gcp-auditmate

gcp-auditmate-mcp-server is a self-service audit toolkit for Google Cloud Platform built on the Spring AI MCP Server framework and compiled to a GraalVM native image for fast cold starts. It automates scanning and reporting of GCP resources and is designed to produce structured audit outputs and SBOM information to support compliance. The server integrates with Claude Desktop for enhanced conversational capabilities and can be accessed as a local or containerized service. Use it to generate comprehensive reports on your GCP environment, including resource inventories, compliance checks, and risk indicators, all driven by the MCP tooling.

Prerequisites:

• Docker installed on your host (recommended for easiest setup)
• Optional: access to Google Cloud credentials if you want live scans against a GCP project
• Network access to pull the container image

Installation steps (Docker-based):

1. Pull the container image (or use your preferred registry): docker pull samzhu/gcp-auditmate-mcp-server

2. Run the MCP server container: docker run -i samzhu/gcp-auditmate-mcp-server

3. If you need to pass credentials or config, mount environment variables or volumes as needed: docker run -e GCP_CREDENTIALS_JSON='{"type":"service_account",...}'
   -v /path/to/config:/app/config
   -i samzhu/gcp-auditmate-mcp-server

4. Verify the server is up by checking logs or hitting the provided MCP endpoint (see documentation for exact port and endpoint).

Notes:

• The server is described as a GraalVM native image for sub-100 ms cold starts. If you encounter startup issues, ensure your container runtime allows native image execution.
• SBOM is provided for compliance; consider enabling SBOM generation in your deployment if you require it in your build pipeline.
• If integrating with Claude Desktop or other tools, follow the MCP server integration docs for authentication and API usage.
• Provide GCP credentials securely and avoid embedding secrets in images or code.
• If you run locally without Docker, you would adapt to the Java/MCP hosting method described in any project-specific docs (e.g., running a Java-based MCP server directly).