You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Read README.md, PLAN.md, and AGENTS.md first.
Follow the project rules strictly.
Start with Phase 0 and Phase 1 only:
- create the project structure
- create requirements.txt
- implement webcam preview using OpenCV
- keep main.py minimal
- do not implement MediaPipe yet
- do not build UI yet
- do not add unrelated features
After coding, explain:
1. files created/changed
2. how to run
3. expected result
Phase 2 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 2 only:
- implement MediaPipe Face Mesh detection
- create vision/face_mesh.py
- draw face landmarks on webcam preview
- keep code modular
- do not implement scrolling yet
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result
Phase 3 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 3 only:
- implement iris tracking in vision/iris_tracker.py
- calculate iris vertical position relative to the eye region
- print left, right, and average y_ratio
- do not implement scrolling yet
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result
Phase 4 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 4 only:
- implement gaze direction detection in vision/gaze_detector.py
- support LOOK_UP, LOOK_CENTER, LOOK_DOWN, UNKNOWN
- use threshold-based detection
- print current gaze state
- do not implement scrolling yet
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result
Phase 5 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 5 only:
- implement smoothing in utils/smoothing.py
- add moving average
- add hold duration logic
- add cooldown logic
- keep behavior configurable
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result
Phase 6 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 6 only:
- implement scroll engine in control/scroll_engine.py
- use PyAutoGUI only for mouse wheel scrolling
- LOOK_DOWN scrolls down
- LOOK_UP scrolls up
- LOOK_CENTER and UNKNOWN stop
- do not click, type, drag, or perform destructive actions
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result
Phase 7 Prompt
Read README.md, PLAN.md, and AGENTS.md first.
Continue with Phase 7 only:
- implement hotkey toggle in control/hotkeys.py
- use Ctrl + Alt + S to enable/disable LazyScroll
- print enabled/disabled status
- do not build UI yet
After coding, explain:
1. files changed
2. how to run
3. expected result