---

name: mobile-testing description: > Use this skill when the user asks about mobile testing strategies, unit tests, widget tests, component tests, integration tests, E2E, golden/snapshot tests, mocking, or CI test integration. version: "1.0.0" author: "j4flmao" license: "MIT" compatibility: claude-code: true cursor: true codex: true windsurf: true tags: [mobile, testing, phase-4, universal]

Mobile Testing

Purpose

Design comprehensive mobile test strategies following the test pyramid — unit, widget/component, integration, and E2E — with proper mocking, golden tests, and CI integration.

Agent Protocol

Trigger

User request includes: mobile test, mobile testing, unit test mobile, widget test, component test, ui test mobile, e2e mobile, golden test, snapshot test mobile.

Input Context

• Platform (iOS, Android, Flutter, React Native)
• Testing framework (XCTest, JUnit, flutter_test, Jest)
• CI provider (GitHub Actions, Bitrise, GitLab CI)

Output Artifact

A markdown document containing:

• Test strategy (pyramid)
• Framework setup
• Key test examples
• CI pipeline integration

Response Format

No preamble. No postamble. No explanations.

Max Response Length

4096 tokens

Decision Trees

Test Type Selection

What are you testing?
├── Business logic (UseCase, calculator, validator)
│   └── Unit test — pure logic, no UI, fast, deterministic
├── ViewModel state management (state transitions, events)
│   └── Unit test with mocked dependencies — no UI rendering
├── UI component rendering (displays correct data in states)
│   ├── Flutter → Widget test (pumpWidget, find, expect)
│   ├── SwiftUI → ViewInspector / snapshot tests
│   ├── Jetpack Compose → compose-test with semijson tree
│   └── React Native → React Native Testing Library
├── Visual regression (design fidelity)
│   └── Golden/snapshot test — compare rendered UI to reference image
├── Feature flow (login → navigate → action)
│   ├── Integration test — multiple components, mocked API/DB
│   └── Avoid full E2E for speed
└── Critical user journey (checkout, signup, payment)
    └── E2E test — real device/emulator, real API/staging

Mock Strategy

What needs mocking?
├── Network calls → Mock API responses (Mockito/mocktail/OHHTTPStubs)
├── Database/DAO → In-memory database or mock DAO
├── Platform channels (Flutter) → Mock channel handlers
├── System services (location, camera, biometrics)
│   ├── Mock per test (unit) OR fake service (integration)
│   └── Avoid mocking what you don't own (platform SDKs)
└── Time/date → Inject clock dependency, use fake timers

CI Test Strategy

What CI stage?
├── Every PR (fast, <10min) → Unit + widget/component tests
├── Nightly (slow, <60min) → Integration tests + golden tests
├── Pre-release (manual trigger) → Full E2E suite on device farm
└── Post-release → Smoke tests on production (canary)

Workflow

Step 1: Define Test Strategy

Apply the mobile test pyramid: many unit/widget tests, medium integration tests, few E2E critical journeys.

Step 2: Write Unit Tests

Cover business logic, use cases, repositories, and utilities with mocked dependencies and parameterized edge cases.

Step 3: Write Widget/Component Tests

Test UI components with mock data, verify rendering, interactions, and state transitions.

Step 4: Write Integration Tests

Test feature flows combining multiple components, real API mocking, and database interactions.

Step 5: Write E2E Tests

Cover critical user journeys (login, purchase, order flow) on real devices or emulators.

Step 6: Integrate in CI

Run unit + widget tests on every PR, integration nightly, E2E before release.

Rules

• Unit tests cover business logic — never test framework behavior
• Mock external dependencies (network, DB, platform channels) in unit tests
• One test file per source file — mirror the source structure
• Widget tests verify states: loading, error, empty, data
• Golden/snapshot tests for visual regression on critical screens
• E2E tests are for critical journeys only — max 5-10 per platform
• Tests must be deterministic — no time-dependent assertions without mocking time
• Code coverage threshold: 80% unit, 70% widget, 60% integration
• Never test private methods directly — test through public API
• Test boundary conditions: empty lists, null values, max values

Test Pyramid

      ╱╲
     ╱E2E╲          Few — critical journeys only (5-10)
    ╱─────╲
   ╱Integration╲    Medium — feature flows (20-50)
  ╱─────────────╲
╱  Unit / Widget  ╲ Many — every class, every widget state (200+)
╱─────────────────╲

Unit Tests

Android — JUnit + MockK

class GetOrdersUseCaseTest {
  private val repository = mockk<OrderRepository>()
  private val useCase = GetOrdersUseCase(repository)

  @Test
  fun `returns orders from repository`() = runTest {
    val orders = listOf(Order("1", "Test"))
    coEvery { repository.getOrders() } returns Result.success(orders)

    val result = useCase()

    assertTrue(result.isSuccess)
    assertEquals(orders, result.getOrNull())
  }

  @Test
  fun `returns error when repository fails`() = runTest {
    coEvery { repository.getOrders() } returns Result.failure(NetworkException("timeout"))

    val result = useCase()

    assertTrue(result.isFailure)
    assertTrue(result.exceptionOrNull() is NetworkException)
  }

  @Test
  fun `handles empty list`() = runTest {
    coEvery { repository.getOrders() } returns Result.success(emptyList())

    val result = useCase()

    assertTrue(result.isSuccess)
    assertTrue(result.getOrNull()!!.isEmpty())
  }
}

iOS — XCTest

import XCTest
@testable import MyApp

final class GetOrdersUseCaseTests: XCTestCase {
  var useCase: GetOrdersUseCase!
  var mockRepository: MockOrderRepository!

  override func setUp() {
    mockRepository = MockOrderRepository()
    useCase = GetOrdersUseCase(repository: mockRepository)
  }

  func testReturnsOrders() async throws {
    let expected = [Order(id: "1", title: "Test")]
    mockRepository.stubbedOrders = expected

    let orders = try await useCase()
    XCTAssertEqual(orders, expected)
  }

  func testThrowsOnNetworkError() async {
    mockRepository.stubbedError = NetworkError.timeout

    await XCTAssertThrowsError(try await useCase())
  }
}

Flutter — flutter_test + mocktail

import 'package:flutter_test/flutter_test.dart';
import 'package:mocktail/mocktail.dart';

class MockOrderRepo extends Mock implements OrderRepository {}

void main() {
  late GetOrdersUseCase useCase;
  late MockOrderRepo repo;

  setUp(() {
    repo = MockOrderRepo();
    useCase = GetOrdersUseCase(repo);
  });

  test('returns orders', () async {
    final orders = [Order(id: '1', title: 'Test')];
    when(() => repo.getOrders()).thenAnswer((_) async => orders);

    final result = await useCase();

    expect(result, orders);
  });

  test('throws on error', () async {
    when(() => repo.getOrders()).thenThrow(NetworkException('timeout'));

    expect(() => useCase(), throwsA(isA<NetworkException>()));
  });
}

React Native — Jest

import { renderHook, act } from '@testing-library/react-hooks';
import { useOrderStore } from './orderStore';

jest.mock('../api/orderApi');

describe('useOrderStore', () => {
  it('loads orders', async () => {
    (orderApi.getOrders as jest.Mock).mockResolvedValue(mockOrders);

    const { result } = renderHook(() => useOrderStore());

    await act(async () => {
      await result.current.loadOrders();
    });

    expect(result.current.orders).toEqual(mockOrders);
    expect(result.current.isLoading).toBe(false);
  });

  it('handles error state', async () => {
    (orderApi.getOrders as jest.Mock).mockRejectedValue(new Error('Network error'));

    const { result } = renderHook(() => useOrderStore());

    await act(async () => {
      await result.current.loadOrders();
    });

    expect(result.current.error).toBe('Network error');
    expect(result.current.isLoading).toBe(false);
  });
});

ViewModel Tests

Android — ViewModel Test

@OptIn(ExperimentalCoroutinesApi::class)
class OrderViewModelTest {
  @get:Rule
  val mainDispatcherRule = MainDispatcherRule()

  private val useCase = mockk<GetOrdersUseCase>()
  private lateinit var vm: OrderViewModel

  @Before
  fun setup() {
    vm = OrderViewModel(useCase)
  }

  @Test
  fun `load orders sets loading then data`() = runTest {
    val orders = listOf(Order("1", "Test"))
    coEvery { useCase() } returns Result.success(orders)

    vm.onAction(OrderAction.Refresh)

    // Loading state
    assertEquals(true, vm.state.value.isLoading)
    // Advance past loading
    advanceUntilIdle()
    // Data state
    assertEquals(orders, vm.state.value.orders)
    assertEquals(false, vm.state.value.isLoading)
  }

  @Test
  fun `load orders sets error on failure`() = runTest {
    coEvery { useCase() } returns Result.failure(NetworkException("timeout"))

    vm.onAction(OrderAction.Refresh)
    advanceUntilIdle()

    assertNotNull(vm.state.value.error)
    assertEquals(false, vm.state.value.isLoading)
  }
}

iOS — ViewModel Test

@MainActor
final class OrderViewModelTests: XCTestCase {
  var vm: OrderViewModel!
  var mockUseCase: MockGetOrdersUseCase!

  override func setUp() {
    mockUseCase = MockGetOrdersUseCase()
    vm = OrderViewModel(getOrdersUseCase: mockUseCase)
  }

  func testLoadSetsOrders() async {
    let expected = [Order(id: "1", title: "Test")]
    mockUseCase.stubbedResult = expected

    await vm.loadOrders()

    XCTAssertEqual(vm.orders, expected)
    XCTAssertFalse(vm.isLoading)
  }

  func testLoadSetsError() async {
    mockUseCase.stubbedError = NetworkError.timeout

    await vm.loadOrders()

    XCTAssertNotNil(vm.error)
    XCTAssertFalse(vm.isLoading)
  }
}

Widget / Component Tests

Flutter — Widget Test

testWidgets('OrderCard displays all states', (tester) async {
  // Data state
  await tester.pumpWidget(MaterialApp(home: OrderCard(order: tOrder)));
  expect(find.text('Alice'), findsOneWidget);
  expect(find.text('\$50.00'), findsOneWidget);

  // Loading state
  await tester.pumpWidget(MaterialApp(home: OrderCard(isLoading: true)));
  expect(find.byType(CircularProgressIndicator), findsOneWidget);

  // Error state
  await tester.pumpWidget(MaterialApp(home: OrderCard(error: 'Failed to load')));
  expect(find.text('Failed to load'), findsOneWidget);

  // Empty state
  await tester.pumpWidget(MaterialApp(home: OrderCard(order: null)));
  expect(find.text('No order found'), findsOneWidget);
});

testWidgets('OrderCard tap triggers callback', (tester) async {
  var tapped = false;
  await tester.pumpWidget(MaterialApp(
    home: OrderCard(order: tOrder, onTap: () => tapped = true),
  ));
  await tester.tap(find.text('Alice'));
  expect(tapped, isTrue);
});

React Native — Testing Library

import { render, fireEvent } from '@testing-library/react-native';

describe('OrderCard', () => {
  it('renders order data', () => {
    const { getByText } = render(<OrderCard order={mockOrder} />);
    expect(getByText('Alice')).toBeOnTheScreen();
    expect(getByText('$50.00')).toBeOnTheScreen();
  });

  it('shows loading spinner when loading', () => {
    const { getByTestId } = render(<OrderCard isLoading />);
    expect(getByTestId('loading-spinner')).toBeOnTheScreen();
  });

  it('calls onPress when tapped', () => {
    const onPress = jest.fn();
    const { getByText } = render(<OrderCard order={mockOrder} onPress={onPress} />);
    fireEvent.press(getByText('Alice'));
    expect(onPress).toHaveBeenCalledTimes(1);
  });
});

Jetpack Compose — compose-ui-test

@Test
fun orderCardDisplaysCustomerName() {
  composeTestRule.setContent {
    OrderCard(order = testOrder)
  }
  composeTestRule.onNodeWithText("Alice").assertIsDisplayed()
  composeTestRule.onNodeWithText("$50.00").assertIsDisplayed()
}

@Test
fun orderCardTapTriggersCallback() {
  var tapped = false
  composeTestRule.setContent {
    OrderCard(order = testOrder, onTap = { tapped = true })
  }
  composeTestRule.onNodeWithText("Alice").performClick()
  assertTrue(tapped)
}

Golden / Snapshot Tests

Flutter — Golden Test

testWidgets('OrderCard golden test', (tester) async {
  await tester.binding.setSurfaceSize(const Size(400, 200));
  await tester.pumpWidget(MaterialApp(home: OrderCard(order: tOrder)));
  await expectLater(
    find.byType(OrderCard),
    matchesGoldenFile('goldens/order_card.png'),
  );
});

iOS — iOSSnapshotTestCase

func testOrderCardSnapshot() {
  let card = OrderCard(order: testOrder)
  let vc = UIHostingController(rootView: card)
  vc.view.frame = CGRect(x: 0, y: 0, width: 400, height: 200)
  FBSnapshotVerifyView(vc.view)
}

Android — Paparazzi

@Test
fun orderCardSnapshot() {
  paparazzi.snapshot {
    OrderCard(order = testOrder)
  }
}

Integration Tests

Flutter — Integration Test

// test_driver/app_test.dart
void main() {
  IntegrationTestWidgetsFlutterBinding.ensureInitialized();

  testWidgets('create order flow', (tester) async {
    await tester.pumpWidget(MyApp());
    await tester.tap(find.byKey(const Key('add-order')));
    await tester.enterText(find.byKey(const Key('name-input')), 'Alice');
    await tester.tap(find.byKey(const Key('save')));
    await tester.pumpAndSettle();
    expect(find.text('Alice'), findsOneWidget);
  });
}

Android — Compose Test (Integration)

@Test
fun createOrderFlow() {
  composeTestRule.setContent {
    MyApp()
  }
  composeTestRule.onNodeWithTag("add-order").performClick()
  composeTestRule.onNodeWithTag("name-input").performTextInput("Alice")
  composeTestRule.onNodeWithTag("save").performClick()
  composeTestRule.onNodeWithText("Alice").assertIsDisplayed()
}

E2E Tests

Detox (React Native)

describe('Order Flow', () => {
  beforeAll(async () => {
    await device.launchApp({ newInstance: true });
  });

  it('creates order end-to-end', async () => {
    await element(by.id('add-order')).tap();
    await element(by.id('name-input')).typeText('Alice\n');
    await element(by.id('save')).tap();
    await expect(element(by.text('Alice'))).toBeVisible();
  });

  it('handles network error gracefully', async () => {
    await device.setURLBlacklist(['.*api.example.com.*']);
    await element(by.id('add-order')).tap();
    await element(by.id('save')).tap();
    await expect(element(by.text('Network error'))).toBeVisible();
  });
});

XCUITest (iOS)

final class OrderFlowTests: XCTestCase {
  let app = XCUIApplication()

  override func setUp() {
    continueAfterFailure = false
    app.launch()
  }

  func testCreateOrder() {
    app.buttons["add-order"].tap()
    let nameField = app.textFields["name-input"]
    nameField.tap()
    nameField.typeText("Alice")
    app.buttons["save"].tap()
    XCTAssertTrue(app.staticTexts["Alice"].waitForExistence(timeout: 5))
  }

  func testEmptyNameShowsError() {
    app.buttons["add-order"].tap()
    app.buttons["save"].tap()
    XCTAssertTrue(app.staticTexts["Name is required"].exists)
  }
}

Espresso (Android) + Compose

@Test
fun createOrderE2E() {
  composeTestRule.onNodeWithTag("add-order").performClick()
  composeTestRule.onNodeWithTag("name-input").performTextInput("Alice")
  composeTestRule.onNodeWithTag("save").performClick()
  composeTestRule.onNodeWithText("Alice").assertIsDisplayed()
}

Maestro (Cross-platform E2E)

# .maestro/create-order.yaml
appId: com.example.app
---
- tapOn: "add-order"
- inputText: "Alice"
- tapOn: "save"
- assertVisible: "Alice"

CI Integration

GitHub Actions — Flutter

name: Tests
on: [pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: subosito/flutter-action@v2
        with:
          flutter-version: '3.x'
      - run: flutter pub get
      - run: flutter analyze
      - run: flutter test --coverage --machine > report.log
      - uses: VeryGoodOpenSource/very_good_coverage@v2
        with:
          path: ./coverage/lcov.info
          min_coverage: 80

GitHub Actions — Android

name: Android Tests
on: [pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-java@v4
        with:
          distribution: 'temurin'
          java-version: '17'
      - run: ./gradlew testDebugUnitTest
      - run: ./gradlew ktlintCheck
      - uses: actions/upload-artifact@v4
        with:
          path: app/build/reports/tests

GitHub Actions — iOS

name: iOS Tests
on: [pull_request]
jobs:
  test:
    runs-on: macos-14
    steps:
      - uses: actions/checkout@v4
      - run: xcodebuild test -scheme App -destination 'platform=iOS Simulator,name=iPhone 16'
      - run: swiftlint

GitHub Actions — React Native

name: RN Tests
on: [pull_request]
jobs:
  test:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: actions/setup-node@v4
      - run: npm ci
      - run: npm test -- --coverage
      - run: npx eslint .

Code Coverage

Configuration

// Android — build.gradle.kts
android {
  testOptions {
    unitTests.all {
      jacoco {
        includeNoLocationClasses = true
      }
    }
  }
}

tasks.register("jacocoTestReport", JacocoReport::class) {
  dependsOn("testDebugUnitTest")
  reports { xml.required.set(true) }
}

// iOS — scheme configuration
// Edit Scheme > Test > Options > Code Coverage
// Or xccov:
// xcrun xccov view --report --json DerivedData/*.xcresult

Threshold Enforcement

# Flutter: very_good_coverage
min_coverage: 80
exclude:
  - "**/*.g.dart"
  - "**/*.freezed.dart"
  - "**/generated/**"

Testing Anti-Patterns

• Testing implementation details: Tests break when refactoring. Test behavior, not internal state
• Over-mocking: Mocking 10+ dependencies means the test verifies mocks, not real code. Use fakes for complex deps
• No edge case coverage: Only testing happy path. Include empty, error, loading, boundary, and timeout cases
• Flaky tests from shared mutable state: Tests leak state between runs. Use setUp/tearDown for clean state
• UI tests that depend on timing: waitFor instead of sleep. Never use fixed delays
• Endless E2E tests: 100 E2E tests = CI takes 2 hours. Keep 5-10 critical paths only
• No golden test review process: Golden images merged blindly catch no regressions. Require human review
• Skipping widget tests for "fast iteration": UI regressions multiply. At minimum test critical widgets
• Testing framework code: Verifying Room DAO SQL generation or Compose rendering is framework's job
• No CI test enforcement: Tests that pass locally but not on CI are worse than no tests. Enforce on every PR
• Gitignoring test reports: CI test failures need debugging. Upload reports as CI artifacts
• Ignoring flaky tests: Marking flaky tests as "known" lets them pile up. Fix or remove

Performance Testing

// Android — Macrobenchmark for startup/scrolling
@RunWith(AndroidJUnit4::class)
class StartupBenchmark {
  @get:Rule
  val benchmarkRule = MacrobenchmarkRule()

  @Test
  fun startup() = benchmarkRule.measureRepeated(
    packageName = "com.example.app",
    metrics = listOf(StartupTimingMetric()),
    iterations = 5,
    startupMode = StartupMode.COLD,
  ) {
    startActivityAndWait()
  }
}

// iOS — XCTMetric for performance
func testScrollPerformance() {
  let metric = XCTOSSignpostMetric.scrolling
  measure(metrics: [metric]) {
    app.tables.element.swipeUp()
  }
}

Accessibility Testing

testWidgets('OrderCard has accessibility labels', (tester) async {
  await tester.pumpWidget(MaterialApp(home: OrderCard(order: tOrder)));
  expect(find.bySemanticsLabel('Order for Alice'), findsOneWidget);
});

func testAccessibilityLabels() {
  XCTAssertTrue(app.buttons["add-order"].exists)
  XCTAssertTrue(app.staticTexts["order-status"].label == "Pending")
}

References

• references/integration-testing.md — Mobile Integration Testing
• references/manual-testing.md — Manual Testing — Mobile
• references/mobile-testing-strategies.md — Mobile Testing Strategies
• references/test-automation-frameworks.md — Test Automation Frameworks
• references/ui-testing.md — Mobile UI Testing
• references/unit-testing.md — Mobile Unit Testing

Handoff

After testing setup, hand off to:

• mobile/universal/performance — Performance benchmarking tests
• mobile/universal/security — Security testing, penetration testing
• mobile/universal/deployment — CI/CD pipeline integration for tests
• mobile/universal/analytics — Test analytics for flaky detection
• mobile/android — Espresso, Robolectric, Compose test
• mobile/ios — XCUITest, XCTest, snapshot testing
• mobile/flutter — flutter_test, integration_test
• mobile/react-native — Jest, Detox, React Native Testing Library