MCP Unit Test Sensei

MCP Testing Sensei

This project implements an MCP (Model Context Protocol) stdio server designed to enforce and guide agentic coding tools (like Gemini CLI or Claude Code) in adhering to language agnostic unit testing principles.

Core Principles Enforced

This tool aims to promote the following unit testing principles:

Tests should be written before implementation. (BDD/TDD for the win)
Tests should document the behavior of the system under test.
Tests should be small, clearly written, and have a single concern.
Tests should be deterministic and isolated from the side effects of their environment and other tests.
Tests should be written in a declarative manner and never have branching logic.

Features

lint_code tool: Analyzes provided code snippets for violations of the defined unit testing standards.
get_testing_principles tool: Provides the core unit testing principles to guide LLMs in generating better tests.
unit-testing-principles resource: Exposes testing principles as an MCP resource.

Installation

Option 1: Install from PyPI (Recommended)

The easiest way to install MCP Testing Sensei is via pip:

pip install mcp-testing-sensei

This installs the mcp-testing-sensei command globally.

Option 2: Install from npm

If you prefer using npm:

npm install -g @kourtni/mcp-testing-sensei

Note: This still requires Python to be installed on your system.

Option 3: Using Docker

docker pull kourtni/mcp-testing-sensei

Option 4: Development Setup with Nix

For development or if you want to build from source:

Prerequisites

Nix (for reproducible development environment)

Development Environment Setup

To enter the development environment with all dependencies:

nix develop

Building the Standalone Executable

To build the standalone executable using Nix, run the following command:

nix build

This will create a result symlink in your project root, pointing to the built executable.

Running the Server

Using the Standalone Executable

After building, you can run the MCP stdio server directly from the result symlink:

./result/bin/mcp-testing-sensei

This will start the MCP server that communicates via standard input/output.

Running from Development Environment

Alternatively, if you are in the nix develop shell, you can run the MCP server:

python mcp_server.py

The server communicates via stdio, reading JSON-RPC messages from stdin and writing responses to stdout.

Using with MCP Clients

The server can be integrated with MCP-compatible clients like Claude Desktop or other tools that support the Model Context Protocol.

Configuration for Claude Desktop

If installed via pip:

{
  "mcpServers": {
    "testing-sensei": {
      "command": "mcp-testing-sensei"
    }
  }
}

If installed via npm:

{
  "mcpServers": {
    "testing-sensei": {
      "command": "npx",
      "args": ["@kourtni/mcp-testing-sensei"]
    }
  }
}

If using Docker:

{
  "mcpServers": {
    "testing-sensei": {
      "command": "docker",
      "args": ["run", "-i", "kourtni/mcp-testing-sensei"]
    }
  }
}

If running from source:

{
  "mcpServers": {
    "testing-sensei": {
      "command": "python",
      "args": ["/path/to/mcp-testing-sensei/mcp_server.py"]
    }
  }
}

Testing the Server

To verify the server is working correctly, you can use the integration test script:

# For development testing
python test_mcp_integration.py

This will:

Start the MCP server
Send test requests to verify the tools are working
Display the responses

The server itself doesn't have a standalone test mode - it's designed to be used with MCP clients.

Development

Running Tests

To run the unit tests locally, first ensure you are in the Nix development environment:

nix develop

Then, execute pytest:

pytest

Project Structure

flake.lock
flake.nix
linter.py           # Core linting logic
mcp_server.py       # MCP stdio server implementation
main.py             # Legacy HTTP server (can be removed)
pyproject.toml
test_mcp_integration.py  # Integration test script for the MCP server
tests/
    test_linter.py  # Unit tests for the linter logic

Contributing

Contributions are welcome! Please ensure your changes adhere to the established unit testing principles and project conventions.

Additional Documentation

DISTRIBUTION.md - Detailed guide for all distribution methods
RELEASE.md - Release process and version management

This server cannot be installed

security - not tested

license - not found

quality - not tested

How are these scores calculated?

hybrid server

The server is able to function both locally and remotely, depending on the configuration or use case.

An MCP server that enforces unit testing standards by linting code and guiding agentic coding tools (like Gemini CLI or Claude Code) to follow best practices in test-driven development.

Related MCP Servers

MCP Claude Code
SDGLBL
-
security
A
license
-
quality
An MCP server that implements Claude Code-like functionality, allowing the AI to analyze codebases, modify files, execute commands, and manage projects through direct file system interactions.
Last updated -
224
Python
MIT License
MCP Everything
s2005
-
security
A
license
-
quality
A test server implementing all features of the MCP protocol, including prompts, tools, resources, and sampling, designed for testing MCP clients rather than practical applications.
Last updated -
Python
MIT License
Pentest Tools MCP Server
ch1nhpd
-
security
F
license
-
quality
An MCP server that integrates various penetration testing tools, enabling security professionals to perform reconnaissance, vulnerability scanning, and API testing through natural language commands in compatible LLM clients like Claude Desktop.
Last updated -
3
Python
SystemPrompt Coding Agentofficial
systempromptio
-
security
A
license
-
quality
An MCP server that orchestrates AI coding assistants (Claude Code CLI and Gemini CLI) to perform complex programming tasks autonomously, allowing remote control of your local development environment from anywhere.
Last updated -
7
93
TypeScript
MIT License

View all related MCP servers