Coding Standards MCP Server

Automatically enforces team coding standards in AI-assisted development

🎯 What This Does

Your AI assistant (Cursor, Claude, etc.) will automatically follow your team's coding standards when generating code.

Benefits:

• ✅ Consistent code across your entire team

• ✅ No more reminding AI about your standards

• ✅ Centralized standards - update once, apply everywhere

• ✅ Works with Python, Java, Node.js, React, and more

📖 Read the blog post: From AI Coding Chaos to Team Standards - Learn about the challenges with file-based standards and why MCP might be the ideal solution for teams (as of now, landscape changing very fast).

Related MCP server: MCP Standards Server

🚀 How to Use

Step 1: Add MCP Server to Cursor

Edit ~/.cursor/mcp.json (create if it doesn't exist):

{
  "mcpServers": {
    "coding-standards": {
      "name": "Coding Standards",
      "description": "IMPORTANT: Before generating ANY code, ALWAYS check this server for coding standards. Provides mandatory team coding standards, style guides, best practices, and patterns for Python, Java, Node.js, React/Next.js, JavaScript, general coding, and debugging. Call get_coding_standard() with the language name before writing code to ensure compliance with team standards. And always use the engineering manager ai tool to get the coding standards before writing code.",
      "url": "https://web-production-ad318.up.railway.app/sse"
    }
  }
}

Step 2: Restart Cursor

Close and reopen Cursor. That's it! 🎉

✨ How It Works

Automatic Mode (Recommended):

Just start coding! The AI will:

1. Detect the language you're using

2. Automatically fetch your team's standards

3. Generate code that follows YOUR rules

Manual Mode:

Ask the AI:

• "Show me the Python coding standards"

• "List all available coding standards"

• "What are our React best practices?"

📋 Available Standards

The server provides standards for:

Ask the AI: "List all available coding standards" to see the full list with descriptions.

🔧 Available MCP Tools

The AI can use these tools:

• list_coding_standards() - See all standards (mandatory vs optional)

• get_coding_standard(category, name) - Get specific standards

• get_standards_for_project([languages]) - Get multi-language standards

You don't need to call these manually - the AI knows when to use them!

💡 Quick Tips

Want the AI to follow a specific standard?

"Use our FastAPI standards to create a new API endpoint"

Check what standards apply to your current code?

"What coding standards should I follow for this Python project?"

See all standards at once?

"Show me all our team's coding standards"

🔄 Standards Update Automatically

When standards are updated on the server, your AI automatically gets the latest version. No configuration needed!

👨‍💻 Development Guide

For developers who maintain or deploy this server

📁 Project Structure

ai-sdlc/
├── server.py                           # MCP server implementation
├── coding-standards/                   # Standards directory
│   ├── code_complexity.md             # Mandatory for all code
│   ├── debugging.md                   # General debugging
│   ├── development_process.md         # Process guidelines
│   ├── server_operations.md           # Server ops
│   ├── python/
│   │   ├── fastapi.md                 # FastAPI standards
│   │   └── general.md                 # Python standards
│   ├── nodejs/
│   │   ├── expressjs.md               # Express.js standards
│   │   └── general.md                 # Node.js standards
│   ├── react_and_nextjs/
│   │   └── general.md                 # React/Next.js standards
│   └── vanilla_js/
│       └── general.md                 # Vanilla JS standards
├── requirements.txt                    # Python dependencies
├── pyproject.toml                     # Python project config
├── uv.lock                            # UV lock file
├── Procfile                           # Railway deployment config
└── railway.json                       # Railway configuration

🚀 Local Development

Prerequisites

• Python 3.11+

• UV (recommended) or pip

Setup & Run

Option 1: Using UV (Recommended)

# Clone the repo
git clone <your-repo-url>
cd ai-sdlc

# Run server (UV handles dependencies automatically)
uv run server.py

Option 2: Using pip

# Install dependencies
pip install -r requirements.txt

# Run server
python server.py

The server will start on http://localhost:8000/sse

Local MCP Configuration

For local development, use this in ~/.cursor/mcp.json:

{
  "mcpServers": {
    "coding-standards": {
      "name": "Coding Standards (Local)",
      "description": "Local development coding standards server",
      "url": "http://localhost:8000/sse"
    }
  }
}

📝 Adding/Editing Standards

Standard File Format

Each standard file uses YAML frontmatter:

---
description: Brief description of the standard
status: active
mandatory: false
---

# Your Standard Title

[Your content here...]

Frontmatter Fields

Adding a New Standard

1. Create the file:

# For language-specific standard
mkdir -p coding-standards/golang
touch coding-standards/golang/general.md

# For general standard
touch coding-standards/testing.md

2. Add frontmatter and content:

---
description: Go coding standards and best practices
status: active
mandatory: false
---

# Go Coding Standards

## Code Organization
...

3. Test:

# Restart server (changes are auto-detected)
# Ask AI: "List all coding standards"

Making a Standard Mandatory

Edit the frontmatter:

---
description: Must follow for ALL code
status: active
mandatory: true  # ← This makes it mandatory
---

Mandatory standards appear in the "🚨 MANDATORY STANDARDS" table and apply to ALL code, regardless of language.

🧪 Testing

Manual Testing

# Start server
uv run server.py

# In another terminal, test the endpoint
curl http://localhost:8000/sse

Testing with MCP Inspector

# Install and run inspector
npx @modelcontextprotocol/inspector

Open http://localhost:6274/ and configure:

• Transport: SSE

• URL: http://localhost:8000/sse

• Connection Type: Via Proxy

Test all MCP tools visually!

Testing in Cursor

1. Configure local MCP server in ~/.cursor/mcp.json

2. Restart Cursor

3. Ask: "List all coding standards"

4. Verify output matches your files

🔍 MCP Server Implementation

Key Functions

parse_frontmatter(content: str)

• Parses YAML frontmatter from markdown

• Converts boolean strings to Python booleans

get_available_standards()

• Scans coding-standards/ directory

• Filters by status: active

• Returns dict with general and language-specific standards

list_coding_standards() (MCP Tool)

• Lists all active standards

• Separates mandatory vs language-specific

• Returns formatted markdown table

get_coding_standard(category, name) (MCP Tool)

• Fetches specific standard content

• Returns full markdown content

get_standards_for_project(languages) (MCP Tool)

• Gets standards for multiple languages

• Always includes general standards

• Returns combined markdown

📊 Monitoring

Check server health:

# Railway logs
railway logs

# Or visit your app URL
curl https://your-app.railway.app/sse

🐛 Troubleshooting

Standards not showing up?

• Check status: active in frontmatter

• Verify file is in correct directory

• Restart server

MCP not working in Cursor?

• Verify ~/.cursor/mcp.json syntax

• Check URL is correct

• Restart Cursor completely

Server not starting?

• Check Python version (3.11+)

• Verify all dependencies installed

• Check port 8000 is available

📚 Learn More

From AI Coding Chaos to Team Standards - Read our journey from file-based standards to MCP, including:

• The challenges with agents.md and .cursorrules

• Why AI agents ignore file-based rules

• How MCP solves team-scale standards enforcement

• Real-world implementation guide

Research Documentation - Deep dive into our findings:

• agents.md reliability issues - Community reports and root causes

• MCP architecture advantages - Technical comparison and benefits

• Key statistics and quotes - Data points and evidence

Questions? Open an issue or ask the AI to help you use the coding standards! 🚀