MCP Code Expert System

Let an expert review your code changes

python

3

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "tomsiwik-mcp-experts": {
      "command": "python",
      "args": [
        "server.py"
      ],
      "env": {
        "OLLAMA_HOST": "http://localhost:11434",
        "OLLAMA_MODEL": "llama3:8b",
        "KNOWLEDGE_GRAPH_PATH": "data/knowledge_graph.json"
      }
    }
  }
}

You can run the Code Expert System MCP server to perform AI-assisted code reviews using expert personas inspired by Martin Fowler and Robert C. Martin. It stores review data and relationships in a knowledge graph, integrates with Ollama for AI reviews, and supports server-sent events for web integration. This guide shows practical steps to install, configure, and run the server, plus how to interact with its features.

How to use

Launch the MCP server in standard mode to interact through a local client or integrate via Cursor-like tooling. Use the SSE transport mode to enable web-based streams of review results. You can prompt the system to review code snippets, read and navigate the knowledge graph, and run reviews by the Martin Fowler and Uncle Bob personas.

How to install

Prerequisites you need before installing and running the server:

# Install Python 3.10 or newer if not already installed
# Then install the dependencies and set up the virtual environment
chmod +x setup.sh
./setup.sh

# Ensure a Python virtual environment is active
source .venv/bin/activate

Additional sections

Configuration and runtime options are described below. Create a .env file from the example if needed, and adjust paths and model settings to suit your environment.

# Knowledge Graph path
KNOWLEDGE_GRAPH_PATH=data/knowledge_graph.json

# Ollama configuration for local AI models
OLLAMA_HOST=http://localhost:11434
OLLAMA_MODEL=llama3:8b

Usage notes and end-to-end flow

Start the server in standard mode for cursor-based integration. 2) Start the server with SSE transport for web integration. 3) Interact with tools to review code, read and search the knowledge graph, and run expert reviews.

# Standard mode
source .venv/bin/activate
python server.py

# HTTP/SSE mode
source .venv/bin/activate
python server.py --transport sse

Use the above commands to begin listening on the defaults or on a custom port with --port.


## Available tools

### ask\_martin

Ask Martin Fowler to review code and suggest refactorings using Fowler-inspired principles.

### ask\_bob

Ask Robert C. Martin (Uncle Bob) to review code based on Clean Code principles.

### read\_graph

Read and traverse the knowledge graph that stores code, reviews, and relationships.

### search\_nodes

Search for specific nodes within the knowledge graph.

### open\_nodes

Open and view specific nodes by their names in the knowledge graph.