Testing Workflow Agents

pytest plus InMemoryRunner. Everything below uses the published google-adk package — no test-internal helpers.

Setup

bash

uv add "google-adk>=2.0" pytest pytest-asyncio

toml

[tool.pytest.ini_options]
asyncio_mode = "auto"

asyncio_mode = "auto" saves marking every test @pytest.mark.asyncio. Omit it if you prefer explicit marks.

Imports

python

import pytest
from google.adk import Workflow
from google.adk.agents import LlmAgent
from google.adk.apps import App, ResumabilityConfig
from google.adk.events import Event, RequestInput
from google.adk.runners import InMemoryRunner
from google.genai import types

Two helpers worth having

python

async def run(agent, text='hi', app_name='test_app'):
    runner = InMemoryRunner(agent=agent, app_name=app_name)
    session = await runner.session_service.create_session(
        app_name=app_name, user_id='u1'
    )
    msg = types.Content(role='user', parts=[types.Part(text=text)])
    events = []
    async for event in runner.run_async(
        user_id='u1', session_id=session.id, new_message=msg
    ):
        events.append(event)
    return runner, session, events


def node_name(event):
    """'workflow@1/step@1' -> 'step'."""
    if not event.node_info:
        return None
    return event.node_info.path.split('/')[-1].split('@')[0]

event.author is the enclosing workflow's name, not the node's, so filtering on it silently matches the wrong events. event.node_info.path is the one that identifies the node.

A workflow

python

async def test_simple_workflow():
    def step_one(node_input: str) -> str:
        return 'step 1 done'

    def step_two(node_input: str) -> str:
        return 'step 2 done'

    agent = Workflow(
        name='test_workflow', edges=[('START', step_one, step_two)]
    )

    _, _, events = await run(agent)
    final = [e for e in events if node_name(e) == 'step_two' and e.output][-1]
    assert final.output == 'step 2 done'

Routing

python

async def test_routing():
    def router(node_input: str):
        route = 'error' if 'error' in node_input else 'success'
        return Event(output=node_input, route=route)

    agent = Workflow(
        name='routing_test',
        edges=[
            ('START', router),
            (router, {'success': success_handler, 'error': error_handler}),
        ],
    )

    _, _, ok = await run(agent, text='all good')
    assert any(node_name(e) == 'success_handler' for e in ok)

    _, _, err = await run(agent, text='error case')
    assert any(node_name(e) == 'error_handler' for e in err)

Pause and resume

python

async def test_hitl_workflow():
    async def ask_user(ctx, node_input: str):
        yield RequestInput(message='Approve?', interrupt_id='ask')

    def after_approval(node_input) -> str:
        return f'Approved: {node_input}'

    agent = Workflow(
        name='hitl_test', edges=[('START', ask_user, after_approval)]
    )
    app = App(
        name='hitl_test_app',
        root_agent=agent,
        resumability_config=ResumabilityConfig(is_resumable=True),
    )
    runner = InMemoryRunner(app=app)
    session = await runner.session_service.create_session(
        app_name='hitl_test_app', user_id='u1'
    )

    msg = types.Content(role='user', parts=[types.Part(text='start')])
    paused = [
        e
        async for e in runner.run_async(
            user_id='u1', session_id=session.id, new_message=msg
        )
    ]
    fc_events = [e for e in paused if e.get_function_calls()]
    assert fc_events, 'expected an interrupt function call'
    fc = fc_events[-1].get_function_calls()[0]

    response = types.Content(
        role='user',
        parts=[types.Part(function_response=types.FunctionResponse(
            id=fc.id, name=fc.name, response={'result': 'yes'},
        ))],
    )
    resumed = [
        e
        async for e in runner.run_async(
            user_id='u1', session_id=session.id, new_message=response
        )
    ]
    final = [e for e in resumed if node_name(e) == 'after_approval'][-1]
    assert final.output == 'Approved: yes'

State

Prefer reading the session back after the run over inspecting state mid-flight.

python

async def test_state_management():
    def writer(node_input: str):
        return Event(output=node_input, state={'counter': 1})

    def reader(ctx, node_input):
        return f"counter={ctx.state['counter']}"

    agent = Workflow(name='state_test', edges=[('START', writer, reader)])

    runner, session, events = await run(agent)
    final = [e for e in events if node_name(e) == 'reader' and e.output][-1]
    assert final.output == 'counter=1'

    after = await runner.session_service.get_session(
        app_name='test_app', user_id='u1', session_id=session.id
    )
    assert after.state['counter'] == 1

Parallel workers

python

from google.adk.workflow import node


async def test_parallel_worker():
    def produce(node_input: str) -> list:
        return [1, 2, 3]

    @node(parallel_worker=True)
    def double(node_input: int) -> int:
        return node_input * 2

    def collect(node_input: list) -> str:
        return f'results: {node_input}'

    agent = Workflow(
        name='parallel_test', edges=[('START', produce, double, collect)]
    )

    _, _, events = await run(agent)
    final = [e for e in events if node_name(e) == 'collect' and e.output][-1]
    assert final.output == 'results: [2, 4, 6]'

Faking the model

BaseLlm has exactly one abstract method, so a fake is short:

python

from google.adk.models.base_llm import BaseLlm
from google.adk.models.llm_response import LlmResponse


class FakeLlm(BaseLlm):

    def __init__(self, *, responses: list[str]):
        super().__init__(model='fake')
        self._responses = list(responses)

    async def generate_content_async(self, llm_request, stream=False):
        yield LlmResponse(content=types.Content(
            role='model', parts=[types.Part(text=self._responses.pop(0))],
        ))


async def test_llm_agent_with_fake():
    agent = LlmAgent(name='x', model=FakeLlm(responses=['ok']), instruction='Help.')
    _, _, events = await run(agent, text='hi')
    assert events[-1].content.parts[0].text == 'ok'

To assert on the request shape instead, monkeypatch the agent's canonical_model.generate_content_async.

Do not assert on event.output for an LLM agent's own event — the runner clears it before you see it. Assert on the downstream node's output, on session.state[output_key], or on event.content.parts[*].text.

Tests that hit a real model

python

import os

import pytest


@pytest.fixture(scope='session', autouse=True)
def adk_env():
    if 'GOOGLE_API_KEY' not in os.environ:
        pytest.skip('GOOGLE_API_KEY not set')
    os.environ.setdefault('GOOGLE_GENAI_USE_ENTERPRISE', 'FALSE')


@pytest.mark.integration
async def test_real_model():
    ...

pytest -m integration runs them; pytest -m "not integration" skips them.

Habits that avoid flakes

One InMemoryRunner and one session per test — runners carry state.
A unique app_name per test (request.node.name works) so parallel pytest workers do not collide.
event.is_final_response() filters for "the agent's last word".
Any LLM agent feeding a JoinNode needs output_schema=, or the join buffer fails to serialize under DatabaseSessionService.
pytest -xvs while iterating: stop at the first failure, verbose, show prints.