finish the testing framework

ROADMAP.md

@@ -28,7 +28,7 @@ Mosis is a **virtual smartphone OS** for VR games and applications. It provides
 | # | Milestone | Status | Description |
 |---|-----------|--------|-------------|
 | 1 | Cross-Platform Kernel | ✅ Complete | Desktop designer with shared kernel code |
-| 2 | Testing Framework | 🔶 80% | Automated UI testing and inspection |
+| 2 | Testing Framework | ✅ Complete | Automated UI testing and inspection |
 | 3 | Virtual Hardware | ❌ Not started | Camera, mic, speaker, filesystem APIs |
 | 4 | App Sandboxing | ❌ Not started | Lua/WASM runtime, package format |
 | 5 | WebRTC Bridge | ❌ Not started | Phone-to-phone communication |
@@ -59,7 +59,7 @@ Mosis is a **virtual smartphone OS** for VR games and applications. It provides
 
 ---
 
-## Milestone 2: Testing Automation 🔶 80% COMPLETE
+## Milestone 2: Testing Automation ✅ COMPLETE
 
 **Goal**: Automated UI testing for rapid iteration and AI agent verification.
 
@@ -74,23 +74,10 @@ Mosis is a **virtual smartphone OS** for VR games and applications. It provides
   - LogParser (navigation event verification)
 - [x] All 5 navigation tests passing
 - [x] Android event injection via ADB broadcast
-
-### Remaining Tasks
-
-- [ ] **Action Recording**
-  - Capture tap, swipe, long_press to JSON
-  - Record timestamps for replay timing
-  - Save to `test-recordings/*.json`
-
-- [ ] **Action Playback**
-  - Load recorded JSON
-  - Replay with timing
-  - Capture results at each step
-
-- [ ] **Screenshot Diff**
-  - Compare two PNG screenshots
-  - Highlight pixel differences
-  - Report diff percentage
+- [x] **Action Recording** - Capture tap, swipe, long_press to JSON with timestamps
+- [x] **Action Playback** - Load recorded JSON and replay with timing
+- [x] **Screenshot Diff** - Pixel-level PNG comparison with configurable tolerance
+- [x] **GLFW Input Hooks** - Automatic mouse/key recording via forked backend
 
 ### Test Recording Format
 
@@ -435,10 +422,11 @@ cmake --build build --config Debug
 
 ## Current Sprint: Complete Partial Tasks
 
-### Priority 1: Testing Framework Completion
-- [ ] Action recording (capture interactions to JSON)
-- [ ] Action playback (replay with timing)
-- [ ] Screenshot diff (visual regression)
+### Priority 1: Testing Framework Completion ✅ DONE
+- [x] Action recording (capture interactions to JSON)
+- [x] Action playback (replay with timing)
+- [x] Screenshot diff (visual regression)
+- [x] GLFW input hooks for automatic recording
 
 ### Priority 2: Remaining System Apps
 - [ ] Store app (UI only - browse, install)
@@ -472,4 +460,4 @@ cmake --build build --config Debug
 
 ---
 
-*Last updated: 2024-01-16*
+*Last updated: 2026-01-16*