeping the SKILL.md file under 500 lines. The distillation process requires removing conversation artifacts and generalizing specific steps to ensure the captured knowledge remains scalable and contextually relevant.