feat: window-diff-based Type Agent locating for Open Dental appointment flow

Replaces fragile whole-screen text matching with a screenshot-diff step that
finds each dialog's actual pixel bounds, then restricts every subsequent AI
locate/click to that cropped region — eliminating false matches from text
elsewhere on screen (title bars, side panels). Adds column-boundary and
row-height detection so the patient row and Exam procedure click positions
are computed geometrically instead of relying on repeated fuzzy AI guesses
for visually similar neighbors. Also adds per-run screenshot/debug-crop
backups and a structured run.log for diagnosing failed runs, plus a
cmd:move primitive on the Windows agent for pre-click confirmation crops.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

apps/WindowsAgent/agent.py

@@ -107,6 +107,13 @@ class DentalAgent:
             encoded = base64.b64encode(buf.getvalue()).decode("ascii")
             return {"image": encoded}
 
+        @sio.on("cmd:move", namespace="/agent")
+        def on_move(data):
+            # Moves the cursor without clicking — used to double-check a location (via a
+            # follow-up cropped screenshot) before committing to a click.
+            pyautogui.moveTo(data["x"], data["y"])
+            return {"ok": True}
+
         @sio.on("cmd:click", namespace="/agent")
         def on_click(data):
             pyautogui.click(data["x"], data["y"])