Align recognition review intervention continuity

UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingRepositoryLibrary.cpp

@@ -29651,7 +29651,13 @@ FHyperTwistTrainingReviewFlowStatus UHyperTwistTrainingRepositoryLibrary::Derive
 		}
 	}
 
-	if (ReviewPlan.Policy.PolicyKind == EHyperTwistTrainingReviewPolicyKind::CoachReview && FlowStatus.RepeatCount > 0)
+	const bool bCoachFacingRepeatPressure =
+		FlowStatus.RepeatCount > 0
+		&& (
+			ReviewPlan.Policy.PolicyKind == EHyperTwistTrainingReviewPolicyKind::CoachReview
+			|| ReviewPlan.Policy.bPreferRecognitionReplayReview
+		);
+	if (bCoachFacingRepeatPressure)
 	{
 		FlowStatus.bNeedsCoachIntervention = true;
 	}

docs/arch/HYPERTWIST_PHASE4_RECOGNITION_REVIEW_INTERVENTION_CONTINUITY_PACKET_2026-05-06.md

@@ -0,0 +1,93 @@
+# HyperTwist Phase 4 recognition review intervention continuity packet
+
+Created on `2026-05-06`
+
+Status:
+
+- first-party HyperTwist packet
+- bounded Phase `4` coach-to-review runtime validation slice
+
+## Purpose
+
+This packet closes a review-flow continuity gap inside the now-live recognition-assisted coach-to-review loop.
+
+The open task is:
+
+- make recognition-biased review plans escalate into the same coach-intervention lane as explicit coach-review plans when repeat pressure appears
+
+It is not:
+
+- a new recognition transport packet
+- a new queue persistence packet
+- a new dashboard packet
+- a broad review-system rewrite
+
+## Scope
+
+Bounded lane:
+
+- adjust review-flow derivation so repeat pressure triggers coach intervention when the active review policy is either:
+  - explicit `CoachReview`
+  - recognition-biased through `bPreferRecognitionReplayReview`
+- let existing review-program summary, coach brief, and dashboard inspection derive the corrected state through the already-landed continuity path
+
+Out of scope:
+
+- new review-plan schema
+- new coach-signal heuristics
+- replay scoring changes
+- broader UI changes
+
+## Why this was the right next packet
+
+Before this slice:
+
+- recognition replay pressure already flowed into coach signals and briefs
+- recognition-review preference already persisted through queue materialization and active review policy
+- dashboard inspection already surfaced active review policy and intervention state
+
+But one live derivation still lagged behind:
+
+- review-flow intervention treated repeat pressure as coach-facing only when policy kind was explicit `CoachReview`
+
+That meant a recognition-biased review plan could:
+
+- carry recognition-review preference correctly
+- accumulate repeat pressure correctly
+- still fail to trip the same coach-intervention lane unless it also used the explicit `CoachReview` kind
+
+So the next honest move was:
+
+- align review-flow intervention semantics with the already-landed recognition-review continuity path
+
+## What landed
+
+Primary code changes:
+
+- `UnrealHyperTwist/Source/UnrealHyperTwist/Private/HyperTwistTraining/HyperTwistTrainingRepositoryLibrary.cpp`
+  - review-flow derivation now treats recognition-biased repeat pressure as coach-facing intervention pressure alongside explicit `CoachReview`
+
+## Product effect
+
+The coach-to-review loop is now more coherent under recognition-assisted review:
+
+- recognition-biased review plans with repeats now surface `coach-review-needed`
+- review-program summaries now inherit that intervention pressure through the existing derivation path
+- coach brief generation now sees the same intervention state without needing a new heuristic packet
+- dashboard inspection remains truthful because it already renders the derived intervention flag
+
+## Acceptance criteria
+
+- a recognition-biased review plan with repeat pressure sets `bNeedsCoachIntervention`
+- `NextActionLabel` resolves to `coach-review-needed` for that case
+- review-program summary inherits the corrected intervention state without schema changes
+- full product build succeeds
+
+## Validation checklist
+
+1. build `UnrealHyperTwist.sln` / `UnrealHyperTwistEditor`
+2. confirm review-flow derivation treats recognition-biased repeat pressure as coach-facing
+3. confirm the derived intervention state continues through review-program summary and coach brief generation without reopening a broader packet
+4. confirm dashboard inspection remains aligned because it reads the existing derived flow state
+
+That is the packet.