import-skilllisted

# Import Skill Import skills from GitHub directory URLs or pasted content. Supports single-skill copy (preserves full directory structure) and multi-skill merge (intelligently combines into one). Only public GitHub repositories are supported. ## Speed Guidelines This workflow involves multiple GitHub API calls and file writes. Minimize turns by following these rules: - **Batch independent API calls** into single turns with parallel Bash calls - **Use `curl`** for all GitHub content fetches — WebFetch summarizes HTML instead of returning raw content, making it unsuitable for fetching source files - **Combine download + write** with `curl -s URL > path` instead of fetch-then-write-separately ## Process ### Step 1: Parse Source Accept one of: 1. **One GitHub directory URL** — single skill copy 2. **Multiple GitHub directory URLs** — multi-skill merge 3. **Pasted content** — user pastes SKILL.md content (and optionally other files) directly Expected URL format: `https://github.com/{owner}/{repo}/tree/{branch}/{path}` Parse to extract `owner`, `repo`, `branch`, and `path`. If the format doesn't match, ask for clarification. For pasted content, skip to Step 4. ### Step 2: Parallel Discovery **Launch ALL of these in a single turn using parallel tool calls:** ```bash # 1. Commit SHA curl -s "https://api.github.com/repos/{owner}/{repo}/commits?sha={branch}&per_page=1" | python3 -c "import sys,json; print(json.load(sys.stdin)[0]['sha'])" # 2. Directory listing curl -s "