Railagent - MCP Server for AI Agent Workflows

A Model Context Protocol (MCP) server that provides structured workflows and tools for AI agents working with software development projects, with specialized focus on Rails applications.

What This Does

Railagent provides AI agents with a comprehensive workflow system for software development. It guides agents through the entire development lifecycle from requirements gathering to implementation, ensuring consistent and high-quality outcomes.

How It Works

For New Projects

1. Functional Requirements: Use the functional requirements tool to collect necessary information about the project (purpose, users, capabilities, workflows, context, edge cases)
2. Product Requirements Document (PRD): Generate a complete PRD from the functional requirements, breaking down the project into features and user stories
3. Feature Requirements Document (FRD): For each feature, create a detailed FRD that breaks large features into individual PR-sized units
4. Task Design Document (TDD): Transform the FRD into an implementation plan with specific subtasks, where each subtask corresponds to a commit

For Existing/Complex Systems

• Skip directly to step 4 (TDD) above when you have a clear understanding of what needs to be built

Implementation Workflow

1. Initialize Tasks: Call the initialize tool to set up subtasks based on your TDD implementation plan (stored in .cursor/scratch/tasks/)
2. Execute Tasks: Use the execute tool to work on each task individually. Each task represents exactly one commit, and the agent waits for review before committing
3. Track Progress: Progress is maintained in .cursor/scratch/todo.md with notes and status updates

All documents and tasks are automatically stored in the .cursor/scratch/ folder for easy access and organization.

Available Tools

Documentation Tools

• build_functional_requirements: Collects comprehensive project requirements
• build_prd: Generates Product Requirements Documents from functional requirements
• build_frd: Creates Feature Requirements Documents for specific features
• build_tdd: Breaks down features into detailed implementation plans with commit-level subtasks
• build_architecture: Generates comprehensive architecture documentation

Workflow Tools

• initialize_task: Sets up implementation subtasks from TDD plans
• execute_task: Executes individual tasks with review checkpoints

Setup Instructions

1. Configure MCP Client

Add to your mcp.json file:

{
  "mcpServers": {
    "Railagent": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "cestbalez/railagent:latest"
      ]
    }
  }
}

2. Restart Your MCP Client

Restart your MCP client to load the new server.

Note: The Docker image will be automatically pulled from DockerHub on first use. No manual installation or repository cloning required!

Project Structure

railagent/
├── app.rb                      # Main application entry point
├── lib/                        # Tool implementations
│   ├── docs/                   # Documentation tools
│   └── workflow/               # Workflow management tools
├── prompts/                    # Workflow prompts and templates
│   ├── docs/                   # Documentation workflow prompts
│   └── workflow/               # Implementation workflow templates
├── Dockerfile                  # Docker configuration
├── Gemfile                     # Ruby dependencies
└── README.md                   # This file

Development

Using the Pre-built Image

The easiest way to use Railagent is with the pre-built Docker image:

docker run -i --rm cestbalez/railagent:latest

Local Development

If you want to contribute or modify the code:

git clone <your-repo-url>
cd railagent
bundle install
ruby app.rb

Building Your Own Image

git clone <your-repo-url>
cd railagent
docker build -t railagent .

Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

This project is licensed under the terms specified in the LICENSE file.