Gemini CLI MCP Server

Gemini CLI MCP Server

Enables comprehensive codebase analysis using Google's Gemini AI through CLI integration. Provides architectural reviews and targeted code analysis with code2prompt integration for efficient context extraction.

Category
访问服务器

README

Gemini CLI MCP Server

A Model Context Protocol (MCP) server that integrates Gemini CLI with Claude Code for comprehensive codebase analysis using code2prompt.

Features

  • Comprehensive Codebase Review: Full architectural analysis using Gemini's 1M token context window
  • Targeted Component Analysis: Focused analysis on specific files/folders
  • Code2Prompt Integration: Efficient codebase context extraction with token optimization
  • OAuth Support: Works with your existing Gemini CLI OAuth authentication
  • Claude Code Sub-Agent: Specialized agent for intelligent analysis orchestration

Prerequisites

Required Tools

  1. Node.js 18+ - For running the MCP server
  2. Gemini CLI - Google's CLI tool with OAuth authentication
  3. code2prompt - Rust CLI tool for codebase extraction
  4. Claude Code - For sub-agent integration

Installation Steps

1. Install Gemini CLI

npm install -g @google/gemini-cli

Authenticate with your Google account:

gemini
# Follow OAuth setup prompts

2. Install code2prompt

# macOS/Linux with Homebrew
brew install mufeedvh/tap/code2prompt

# Or build from source
cargo install code2prompt

3. Install MCP Server Dependencies

cd /Users/ryangould/Dev/gemini-cli-mcp
npm install

Setup & Configuration

1. Build the MCP Server

npm run build

2. Configure Claude Code MCP Settings

Add to your Claude Code MCP settings (usually ~/.config/claude-desktop/mcp_settings.json):

{
  "mcpServers": {
    "gemini-cli-mcp": {
      "command": "node",
      "args": [
        "/Users/ryangould/Dev/gemini-cli-mcp/dist/server.js"
      ],
      "env": {}
    }
  }
}

3. Restart Claude Code

Restart Claude Code to load the new MCP server.

Available Tools

gemini_comprehensive_review

Performs a complete codebase architectural review using Gemini CLI.

Parameters:

  • workingDir (string): Path to the codebase directory
  • excludePatterns (array, optional): Additional patterns to exclude

Usage Example:

// Called by Claude Code sub-agent
gemini_comprehensive_review({
  workingDir: "/path/to/your/project"
})

gemini_targeted_analysis

Analyzes specific files or folders with focused insights.

Parameters:

  • workingDir (string): Path to the codebase directory
  • targetPaths (array): Specific files/folders to analyze
  • analysisContext (string, optional): Additional context for the analysis

Usage Example:

// Called by Claude Code sub-agent
gemini_targeted_analysis({
  workingDir: "/path/to/your/project",
  targetPaths: ["src/api/", "src/controllers/auth.js"],
  analysisContext: "Security review of authentication components"
})

Claude Code Integration

Sub-Agent Usage

The gemini-code-analyzer sub-agent is automatically available in Claude Code once the MCP server is configured.

Example Commands:

  • "Please analyze this entire codebase for architectural issues"
  • "Review the authentication module in src/auth/ for security vulnerabilities"
  • "Analyze the API endpoints in src/controllers/ for performance bottlenecks"

Manual Tool Access

You can also call the MCP tools directly through Claude Code's tool interface:

  1. Use /tools to see available MCP tools
  2. Select gemini_comprehensive_review or gemini_targeted_analysis
  3. Provide required parameters

Development

Running in Development Mode

npm run dev

Building for Production

npm run build
npm start

Project Structure

gemini-cli-mcp/
├── src/
│   ├── server.ts              # Main MCP server
│   ├── tools/                 # MCP tool implementations
│   │   ├── comprehensive-review.ts
│   │   └── targeted-analysis.ts
│   └── utils/                 # Utility functions
│       ├── code2prompt-executor.ts
│       └── gemini-executor.ts
├── prompts/                   # Prompt templates
│   ├── comprehensive-review.md
│   └── targeted-analysis.md
├── dist/                      # Built TypeScript output
└── package.json

Troubleshooting

Common Issues

"gemini command not found"

Ensure Gemini CLI is installed and in your PATH:

npm install -g @google/gemini-cli
which gemini

"code2prompt command not found"

Install code2prompt:

cargo install code2prompt
# or
brew install mufeedvh/tap/code2prompt

Authentication Issues

Re-authenticate with Gemini CLI:

gemini
# Follow OAuth prompts again

MCP Server Not Connecting

  1. Check Claude Code MCP settings configuration
  2. Verify the server path is correct
  3. Check console logs for errors
  4. Restart Claude Code after configuration changes

Logging

The MCP server logs to stderr. Monitor logs when running:

npm run dev

Configuration Options

Excluding Files/Patterns

The comprehensive review automatically excludes common patterns:

  • node_modules/**
  • .git/**
  • *.log
  • dist/**, build/**
  • .next/**
  • coverage/**

Add custom exclusions via the excludePatterns parameter.

Token Management

  • code2prompt provides token counting for large codebases
  • Gemini CLI handles the 1M token context window automatically
  • Large codebases are efficiently processed through code2prompt's optimization

Security Considerations

  • Uses your existing Gemini CLI OAuth authentication
  • No API keys stored in the MCP server
  • Codebase content is sent to Gemini for analysis
  • Consider data sensitivity when analyzing proprietary code

License

MIT License - See LICENSE file for details

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选