Lazy Mcp banner
machjesusmoto machjesusmoto

Lazy Mcp

Development community

Description

CLI tool to manage Claude Code MCP servers and memory files

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

Lazy MCP

[![npm version](https://img.shields.io/npm/v/lazy-mcp.svg)](https://www.npmjs.com/package/lazy-mcp) [![CI](https://github.com/machjesusmoto/lazy-mcp/actions/workflows/ci.yml/badge.svg)](https://github.com/machjesusmoto/lazy-mcp/actions/workflows/ci.yml) [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)

Intelligent lazy-loading for Claude Code MCP tools. Let the agent decide what tools to load based on usage patterns and context needs.

Overview

Lazy MCP is a Claude Code plugin that enables **agent-driven tool loading** with intelligent context optimization. Instead of loading all available MCP servers upfront (consuming precious context window space), Lazy MCP uses a registry-based approach to inform Claude about available tools and loads them on-demand based on your actual needs.

**What It Does**:

  • Agent-Driven Loading: Claude intelligently requests tools based on conversation analysis
  • Registry-Informed: Maintains a registry of tool capabilities, keywords, and usage patterns
  • Lazy Loading: Tools are loaded only when needed, then purged when usage ceases
  • Context Optimization: Automatically manages context window to prevent limitations
  • Profile Support: Save and switch between different tool configurations

The Problem

Traditional MCP integration loads all configured servers at session start, quickly consuming your context window with tools you may never use in a particular conversation. This leads to:

  • Frequent context limit errors
  • Wasted tokens on unused tool definitions
  • Poor agent performance due to context clutter

The Solution

Lazy MCP uses a **registry + lazy loading** approach:

  1. Registry: Maintains metadata about all available tools (capabilities, keywords, usage patterns)
  2. Agent Analysis: Claude analyzes your request and identifies potentially useful tools
  3. On-Demand Loading: Only requested tools are loaded into context
  4. **I