fix(rectification): stop a stale candidate-set prefix from hiding the choice card

Scoring scoped the contrast packet to the previous round's candidate set while
minting probes for the new one, so every later read re-prefixed the stored hash
and the persisted focus schema could never match. Mint one prefix per set and
compare splits by probe identity.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LVapmh2oGNyr6ECHKjPJY8

frontend/src/mastra/rectification-v9-tools.ts

@@ -105,6 +105,7 @@ import {
   inspectDiscriminatorProbes,
   mentionedVargaKeysFromLedgerEvidence,
 } from "@/lib/rectification-agentic/core/candidate-contrast-packet";
+import { candidateSetId } from "@/lib/rectification-agentic/core/build-state";
 import { offerSessionKinds } from "@/lib/rectification-agentic/core/decide-next-action";
 import {
   candidateSnapshotSource,
@@ -882,9 +883,16 @@ export function createRectificationV9Tools(ctx: RectificationV9Context) {
     });
     const refinement = refinementFromDecisionReceipt(receipt);
     const windowScan = windowScanFromDecisionReceipt(receipt);
+    // The packet must be scoped to the candidate set this round produced. Using
+    // the previous round's set id mints split hashes that no later read can
+    // match, which silently suppresses the persisted choice card.
+    const scoredCandidateSetId = candidateSetId(
+      parsed.case.candidateRange.start_time,
+      parsed.case.candidateRange.end_time,
+      score.candidates.map((item) => item.time),
+    );
     const contrastPacket = buildCandidateContrastPacket({
-      candidateSetVersion: previousInferenceFromReceipt(dossier.latestResult?.decisionReceipt ?? null)?.candidate_set_id
-        ?? score.engineResultId,
+      candidateSetVersion: scoredCandidateSetId,
       calculationResultId: score.engineResultId,
       engineProbes: refinement.discriminating_event_probes,
       vargaDifferences: [