Check page band legends against bands.py; amber is verified, not sampled

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

rpa-dashboard-ai/templates/web/bands.py

@@ -5,14 +5,17 @@ an exported sheet, the approve gate) imports from here, so they never disagree
 about what counts as red.
 
     GREEN  >= 0.95   clean print, straight through
-    AMBER  0.90-0.95 slightly unclear, sampled
-    RED    <  0.90   handwriting or a messy scan, always reviewed
-    BLANK            a mandatory field with no value, always flagged
+    AMBER  0.90-0.95 slightly unclear, check it
+    RED    <  0.90   handwriting or a messy scan
+    BLANK            a mandatory field with no value
     UNSCORED         no word scores came back for the value
 
+The approve gate: every field that isn't green must be verified by a person before the
+run can be approved (needs_review). Amber is not sampled; it is checked like red.
+
 Calibration (from the prototype these rules came from): crisp print scored 0.90 and
 up, handwriting clustered in 0.80-0.89, and 0.70-0.79 was empty. 0.90 sits in that
-gap. 0.95 splits off the 0.90-0.99 band that reviewers sample rather than trust.
+gap. 0.95 splits off the slightly unclear 0.90-0.95 reads from clean print.
 
 What a band claims: the score is the model's own legibility rating. Green means the
 text was read cleanly, NOT that the value is correct; a confident misread scores