|
| 1 | +# Flutter Skill Roadmap |
| 2 | + |
| 3 | +> Comprehensive optimization and multi-framework expansion plan for Flutter Skill as an AI E2E testing platform. |
| 4 | +
|
| 5 | +--- |
| 6 | + |
| 7 | +## Phase 1: Core Flutter Optimization (v0.7.1) |
| 8 | + |
| 9 | +### Smart Screenshot Optimization |
| 10 | +- Auto-compress screenshots for AI vision models (reduce token cost 3-5x) |
| 11 | +- Configurable quality/resolution presets: `screenshot({ quality: "ai" })` |
| 12 | +- Auto-crop status bar and navigation bar |
| 13 | +- Delta screenshots (only send changed regions) |
| 14 | + |
| 15 | +### Intelligent Wait System |
| 16 | +- Replace fixed timeouts with smart polling: `wait_for_stable()` waits until UI stops changing |
| 17 | +- Auto-detect animations/transitions and wait for completion |
| 18 | +- Network-aware waiting (wait for HTTP requests to complete) |
| 19 | + |
| 20 | +### Enhanced Error Recovery |
| 21 | +- Auto-reconnect on VM Service disconnection |
| 22 | +- Graceful handling of app crashes with diagnostic info |
| 23 | +- Retry logic for flaky operations (tap during animation, etc.) |
| 24 | + |
| 25 | +### Semantic Element Discovery |
| 26 | +- Find elements by semantic meaning: `find({ role: "submit_button" })` instead of exact key/text |
| 27 | +- Fuzzy text matching: `tap({ text: "Submit" })` matches "SUBMIT", "Submit Order", etc. |
| 28 | +- AI-friendly element descriptions in inspect output |
| 29 | + |
| 30 | +--- |
| 31 | + |
| 32 | +## Phase 2: Advanced Testing Features (v0.7.2) |
| 33 | + |
| 34 | +### Visual Regression Testing |
| 35 | +- `screenshot_compare({ baseline: "login_screen" })` - pixel-diff with threshold |
| 36 | +- Auto-generate baselines on first run |
| 37 | +- Highlight visual differences in returned image |
| 38 | +- Golden file management (save/update/compare) |
| 39 | + |
| 40 | +### Test Recording & Playback |
| 41 | +- `start_recording()` / `stop_recording()` - record all interactions |
| 42 | +- Export as reusable test script (Dart test, JSON, or natural language) |
| 43 | +- Parameterized playback with variable substitution |
| 44 | + |
| 45 | +### Performance Monitoring Integration |
| 46 | +- Frame rate tracking during interactions |
| 47 | +- Memory usage snapshots before/after operations |
| 48 | +- Jank detection and reporting |
| 49 | +- Network request timing correlation |
| 50 | + |
| 51 | +### Accessibility Validation |
| 52 | +- `check_accessibility()` - verify semantic labels, contrast ratios, touch targets |
| 53 | +- WCAG compliance checking |
| 54 | +- Screen reader simulation output |
| 55 | + |
| 56 | +--- |
| 57 | + |
| 58 | +## Phase 3: Native Platform Interaction (v0.7.3) |
| 59 | + |
| 60 | +### Native View Support (iOS Simulator) |
| 61 | +- `native_screenshot` - `xcrun simctl io screenshot` for native dialogs (photo picker, permission dialogs) |
| 62 | +- `native_tap` - macOS Accessibility API for tapping native UI elements |
| 63 | +- `native_input_text` - `simctl pbcopy` + paste for text input in native views |
| 64 | +- `native_swipe` - Accessibility API scroll actions for native scroll views |
| 65 | + |
| 66 | +### Native View Support (Android Emulator) |
| 67 | +- `native_screenshot` - `adb shell screencap` for native views |
| 68 | +- `native_tap` - `adb shell input tap x y` |
| 69 | +- `native_input_text` - `adb shell input text` |
| 70 | +- `native_swipe` - `adb shell input swipe` |
| 71 | + |
| 72 | +### Hybrid Detection |
| 73 | +- Auto-detect when native view is presented (VM Service becomes unresponsive) |
| 74 | +- Seamless fallback: try VM Service first, fall back to native driver |
| 75 | +- Unified API - same `tap()` / `screenshot()` works for both Flutter and native |
| 76 | + |
| 77 | +--- |
| 78 | + |
| 79 | +## Phase 4: Multi-Framework Support (v0.7.4) |
| 80 | + |
| 81 | +### Architecture: Universal App Driver |
| 82 | + |
| 83 | +``` |
| 84 | +AbstractAppDriver (interface) |
| 85 | +├── FlutterDriver (existing - VM Service Protocol) |
| 86 | +├── ReactNativeDriver (new - Chrome DevTools Protocol / Hermes) |
| 87 | +├── NativeIOSDriver (new - XCTest / Accessibility API) |
| 88 | +├── NativeAndroidDriver (new - UIAutomator / adb) |
| 89 | +├── WebDriver (new - Chrome DevTools Protocol) |
| 90 | +└── ElectronDriver (new - Chrome DevTools Protocol + Node.js) |
| 91 | +``` |
| 92 | + |
| 93 | +### React Native Support |
| 94 | +- **Protocol**: Chrome DevTools Protocol (CDP) via Hermes debugger |
| 95 | +- **Connection**: `ws://localhost:8081/debugger-proxy` (Metro bundler) |
| 96 | +- **UI Inspection**: React DevTools protocol for component tree |
| 97 | +- **Actions**: CDP `Input.dispatchTouchEvent` / `Input.dispatchKeyEvent` |
| 98 | +- **Screenshots**: CDP `Page.captureScreenshot` |
| 99 | + |
| 100 | +### Native iOS Support |
| 101 | +- **Protocol**: XCTest framework + Accessibility API |
| 102 | +- **Connection**: `xcrun simctl` for simulator, `idevice` tools for physical devices |
| 103 | +- **UI Inspection**: Accessibility tree traversal |
| 104 | +- **Actions**: XCTest gestures or Accessibility API `AXPress` |
| 105 | +- **Screenshots**: `xcrun simctl io screenshot` |
| 106 | + |
| 107 | +### Native Android Support |
| 108 | +- **Protocol**: UIAutomator2 + adb |
| 109 | +- **Connection**: `adb` for emulator and physical devices |
| 110 | +- **UI Inspection**: `uiautomator dump` for view hierarchy |
| 111 | +- **Actions**: `adb shell input` for tap/swipe/text |
| 112 | +- **Screenshots**: `adb shell screencap` |
| 113 | + |
| 114 | +### Web App Support |
| 115 | +- **Protocol**: Chrome DevTools Protocol (CDP) |
| 116 | +- **Connection**: `ws://localhost:9222` (Chrome remote debugging) |
| 117 | +- **UI Inspection**: DOM traversal via CDP |
| 118 | +- **Actions**: CDP Input domain |
| 119 | +- **Screenshots**: CDP `Page.captureScreenshot` |
| 120 | + |
| 121 | +### Unified MCP Interface |
| 122 | + |
| 123 | +All frameworks share the same MCP tool names: |
| 124 | + |
| 125 | +| Tool | Flutter | React Native | Native iOS | Native Android | Web | |
| 126 | +|------|---------|-------------|------------|---------------|-----| |
| 127 | +| `connect_app` | VM Service | Metro/CDP | XCTest | adb | CDP | |
| 128 | +| `screenshot` | VM Service | CDP | simctl/idevice | screencap | CDP | |
| 129 | +| `tap` | Extension | CDP touch | Accessibility | input tap | CDP | |
| 130 | +| `inspect` | Extension | React DevTools | Accessibility tree | uiautomator | DOM | |
| 131 | +| `enter_text` | Extension | CDP key | Accessibility | input text | CDP | |
| 132 | + |
| 133 | +### Auto-Detection |
| 134 | + |
| 135 | +``` |
| 136 | +scan_and_connect() |
| 137 | + → Detect running apps across all frameworks |
| 138 | + → Return: [ |
| 139 | + { framework: "flutter", uri: "ws://...:50000/ws", name: "MyFlutterApp" }, |
| 140 | + { framework: "react-native", uri: "ws://...:8081/debugger-proxy", name: "MyRNApp" }, |
| 141 | + { framework: "web", uri: "ws://...:9222", name: "localhost:3000" } |
| 142 | + ] |
| 143 | + → Connect to selected app with appropriate driver |
| 144 | +``` |
| 145 | + |
| 146 | +--- |
| 147 | + |
| 148 | +## Version Summary |
| 149 | + |
| 150 | +| Version | Focus | Key Deliverables | |
| 151 | +|---------|-------|-----------------| |
| 152 | +| **v0.7.1** | Core Flutter Optimization | Smart screenshots, intelligent waits, error recovery, semantic discovery | |
| 153 | +| **v0.7.2** | Advanced Testing | Visual regression, test recording, performance monitoring, accessibility | |
| 154 | +| **v0.7.3** | Native Platform Interaction | iOS/Android native view support, hybrid detection, unified API | |
| 155 | +| **v0.7.4** | Multi-Framework | React Native, native iOS/Android, Web, Electron support | |
| 156 | + |
| 157 | +--- |
| 158 | + |
| 159 | +## Priority Matrix |
| 160 | + |
| 161 | +| Feature | Impact | Effort | Priority | |
| 162 | +|---------|--------|--------|----------| |
| 163 | +| Smart screenshots | High (token cost) | Low | P0 | |
| 164 | +| Intelligent waits | High (reliability) | Medium | P0 | |
| 165 | +| Error recovery | High (stability) | Medium | P0 | |
| 166 | +| Semantic discovery | High (usability) | Medium | P1 | |
| 167 | +| Native view support | High (coverage) | High | P1 | |
| 168 | +| Visual regression | Medium (testing) | Medium | P1 | |
| 169 | +| React Native support | High (market) | High | P2 | |
| 170 | +| Web support | Medium (market) | Medium | P2 | |
| 171 | +| Test recording | Medium (DX) | Medium | P2 | |
| 172 | +| Accessibility | Medium (quality) | Low | P2 | |
| 173 | +| Performance monitoring | Low (niche) | Medium | P3 | |
| 174 | +| Electron support | Low (niche) | Low | P3 | |
0 commit comments