Update Phase-3 Wave 3E status

PLAN.md

@@ -164,10 +164,11 @@ Phase-2 Status (Wave E):
 - Kandidatenpriorisierung
 - automatische Decoder-Slot-Vergabe
 
-Phase-3 Status (Wave 3C):
+Phase-3 Status (Wave 3E):
 - Wave 3A: Priority-Scoring + BudgetModel scaffolding integriert
 - Wave 3B: Arbitration/Admission + Hold/Displacement-Logik für Refinement/Record/Decode
 - Wave 3C: Pressure-Summaries + Decision-Queue-Stats + Reason-Taxonomie + Tests (Hold/Displacement)
+- Wave 3E: Konservative cross-resource Rebalance-Schicht (Refinement/Record/Decode) mit profil-/intent-spezifischem Schutz, Debug-Surface und Szenario-Tests
 
 ### Phase 4
 - breitbandige Multi-Span-Profile

README.md

@@ -111,10 +111,12 @@ Phase-2 status (Wave E):
 - fused candidate evidence summaries expose role/kind detail
 - surveillance level summaries include role intent for easier debug
 
-Phase-3 status (Wave 3C):
+Phase-3 status (Wave 3E):
 - scheduler/arbitration layer for refinement + record/decode admission
 - hold/displacement policy with normalized reason tags
 - budget pressure summaries and decision queue stats (including displaced-by-hold counters)
+- conservative cross-resource rebalance (refinement/record/decode) with profile/intent-aware protections
+- rebalance state and reasons surfaced in arbitration/debug snapshots
 
 The long-term target is that you describe *what the system should do* (for example broad-span monitoring intent, preferred signal families, auto-record/decode priorities), while the engine decides *how* to allocate surveillance, refinement and decoding budgets.