AI Pair Programmer (Ruby)

AI-powered tools for code review, brainstorming, performance analysis, and security review in Ruby.

A Ruby implementation of an AI Pair Programmer MCP (Model Context Protocol) server that provides AI-powered tools for code review, brainstorming, performance analysis, and security review.

This server provides 5 AI-powered tools:
- pair- General collaboration and problem-solving
- review- Comprehensive code review with actionable feedback
- brainstorm- Creative ideation and solution exploration
- review_performance- Performance analysis and optimization suggestions
- review_security- Security-focused code review and vulnerability detection

The server uses inline bundler for automatic gem installation. Required gems:

- fast-mcp- Ruby MCP server framework
- ruby_llm- Unified AI model interface

export OPENROUTER_API_KEY="your_api_key_here"

- Automatically install missing gems on first run
- Start with STDIO transport for MCP clients
- Log to STDERR which should be saved by your MCP Host.

If the MCP fails to start when lunching Claude Code it's probably due to timeout. The bundler is installing dependecies. To fix that either set an envvarMCP_TIMEOUT=10000(10s) or start the MCP server in the terminal first.

The server supports these AI models via OpenRouter:

- Gemini(default) -google/gemini-2.5-pro-preview
- O3-openai/o3
- Grok-x-ai/grok-3-beta
- DeepSeek-deepseek/deepseek-r1-0528
- Opus-anthropic/claude-opus-4

Add to yourclaude_desktop_config.json:

{ "mcpServers": { "AiPairProgrammer": { "command": "ruby", "args": ["/path/to/server.rb"] } } }

Please review this Ruby code for best practices and potential issues: def process_data(items) items.map { |item| item.upcase } end

Can you analyze this code for performance bottlenecks? def find_duplicates(array) duplicates = [] array.each do |item| if array.count(item) > 1 && !duplicates.include?(item) duplicates << item end end duplicates end

Please check this authentication code for security vulnerabilities: def authenticate(username, password) user = User.find_by(username: username) if user && user.password == password session[:user_id] = user.id true else false end end

I need ideas for improving user onboarding in my Ruby on Rails app. The current flow has a 60% drop-off rate.

I'm struggling with this algorithm problem. Can you help me think through it step by step?

- FastMcp- Ruby MCP server framework with STDIO transport
- RubyLLM- Unified interface to AI models via OpenRouter
- Inline Bundler- Automatic gem installation for easy deployment

AI Assistant MCP Server by Eduardhttps://github.com/eduardm/ai_pairs_with_ai
- Fork the repository
- Create a feature branch
- Make your changes
- Test withruby test_server.rb
- Submit a pull request

Build and deploy full-stack Next.js apps with 98 tools for React, AWS, and MongoDB