Test MCP tools

Test a tool in two steps: discover its advertised schema, then call it with known arguments and check the result that matters to your application.

Requirements

Prepare a Python 3.10+ project with m3 setup. Download the example shipping_server.py and save it as shipping_server.py in the project root. It starts without external services or credentials.

Discover and call the shipping tool

This complete test belongs in tests/test_tools.py:

import sys
from pathlib import Path

import pytest

from m3 import MCPTestKit, StdioServer

pytestmark = pytest.mark.m3(suite_name="shipping-direct")


def test_discover_and_call_shipping_quote() -> None:
    root = Path(__file__).parents[1]
    server = StdioServer(
        name="shipping",
        command=sys.executable,
        args=(str(root / "shipping_server.py"),),
        cwd=str(root),
    )
    with MCPTestKit(env={}) as kit, kit.direct(server) as client:
        tools = client.list_all_tools()
        shipping = next(tool for tool in tools if tool.name == "shipping_quote")
        assert list(shipping.input_schema["required"]) == ["weight_kg", "zone"]
        result = client.call_tool("shipping_quote", {"weight_kg": 2, "zone": "local"})

    assert result.is_error is False
    assert result.structured_content == {"amount": 9.0, "currency": "USD"}

Run it from the example project root:

m3 test -- tests/test_tools.py

list_all_tools() follows pagination until the server has no next page. The test checks that the server advertises both required inputs and returns the expected structured quote. A non-error result by itself would not verify the quote contract.

For invalid inputs and expected tool errors, continue to test errors and schemas. For the server command setup, see stdio servers.