--- name: import-command description: Convert Cursor custom commands (.cursor/commands/*.md) to Claude Code skills (.claude/skills/*/SKILL.md). Use when the user says "import command to skill", "convert command to skill", or "migrate cursor command". --- # import-command Convert Cursor custom commands to Claude Code skills. ## Trigger Examples - "Import command to skill" - "Convert command to skill" - "Migrate cursor command to skill" - "Turn this command into a skill" ## Format Differences ### Cursor Command Format ```markdown # /command-name ## Overview [Description] ## Usage [Usage pattern] ## Prerequisites [Prerequisites] ## Execution Flow (steps) [Steps] ## AI considerations [AI guidance] ## Notes [Notes] ## Examples [Examples] ``` ### Claude Code Skill Format ```markdown --- name: skill-name description: [What it does and when to use it. Include trigger phrases.] --- # skill-name [One-line overview] ## Trigger Examples - "Trigger phrase 1" - "Trigger phrase 2" ## Prerequisites (optional) - [Prerequisite] ## Execution Flow ### 1. [Step name] - [Details] ## Notes - [Note] ``` ## Conversion Rules ### 1. Command Name → Skill Name - Remove leading `/` - Keep kebab-case: `/commit-push-pr` → `commit-push-pr` ### 2. Overview → Description (Frontmatter) - Combine "Overview" and "Usage" into a concise description - Add trigger phrases based on the command name and purpose - This is the most important part: description determines when the skill triggers ### 3. Usage → Trigger Examples - Convert usage patterns to natural language trigger phrases - Example: `/check-secrets [--strict]` → "Check for secrets", "Scan for leaks" ### 4. Execution Flow (steps) → Execution Flow - Keep the step structure - Simplify verbose explanations - Remove redundant context (Claude is smart) ### 5. AI considerations → Merge into Execution Flow or Notes - If actionable: merge into Execution Flow - If general guidance: move to Notes ### 6. Examples → Omit or Simplify - Skills prefer concise instructions over verbose examples - Keep only if essential for understanding ## Execution Flow ### 1. Read Source Command Read the Cursor command file from `.cursor/commands/.md` ### 2. Extract Key Information - Command name - Overview/description - Usage patterns - Prerequisites - Execution steps - Notes and caveats ### 3. Generate Skill Structure Create skill directory and SKILL.md: ``` .claude/skills// └── SKILL.md ``` ### 4. Write Frontmatter ```yaml --- name: description: --- ``` ### 5. Write Body - Concise trigger examples - Simplified execution flow - Essential notes only ### 6. Present for Approval Show the generated skill and ask for approval before creating. ### 7. Update CLAUDE.md Add the new skill to the skills list in CLAUDE.md. ## Notes - Skills should be more concise than commands - Focus on trigger phrases in description - Remove redundant explanations - Deprecated commands should not be converted