What problem does it solve? It removes the guesswork of packaging new capabilities for Claude Code in the CLAUDEMAX ecosystem by defining the exact file structure, metadata fields, validation steps, and registration commands for both declarative skills and MCP servers. ## Core Features & Use Cases - Skill 2.0 scaffolding: Defines the required three-file layout (SKILL.md, skill.yaml, schema.json), bilingual trigger conventions, and the validate-skills.mjs check that must pass before a skill ships. - MCP server pattern: Documents the canonical stdio server built on @modelcontextprotocol/sdk, with ListToolsRequestSchema and CallToolRequestSchema handlers, plus the rule that the MCP must be a thin wrapper over a testable CLI. - Decision guidance: Provides a comparison table for choosing between a knowledge skill, a CLI-backed tool skill, or a full MCP server based on whether you need context injection, shell commands, or protocol-validated tools. - Use Case: When you want to add a new capability to Claude Code, invoke this skill to generate the correct directory structure, pass the repo validator, register the MCP with claude mcp add, and wire symmetric uninstall logic. ## Quick Start Ask the AI to create a new skill or MCP server for Claude Code following the skill-mcp-builder conventions and run the validator until it reports OK.