Protect Sapho priority at the Testing Station

docs/DISCOVERY_PROTOTYPE.md

@@ -64,10 +64,13 @@ In the supported Semuta/Discovery profile, the committed reorder pauses for
 the neutral discard reaction; claim or pass resumes the same next station
 choice. The ordering is a provisional online interpretation of Sapho's
 printed first/last effect; no combined publisher clarification is claimed.
-The current occupant may make its choice before the other occupant plays Sapho,
-including when an automated seat acts first. There is no neutral pre-choice
-priority window yet, so first-position intervention is not guaranteed against
-an automatically acting current occupant.
+New two-occupant Station decisions with Richese seated open a neutral shared
+priority before either adjustment. Each occupant may continue or use its own
+eligible Sapho; an automated occupant passes without consuming the original
+station choice. The window is independent of private card custody and closes
+when both pass or Sapho is used. Already-open unstamped choices retain their
+existing timing. Replay of a recovered Sapho and broader combinations remain
+unfinished.
 
 13 September 2026. The connected Discovery batches are **Prototyped**, with
 partial rules coverage. It connects physical components, real actions, private