GCP-MCP: Google Cloud Platform Model Control Protocol Server

⚠️ EXPERIMENTAL PROJECT ⚠️

This project provides Model Control Protocol (MCP) servers for interacting with Google Cloud Platform resources.

Overview

GCP-MCP allows AI assistants and other MCP clients to interact with Google Cloud Platform services through standardized interfaces.

Installation

Prerequisites

• uv
• Access to Google Cloud Platform with appropriate permissions
• Google Cloud SDK (for authentication)

Setup

1. Clone this repository:

   git clone https://github.com/manuelbernhardt/gcp-mcp.git
   cd gcp-mcp
   

2. Install dependencies using uv:

   uv sync
   

3. Authenticate with Google Cloud:

   gcloud auth application-default login
   

Available Tools

Secret Manager

• list_secrets: List secrets in a GCP project (optionally filter by prefix)
• delete_secret: Delete a secret from Secret Manager
• add_secret: Add or update a secret in Secret Manager
• get_secret_value: Retrieve the value of a secret

Cloud Run

• list_cloud_run_services: List Cloud Run services in a project and region
• delete_cloud_run_service: Delete a Cloud Run service
• get_cloud_run_service_logs: Fetch recent logs for a Cloud Run service

Configuring for Cursor

Create a mcp.json file in ~/.cursor/mcp.json:

{
  "mcpServers": {
    "gcp": {
      "command": "uv",
      "args": ["--directory", "/path/to/gcp-mcp", "run", "main.py"]
    }
  }
}

Testing

Running Tests

# Run all tests
python -m unittest discover

# Run Secret Manager tests
python -m unittest test_secret_manager.py

# Run Cloud Run tests
python -m unittest test_cloud_run.py

# Run a specific test class
python -m unittest test_secret_manager.TestSecretManagerFunctions
python -m unittest test_cloud_run.TestCloudRunFunctions

# Run a specific test
python -m unittest test_secret_manager.TestSecretManagerFunctions.test_list_secrets