[Mohith26]

Summary

Adds slack_bolt.testing, built-in support for unit testing Bolt apps without connecting to Slack (#380). It follows the approach Bolt's own tests/scenario_tests use (a local mock server plus signed requests), packaged so app developers don't have to rebuild it. No new dependencies; the server uses the standard library.

• MockSlackServer: an in-process HTTP server that stands in for the Web API and for response_url. It records every API call (method, parsed args, token) and every message posted to a response_url. auth.test returns a bot installation by default, other methods return {"ok": true}, and set_response() sets canned responses, including ok: false errors that raise SlackApiError like the real API.
• BoltTestClient(app, server): builds signed requests for events, messages, slash commands, block actions, global and message shortcuts, and view submissions, and sends them through app.dispatch(). It returns a DispatchResult with the ack status/body/json, plus the API calls and respond() bodies that request caused.
• AsyncBoltTestClient (in slack_bolt.testing.async_client, so sync users don't import aiohttp) does the same for AsyncApp.
• slack_bolt.testing.payloads: the payload builders on their own, plus to_http_body() and signed_headers().

Example:

with MockSlackServer() as server:
    app = App(client=server.web_client(), signing_secret="test-secret")

    @app.command("/hello")
    def hello(ack, command, respond):
        ack()
        respond(f"Hi <@{command['user_id']}>!")

    result = BoltTestClient(app, server).slash_command("/hello", user_id="U123")
    assert result.status == 200
    assert result.responds == [{"text": "Hi <@U123>!"}]

Requests go through the full middleware chain (request verification, authorization, listener matching). Two design choices worth flagging:

• To make results deterministic, BoltTestClient runs listeners and lazy listeners inline (an InlineExecutor on the app's listener runner), so side effects are done when the call returns even with process_before_response=False. The async client instead waits for the tasks the dispatch started, with a timeout.
• It points app.client.base_url at the mock server, so apps created at import time with a real token (and token_verification_enabled=False) can be tested without restructuring.

There is a new docs page, docs/english/concepts/testing.md, linked from the sidebar. I posted the design on #380 first; happy to rename things, trim the scope (for example ship only the payload builders and server first), or move it to a separate package if you'd rather not add public API to slack_bolt.

Testing

• New tests: tests/slack_bolt/testing/test_testing.py and tests/slack_bolt_async/testing/test_async_testing.py (19 tests): recorded calls and canned responses, error responses, slash command ack + respond(), say() from a message listener, listeners that finish after ack with process_before_response=False, lazy listeners, block actions and views.open, view submission response_action errors, both shortcut types, 404 for unhandled requests and 401 for a bad signature, 500 for a listener error, an app created with a default client, payload encoding and signatures, and the async timeout.
• pytest tests/slack_bolt tests/slack_bolt_async tests/scenario_tests tests/scenario_tests_async: 849 passed.
• ./scripts/format.sh --no-install, ./scripts/lint.sh --no-install: clean. mypy slack_bolt/testing: no issues. The full ./scripts/run_mypy.sh only reports errors in adapter modules whose optional dependencies I don't have installed.
• I did not run ./scripts/install_all_and_run_tests.sh, so the adapter test suites (which need the optional framework dependencies) were not run locally. This change doesn't touch adapters.

I used an AI assistant (Claude) while writing this; I reviewed and tested all of it.

Category

• slack_bolt.App and/or its core components
• slack_bolt.async_app.AsyncApp and/or its core components
• Adapters in slack_bolt.adapter
• Document pages under /docs
• Others

Requirements

• I've read and understood the Contributing Guidelines and have done my best effort to follow them.
• I've read and agree to the Code of Conduct.
• I've run ./scripts/install_all_and_run_tests.sh after making the changes. (See Testing: core and scenario suites run, adapter suites not run locally.)

[salesforce-cla]

Thanks for the contribution! Before we can merge this, we need @Mohith26 to sign the Salesforce Inc. Contributor License Agreement.

[mwbrooks]

Thanks @Mohith26 for putting this together! 🙇 It's a big addition, so we're discussing it as a team and will follow up here.