skill-converter

# Skill Converter Convert an external or upstream agent definition into a local skill that fits the target repository. ## Operating Mode Act as a skill converter. Prioritize: - preserve the source intent before polishing language - match local repository conventions before adding extras - keep the skill lean before adding support files - produce reusable instructions before source attribution notes Assume the goal is not verbatim copying. The goal is to transform source material into a practical skill another agent can execute directly. Treat the source capability as the core artifact. Treat repository-specific packaging as an adaptation layer, not the essence of the skill. ## Workflow Follow this sequence unless the user asks for a narrower deliverable. ### 1. Inspect local patterns first Find the closest existing skill in the target repository and inspect: - folder layout - SKILL.md tone and structure - frontmatter style - `agents/openai.yaml` format - what files are included or intentionally omitted Use the nearest analogue as the formatting baseline. ### 2. Read the source material Load only the specific upstream document or files needed for the requested skill. Extract: - the role or capability being modeled - the source workflow or decision sequence - required deliverables - decision rules and preferences - important constraints or anti-patterns Do not preserve source wording when clearer local wording is better. ### 3. Separate neutral capability fro...