🧠 MCP Server for Enterprise Standards

This project runs an MCP (Model Context Protocol) server designed for enterprise code validation. It integrates with editors like VS Code and Cursor to provide tailored feedback inside your IDE — based on your organization's specific standards.

---

💼 Why Use This?

As an enterprise developer, you often wonder:

• Am I complying with internal policies?
• Can I expose this API endpoint?
• Is this deployment pattern secure or recommended?

GitHub Copilot can help, but doesn’t understand your enterprise context. With an MCP server like this one, you can supercharge Copilot with custom logic and internal rules.

🔗 MCP servers in GitHub Copilot

---

⚙️ How It Works

1. Your org (or you locally) run this MCP server via server.py
2. A developer writes code (e.g. Bicep or Python) in VS Code
3. The editor is configured to use this MCP server
4. GitHub Copilot (in Agent Mode) calls the MCP tool
5. The server checks code against enterprise rules using an LLM endpoint
6. Copilot receives feedback and proposes refactors or improvements

---

🖼️ Visual Overview

graph TD
    A["Developer writing code"] -->|"Asks for feedback"| B["GitHub Copilot"]
    B -->|"Calls"| C["MCP Server"]
    C -->|"Analyzes code against"| D["Enterprise Standards"]
    C -->|"Uses"| E["LLM Endpoint"]
    E -->|"Returns analysis"| C
    C -->|"Returns feedback"| B
    B -->|"Suggests improvements"| A 

Loading

---

🚀 Getting Started

🛠️ All setup instructions have moved to setup.md →

That file includes:

• ✅ System requirements (Python, Git, VS Code, DevContainer, etc.)
• 🐍 Native (non-container) setup using venv and pip
• 🐳 Containerized setup using WSL2 + Docker + DevContainer
• 🔐 .env file setup and configuration
• ▶️ Commands to launch and verify the server

---

🔌 LLM Configuration (Required)

The server uses Semantic Kernel to connect to a Large Language Model.

You'll need a .env file like this:

AZURE_OPENAI_API_KEY=your_api_key_here
AZURE_OPENAI_ENDPOINT=your_endpoint_here
AZURE_OPENAI_DEPLOYMENT_NAME=gpt-4-32k_for_example
AZURE_OPENAI_API_VERSION=2024-05-01-preview_for_example

Create this by copying the sample:

cp .env.sample .env

Then fill in the values for your Azure OpenAI instance.

---

🧪 Testing Locally

After setup:

python src/server.py

Then open: http://localhost:8000/mcp/ You should see the MCP server interface and be ready to integrate with GitHub Copilot.

---

📎 Related Links

• Model Context Protocol Overview
• GitHub Copilot MCP Server Docs
• Semantic Kernel