MCP Test Client

🚀 MCP Test Client

An MCP middleware that serves as a testing tool between two ends (Claude and the server under test), used for testing MCP servers during development.

🚀 Quick Start

The MCP Test Client is a versatile tool designed to simplify the testing process of MCP servers. It allows you to test MCP servers without directly registering on Claude.

✨ Features

• Deploy MCP servers in the test environment.
• Invoke individual tools with custom parameters.
• Run automated test suites.
• View server logs.
• Test servers before official registration on Claude.

📦 Installation

# Install dependencies
npm install

# Build the TypeScript project
npm run build

💻 Usage Examples

Basic Usage

As an MCP Server

The MCP Test Client is registered with Claude via the claude-mcp-local script. You can use the following tools:

1. Deploy a server:

mcp__mcp-test__mcp_test_deploy_server({
  name: "my-server",
  source_path: "/path/to/server",
  env_vars: {
    "API_KEY": "${API_KEY}"
  }
})

2. Call a tool on the deployed server:

mcp__mcp-test__mcp_test_call_tool({
  server_name: "my-server",
  tool_name: "tool_name",
  arguments: {
    // Tool-specific parameters
  }
})

3. Run tests on the server:

mcp__mcp-test__mcp_test_run_tests({
  server_name: "my-server"
})

4. View server logs:

mcp__mcp-test__mcp_test_get_logs({
  server_name: "my-server",
  lines: 100
})

5. List deployed servers:

mcp__mcp-test__mcp_test_list_servers({})

6. Stop a server:

mcp__mcp-test__mcp_test_stop_server({
  server_name: "my-server"
})

As a CLI Tool

Run the CLI interface for testing:

# Use the npm script
npm run test

# Or run directly
node dist/test-runner.js

This provides an interactive menu for deploying, testing, and managing MCP servers.

Advanced Usage

The current implementation is the first phase of the design plan, with enhancements planned for the second and third phases in the future.

📚 Documentation

Architecture

The MCP Test Client has a dual role:

• It is a server registered with Claude and exposes tools for testing.
• It is a client that connects to and tests other MCP servers.

┌─────────────┐          ┌───────────────────┐          ┌────────────────┐
│             │  Tools   │                   │  Client   │                │
│   Claude    │─────────>│  MCP Test Client  │─────────>│  Server Under Test │
│             │          │                   │          │                │
└─────────────┘          └───────────────────┘          └────────────────┘

This architecture allows you to test MCP servers without directly registering on Claude.

Implementation

The MCP Test Client is implemented using the following technologies:

• Process Management: Spawn and manage MCP server processes.
• Official MCP SDK Client: Communicate with the server using the official MCP SDK.
• Custom Transport: Implement custom transport for standard input/output communication.
• Test Execution: Run tests and validate responses.
• CLI Interface: Provide an interactive testing interface.

Development Process

The MCP Test Client supports the following development process:

1. Develop an MCP server in the playground directory.
2. Use the MCP Test Client to deploy it to the test environment.
3. Test functionality, invoke individual tools, and debug.
4. Fix issues and iterate until the server works properly.
5. Migrate the server to the mcp-servers/ directory.
6. Register with Claude via claude-mcp-local.

Future Enhancements

Planned enhancements include:

• Phase 2: Docker-based container management and comprehensive test suites.
• Phase 3: Migration tools and more advanced test validation.

For the full design document, see notes/mcp_test_client_design.md.