| name | skill-distillation-pro |
| description | Distill books, articles, external repositories, expert workflows, team conventions, and repeated project experience into concise executable Codex skills. Inspired by kangarooking/cangjie-skill and adapted for Codex v4.3. Use when converting knowledge into SKILL.md files, improving skill packs, or extracting reusable methods from references. |
Skill Distillation Pro
Purpose
Turn high-value knowledge into compact, executable skills.
Use this for:
- Books and long articles.
- External skill repositories.
- Team conventions.
- Repeated project workflows.
- Expert heuristics.
- Postmortems and lessons learned.
Workflow
- Define the target skill trigger: when should Codex use it?
- Extract reusable principles, procedures, checklists, failure modes, and outputs.
- Remove background explanation that does not guide action.
- Convert knowledge into
SKILL.md with name, description, workflow, rules, and output contract.
- Add references only when details are too large for the main skill.
- Validate that the skill is concise, triggerable, and not redundant with existing skills.
Extraction Checklist
- What task does this knowledge improve?
- What inputs are required?
- What sequence should Codex follow?
- What mistakes should Codex avoid?
- What should the final output include?
- What evidence or verification is expected?
- Which existing skills overlap?
Quality Rules
- Keep
description trigger-rich and specific.
- Keep body procedural, not essay-like.
- Prefer checklists and workflows over theory.
- Do not copy large external text verbatim.
- Preserve licensing awareness for external content.
- If source quality is uncertain, mark the skill as experimental or adapt only the method.
Output
Provide the distilled skill file, source inspiration, overlap analysis, and suggested install location.
