Code Review Server

Name: Code Review Server
Author: crazyrabbitLTC

@crazyrabbitLTC

代码仓库

关于 Code Review Server

A MCP server for code reviews

基本信息

分类

版本控制

运行时

node

传输方式

stdio

发布者

crazyrabbitLTC

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "mcp-code-review-server": {
      "command": "node",
      "args": [
        "build/index.js"
      ]
    }
  }
}

工具

未检测到工具

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is Code Review Server?

A custom MCP server that performs code reviews by flattening repositories with Repomix and analyzing them using Large Language Models (LLMs). It is intended for developers who want structured, AI‑powered code reviews with multiple LLM provider options.

How to use Code Review Server?

Clone the repository, run npm install and npm run build, then create a .env file with your preferred LLM provider and API key. Start the MCP server with node build/index.js or use the CLI tool (node build/cli.js <repo_path> [options]). The server exposes two tools: analyze_repo (flattens a codebase) and code_review (performs an LLM‑based review).

Key features of Code Review Server

Flatten codebases using Repomix
Analyze code with multiple LLMs (OpenAI, Anthropic, Gemini)
Get structured reviews with issues, strengths, and recommendations
Handle chunking for large codebases
Supports CLI and MCP protocol usage

Use cases of Code Review Server

Quickly scan a repository’s structure before a detailed review
Perform comprehensive code quality assessments with severity ratings
Identify security vulnerabilities, performance bottlenecks, or maintenance issues
Focus reviews on specific files, file types, or focus areas (e.g., security, quality)

FAQ from Code Review Server

What LLM providers are supported?

OpenAI (default gpt‑4o), Anthropic (default claude‑3‑opus‑20240307), and Gemini (default gemini‑1.5‑pro). You can also override the default models for each provider.

How do I configure the server?

Set your preferred provider via LLM_PROVIDER and the corresponding API key in a .env file (e.g., OPENAI_API_KEY). Optional model overrides are available.

What does the code review output look like?

A structured JSON object containing a summary, a list of issues (each with type, severity, description, line numbers, and recommendation), a list of strengths, and overall recommendations.