Workload Manager API

You run an MCP Server that exposes a Model Context Protocol interface for the Google Workload Manager API. This server lets MCP clients interact with the API in a standardized way, enabling you to query, update, and manage workload resources through MCP-compatible requests while keeping your tooling decoupled from direct API calls.

How to use

Start the MCP server using a local, self-contained process. You will run a single command that launches the server in stdio mode and then connect your MCP client to it. The server accepts configuration through environment variables or a JSON config file so you can tailor security and behavior to your environment.

How to install

Prerequisites you need before installing and running: Python 3.9 or newer, and the Python package installer (pip) along with the uv runtime for optional execution. You will install dependencies and then start the server.

# Step 1: Install dependencies
pip install -e ".[dev]"

# If you prefer, install via uv
uv pip install --editable ".[dev]"

# Step 2: Run static checks and tests locally when desired
ruff check
ruff format
./scripts/static-analysis.sh
./scripts/test.sh

Additional configuration notes

The server is started with a standard Python invocation in stdio mode. You can also influence runtime behavior using environment variables. The following start command is the definitive runtime for launching the server locally.

python mcp_server/main.py stdio

Available tools

ruff

Linting and formatting tool used to ensure code quality and consistency.

mypy

Static type checking tool to catch type errors before runtime.

pytest

Testing framework used to run unit and integration tests.

pre-commit

Hook-based checks that run before every commit to enforce quality gates.

hatch

Build and publish tool used to package the MCP server.

static-analysis

Script that runs static analysis tools like mypy, bandit, and semgrep for security and quality checks.