diff --git a/docs/BUG_HISTORY.md b/docs/BUG_HISTORY.md
index f422491d..198a4750 100644
--- a/docs/BUG_HISTORY.md
+++ b/docs/BUG_HISTORY.md
@@ -6214,8 +6214,8 @@
 - 用户现象：训练事件已经够、推理层已有高信息量 D24 探针，界面仍出低信息量事业存在题；题干修复后选题内容仍不对。
 - 触发条件：快照 `latest_result.candidates` 为空，或可信区间与推理活跃时刻不一致导致 `authoritativeCandidateProjection` fail-closed；Python `discriminating_event_probes` 仍给出低分事业题；已打开的事业区分卡把后续排序锁在原题上。
 - 根因：BUG-405 把排序公式改成按信息量取最高，但读路径组包不读 `inference_state.probes`，remaining D24 又依赖快照时刻。时刻被投影饿死后目录只剩职业题。评分落库用 `score.candidates` 能把 D24 写进推理层，GET/工具随后丢掉。已打开的低分 distinguish Focus 还在 method-followup 里优先于重新排序。
-- 修复：GET、Agent、工具共用一份 `rectificationFollowupCatalog`。合并未回答的推理探针与 Python 事件探针，按 `semantic_key` 去重留更高信息量。拼 remaining splits 用推理未淘汰时刻，不把 adopt 投影的空 `scores` 当成出题目录。`varga.d24`/`d5` 按质量题、`d9`/`d10` 按风格题补全。method-followup 对 packet 里全部 contrast 探针排序；已打开但 semantic_key 不是当前目录赢家的区分卡让位。评分落库仍用当次 `score.candidates`。Skill 版本保持 `10.0.13`。
-- 验证：`frontend/tests/rectification-decision-authority.test.ts` 空快照仍选出 D24；`frontend/tests/rectification-eight-method.test.ts` 低分事业题与已打开事业卡都不得压过可渲染 D24。
+- 修复：GET、Agent、工具共用一份 `rectificationFollowupCatalog`。合并未回答的推理探针与 Python 事件探针，按 `semantic_key` 去重留更高信息量。拼 remaining splits 用推理未淘汰时刻，不把 adopt 投影的空 `scores` 当成出题目录。已有日期证据的领域不再出存在题（例如已记事业就不再问另一年入职）；D9/D10/D24 等分盘区分题仍进池，按信息量全局取最高，不绑定某个领域。`varga.d24`/`d5` 按质量题、`d9`/`d10` 按风格题补全。已打开但 semantic_key 不是当前目录赢家的区分卡让位。评分落库仍用当次 `score.candidates`。Skill 版本保持 `10.0.13`。
+- 验证：`frontend/tests/rectification-decision-authority.test.ts` 空快照仍选出目录最高分；`frontend/tests/rectification-eight-method.test.ts` 已覆盖领域的存在题让位，D10 与 D24 谁分高问谁。
 - 防复发：出题目录必须来自推理探针加引擎探针，不得只吃 Python 事件探针或快照投影时刻。Adopt/展示投影 fail-closed 不得饿死出题。已打开的低分区分卡不得挡住更高分目录赢家。
 - 相关记录：BUG-405、BUG-406
 - 复发自：BUG-405（排序公式对，目录被投影饿死，已打开低分卡锁题）
diff --git a/frontend/src/lib/rectification-agentic/core/candidate-contrast-packet.ts b/frontend/src/lib/rectification-agentic/core/candidate-contrast-packet.ts
index 9fd0ad0f..e33036cb 100644
--- a/frontend/src/lib/rectification-agentic/core/candidate-contrast-packet.ts
+++ b/frontend/src/lib/rectification-agentic/core/candidate-contrast-packet.ts
@@ -251,6 +251,24 @@ export function askedEventProbeKeysFromLedgerEvidence(
   return [...keys];
 }
 
+export function datedDomainsFromEvidence(
+  evidence: readonly Readonly<{
+    status?: string | null;
+    domain?: string | null;
+    occurredFrom?: string | null;
+    occurredTo?: string | null;
+  }>[],
+): string[] {
+  const domains = new Set<string>();
+  for (const item of evidence) {
+    if (item.status && !LIVE_EVIDENCE.has(item.status)) continue;
+    if (!item.domain) continue;
+    const dated = [item.occurredFrom, item.occurredTo].some((value) => /^\d{4}/.test(value ?? ""));
+    if (dated) domains.add(item.domain);
+  }
+  return [...domains];
+}
+
 export function volunteeredDomainsFromEvidence(
   evidence: readonly Readonly<{
     status?: string | null;
@@ -286,18 +304,21 @@ export function buildCandidateContrastPacket(input: {
   transitions?: readonly WindowScanTransition[];
   askedKeys?: readonly string[];
   volunteeredDomains?: readonly string[];
+  providedDomains?: readonly string[];
 }): CandidateContrastPacket {
   const asked = new Set(input.askedKeys ?? []);
+  const provided = new Set(input.providedDomains ?? []);
   const fromEngine = (input.engineProbes ?? []).flatMap((probe) => {
     const built = probeFromEngine(probe, input.candidateSetVersion, input.calculationResultId ?? null);
     if (!built) return [];
     const eventKey = built.domain && built.year ? `${built.domain}.${built.year}` : null;
-    const isStructured = built.choiceKind === "varga_style" || built.semanticKey.startsWith("varga.");
+    const isStructured = isStructuredDiscriminator(built);
     if (
       asked.has(built.semanticKey)
       || asked.has(built.candidateSplitHash)
       || asked.has(built.probeId)
       || (!isStructured && eventKey && asked.has(eventKey))
+      || (!isStructured && built.domain && provided.has(built.domain))
     ) {
       return [];
     }
@@ -316,13 +337,13 @@ export function buildCandidateContrastPacket(input: {
     transitions: input.transitions ?? [],
   });
   const presentKeys = new Set(fromEngine.map((item) => item.semanticKey));
-  const fromVarga = vargaProbe(
+  const fromVarga = vargaProbes(
     remainingSplits,
     input.candidateSetVersion,
     input.calculationResultId ?? null,
     new Set([...asked, ...presentKeys]),
   );
-  const probes = [...fromEngine, ...(fromVarga ? [fromVarga] : [])]
+  const probes = [...fromEngine, ...fromVarga]
     .sort((left, right) => right.informationGain - left.informationGain);
   return {
     candidateSetVersion: input.candidateSetVersion,
@@ -596,15 +617,23 @@ function styleOptionsFromEngine(
   return parsed.length > 0 ? parsed : undefined;
 }
 
-function vargaProbe(
+export function isStructuredDiscriminator(probe: Pick<CandidateDiscriminatorProbe, "choiceKind" | "semanticKey">): boolean {
+  return probe.choiceKind === "varga_style"
+    || probe.choiceKind === "event_quality"
+    || probe.semanticKey.startsWith("varga.");
+}
+
+function vargaProbes(
   remainingSplits: readonly RemainingVargaSplit[],
   candidateSetVersion: string,
   calculationResultId: string | null,
   asked: ReadonlySet<string>,
-): CandidateDiscriminatorProbe | null {
-  const remaining = remainingSplits.find((item) => !vargaLayerAsked(asked, item.layer));
-  if (!remaining) return null;
-  return vargaProbeFromRemaining(remaining, candidateSetVersion, calculationResultId);
+): CandidateDiscriminatorProbe[] {
+  return remainingSplits.flatMap((item) => {
+    if (vargaLayerAsked(asked, item.layer)) return [];
+    const probe = vargaProbeFromRemaining(item, candidateSetVersion, calculationResultId);
+    return probe ? [probe] : [];
+  });
 }
 
 function vargaLayerAsked(asked: ReadonlySet<string>, layer: string): boolean {
diff --git a/frontend/src/lib/rectification-agentic/v9/decision-from-dossier.ts b/frontend/src/lib/rectification-agentic/v9/decision-from-dossier.ts
index f430afd5..c7332c04 100644
--- a/frontend/src/lib/rectification-agentic/v9/decision-from-dossier.ts
+++ b/frontend/src/lib/rectification-agentic/v9/decision-from-dossier.ts
@@ -7,6 +7,7 @@
 
 import {
   buildCandidateContrastPacket,
+  datedDomainsFromEvidence,
   selectDiscriminatorProbe,
   volunteeredDomainsFromEvidence,
   type CandidateContrastPacket,
@@ -175,6 +176,7 @@ export function contrastPacketFromLatestResult(
     transitions: windowScan?.transitions ?? [],
     askedKeys: askedDiscriminatorKeys(latest?.decisionReceipt, evidence),
     volunteeredDomains: volunteeredDomainsFromEvidence(evidence),
+    providedDomains: datedDomainsFromEvidence(evidence),
   });
 }
 
diff --git a/frontend/src/lib/rectification-agentic/v9/method-followup.ts b/frontend/src/lib/rectification-agentic/v9/method-followup.ts
index 71addedc..c76128d8 100644
--- a/frontend/src/lib/rectification-agentic/v9/method-followup.ts
+++ b/frontend/src/lib/rectification-agentic/v9/method-followup.ts
@@ -26,10 +26,11 @@
  * Method coverage asks for dated events in natural language.
  * Known-event quality probes (exam went badly for a year already
  * in the ledger) stamp a choice card as soon as that year is
- * recorded. Dasha conflict probes wait until training event
- * quality (3 training events in 2 training domains; holdout
- * excluded), then jump ahead of remaining method rotation and
- * block offering time cards so the window can be filtered.
+ * recorded. Dasha existence probes skip a domain once that domain
+ * already has dated evidence. Remaining chart discriminators
+ * (D9/D10/D24 and other varga splits) stay in the pool and the
+ * highest information-gain renderable probe is asked next, with
+ * no preferred domain.
  * If holdout is already reserved but training is still short,
  * keep collecting a dated event instead of discriminating.
  * Once blocking methods are covered, move into candidate discrimination.
@@ -52,6 +53,8 @@ import { overlayChoicePromptFromSpoken } from "./turn-narration.ts";
 import { decideRectification, type HoldoutValidationStatus } from "../core/rectification-decision.ts";
 import {
   askedKeysFromLedgerEvidence,
+  datedDomainsFromEvidence,
+  isStructuredDiscriminator,
   selectDiscriminatorProbe,
   type CandidateContrastPacket,
   type CandidateDiscriminatorProbe,
@@ -203,6 +206,19 @@ function existenceNearbyYears(domain: string): number {
   return domain === "education" ? EDUCATION_EXISTENCE_NEARBY_YEARS : 0;
 }
 
+function probeDomainAlreadyCovered(
+  evidence: readonly MethodFollowupEvidence[],
+  domain: string,
+): boolean {
+  return evidence.some((item) => {
+    if (item.status !== "confirmed" && item.status !== "draft" && item.status !== "pending_confirmation") {
+      return false;
+    }
+    if (item.domain !== domain) return false;
+    return evidenceYear(item) !== null;
+  });
+}
+
 function probeYearAlreadyCovered(
   evidence: readonly MethodFollowupEvidence[],
   domain: string,
@@ -362,7 +378,7 @@ function remainingConflictProbes(
     if (probe.source === "known_event_quality" || probe.role === "clarify") continue;
     if (!isValidDistinguishProbe({ ...probe, role: "distinguish" })) continue;
     if (declined.has(probe.domain)) continue;
-    if (probeYearAlreadyCovered(evidence, probe.domain, probe.year)) continue;
+    if (probeDomainAlreadyCovered(evidence, probe.domain)) continue;
     const semantic = probe.semantic_key ?? `${probe.domain}.${probe.year}`;
     const split = probe.candidate_split_hash ?? "";
     if (askedKeys.has(semantic) || (split && askedKeys.has(split))) continue;
@@ -516,8 +532,10 @@ function rankRenderableDiscriminators(input: {
   contrastProbes: readonly CandidateDiscriminatorProbe[];
   askedKeys: ReadonlySet<string>;
   topCandidateTimes?: readonly string[];
+  providedDomains?: readonly string[];
 }): RankedDiscriminator[] {
   const top = input.topCandidateTimes ?? [];
+  const provided = new Set(input.providedDomains ?? []);
   const rows: RankedDiscriminator[] = [];
   const seen = new Set<string>();
   const push = (row: RankedDiscriminator | null) => {
@@ -533,6 +551,7 @@ function rankRenderableDiscriminators(input: {
     push(renderableEventProbe(probe, input.askedKeys, top));
   }
   for (const probe of input.contrastProbes) {
+    if (!isStructuredDiscriminator(probe) && probe.domain && provided.has(probe.domain)) continue;
     push(renderableContrastProbe(probe, input.askedKeys, top));
   }
   return rows.sort((left, right) => right.score - left.score || (right.eventProbe?.information_gain ?? right.contrastProbe?.informationGain ?? 0) - (left.eventProbe?.information_gain ?? left.contrastProbe?.informationGain ?? 0));
@@ -912,6 +931,7 @@ export function buildMethodFollowupPlan(input: {
       contrastProbes,
       askedKeys,
       topCandidateTimes: input.topCandidateTimes,
+      providedDomains: datedDomainsFromEvidence(input.evidence),
     })
     : [];
   const bestDiscriminator = rankedDiscriminators[0] ?? null;
diff --git a/frontend/src/mastra/rectification-v9-tools.ts b/frontend/src/mastra/rectification-v9-tools.ts
index ed4cded1..c2703280 100644
--- a/frontend/src/mastra/rectification-v9-tools.ts
+++ b/frontend/src/mastra/rectification-v9-tools.ts
@@ -99,6 +99,7 @@ import {
 import {
   buildCandidateContrastPacket,
   conflictProbesFromContrast,
+  datedDomainsFromEvidence,
   selectDiscriminatorProbe,
   volunteeredDomainsFromEvidence,
 } from "@/lib/rectification-agentic/core/candidate-contrast-packet";
@@ -1002,6 +1003,7 @@ export function createRectificationV9Tools(ctx: RectificationV9Context) {
         parsed.evidence,
       ),
       volunteeredDomains: volunteeredDomainsFromEvidence(parsed.evidence),
+      providedDomains: datedDomainsFromEvidence(parsed.evidence),
     });
     const inference = buildCaseInferenceState({
       range: parsed.case.candidateRange,
diff --git a/frontend/tests/rectification-candidate-contrast-packet.test.ts b/frontend/tests/rectification-candidate-contrast-packet.test.ts
index d9291e12..3e4b61a3 100644
--- a/frontend/tests/rectification-candidate-contrast-packet.test.ts
+++ b/frontend/tests/rectification-candidate-contrast-packet.test.ts
@@ -24,6 +24,8 @@ test("remaining D10 three-way outranks a skewed D24 split", () => {
   assert.equal(probe.choiceKind, "varga_style");
   assert.match(probe.semanticKey, /varga\.d10/);
   assert.doesNotMatch(probe.semanticKey, /varga\.d24/);
+  assert.ok(packet.probes.some((item) => item.semanticKey.startsWith("varga.d24.")));
+  assert.ok(packet.probes.some((item) => item.semanticKey.startsWith("varga.d10.")));
   assert.equal(probe.styleOptions?.length, 4);
   assert.equal(probe.expectedOutcomes.length, 4);
   assert.deepEqual(probe.expectedOutcomes.filter((row) => row.outcomeId !== "unsure").map((row) => row.supportsCandidateIds), [
@@ -83,3 +85,28 @@ test("finance remaining splits stay out unless volunteered", () => {
   );
   assert.equal(shown[0]?.layer, "d2");
 });
+
+test("existence probes leave the catalog once that domain already has dated evidence", () => {
+  const packet = buildCandidateContrastPacket({
+    candidateSetVersion: "05:00-05:14",
+    engineProbes: [{
+      semantic_key: "career.2023.dasha_activation",
+      domain: "career",
+      year: 2023,
+      user_meaning: "时间范围锁定 2023 年前后。",
+      information_gain: 0.56,
+      expected_outcomes: [
+        { answer_class: "yes", supports: ["05:00"], conflicts: ["05:14"] },
+        { answer_class: "no", supports: ["05:14"], conflicts: ["05:00"] },
+      ],
+    }],
+    providedDomains: ["career"],
+    candidateTimes: ["05:00", "05:07", "05:14"],
+    transitions: [
+      { layer: "d24", at: "05:07", from_sign: "白羊座", to_sign: "金牛座" },
+      { layer: "d24", at: "05:14", from_sign: "金牛座", to_sign: "双子座" },
+    ],
+  });
+  assert.equal(packet.probes.some((item) => item.semanticKey.includes("career.2023")), false);
+  assert.match(selectDiscriminatorProbe(packet)?.semanticKey ?? "", /^varga\.d24\./);
+});
diff --git a/frontend/tests/rectification-decision-authority.test.ts b/frontend/tests/rectification-decision-authority.test.ts
index 26d7152d..0134e839 100644
--- a/frontend/tests/rectification-decision-authority.test.ts
+++ b/frontend/tests/rectification-decision-authority.test.ts
@@ -265,6 +265,7 @@ test("scored inference catalog outranks a low-gain Python career probe when snap
   };
   const packet = contrastPacketFromDossier(dossier);
   const selected = selectDiscriminatorProbe(packet);
+  assert.equal(packet.probes.some((probe) => probe.semanticKey.includes("career.2023")), false);
   assert.match(selected?.semanticKey ?? "", /^varga\.d24\./);
   assert.ok((selected?.informationGain ?? 0) > 2);
   assert.doesNotMatch(selected?.semanticKey ?? "", /career\.2023/);
diff --git a/frontend/tests/rectification-eight-method.test.ts b/frontend/tests/rectification-eight-method.test.ts
index 1e9ccf25..daa8c83f 100644
--- a/frontend/tests/rectification-eight-method.test.ts
+++ b/frontend/tests/rectification-eight-method.test.ts
@@ -1057,19 +1057,19 @@ test("evidence batch returns the persisted choice prompt as open_question", asyn
       discriminating_event_probes: [{
         year: 2023,
         year_label: "2023 年前后",
-        domain: "career",
-        event_family: "入职、升职或职责明显加重",
+        domain: "relocation",
+        event_family: "搬家、离乡或长期异地",
         source: "dasha_activation",
         tracks: ["vimshottari", "narayana"],
         tracks_agree: false,
         unique_minute_claim: false,
-        user_meaning: "年份锁定 2023 年前后。事件家族：入职、升职或职责明显加重。",
+        user_meaning: "年份锁定 2023 年前后。事件家族：搬家、离乡或长期异地。",
         role: "distinguish",
         phase: "candidate_discriminator",
         information_gain: 1.09,
-        semantic_key: "career.2023.dasha_activation",
+        semantic_key: "relocation.2023.dasha_activation",
         candidate_set_version: "set-test",
-        candidate_split_hash: "set-test:career:2023",
+        candidate_split_hash: "set-test:relocation:2023",
         candidate_ids: ["04:50", "05:20"],
         expected_outcomes: [
           { answer_class: "yes", supports: ["04:50"], conflicts: ["05:20"] },
@@ -1158,7 +1158,7 @@ test("evidence batch returns the persisted choice prompt as open_question", asyn
     const schema = setFocus?.args.p_expected_answer_schema as { choice?: { prompt?: string } } | undefined;
     assert.match(schema?.choice?.prompt ?? "", /2023 年前后/);
     assert.match(result.open_question?.prompt ?? "", /2023 年前后/);
-    assert.match(result.open_question?.prompt ?? "", /入职、升职或职责明显加重/);
+    assert.match(result.open_question?.prompt ?? "", /搬家、离乡或长期异地/);
     assert.doesNotMatch(result.open_question?.prompt ?? "", /高考/);
   } finally {
     restore();
@@ -1548,6 +1548,30 @@ test("confirmed relationship evidence skips generic D9 followups unless a real p
   const probed = buildMethodFollowupPlan({
     evidence,
     precisionStage: "d9_refine",
+    contrastPacket: {
+      candidateSetVersion: "05:00-05:14",
+      vargaDifferences: [],
+      probes: [{
+        probeId: "contrast:varga.d9.05:00|05:14",
+        candidateSetVersion: "05:00-05:14",
+        question: "当前几个候选在关系盘上还分得开。",
+        expectedOutcomes: [
+          { outcomeId: "yes", supportsCandidateIds: ["05:00"], conflictsCandidateIds: ["05:14"] },
+          { outcomeId: "weak_yes", supportsCandidateIds: ["05:14"], conflictsCandidateIds: ["05:00"] },
+        ],
+        candidateSplitHash: "varga.d9.05:00|05:14",
+        informationGain: 1.4,
+        sourceFeatures: [{ technique: "D9", calculationResultId: RESULT_ID }],
+        domain: "relationship",
+        year: null,
+        semanticKey: "varga.d9.05:00|05:14",
+        choiceKind: "varga_style",
+        styleOptions: [
+          { label: "相处更主动热情", answerClass: "yes", sign: "白羊座" },
+          { label: "相处更深刻占有", answerClass: "weak_yes", sign: "天蝎座" },
+        ],
+      }],
+    },
     eventProbes: [{
       ...CAREER_CONFLICT_PROBE,
       year: 2021,
@@ -1560,7 +1584,8 @@ test("confirmed relationship evidence skips generic D9 followups unless a real p
   });
   assert.equal(probed.next_followup?.domain, "relationship");
   assert.equal(probed.next_followup?.source, "event_probe");
-  assert.equal(probed.next_followup?.semantic_key, "relationship.2021.dasha_activation");
+  assert.equal(probed.next_followup?.semantic_key, "varga.d9.05:00|05:14");
+  assert.doesNotMatch(probed.next_followup?.semantic_key ?? "", /relationship\.2021/);
 });
 
 test("d9_refine after relationship still asks uncovered career first", () => {
@@ -2065,22 +2090,88 @@ test("structured paused state ends evidence collection without parsing user copy
 });
 
 
-test("same domain different year still asks a conflict probe", () => {
-  const plan = buildMethodFollowupPlan({
-    evidence: [
-      datedEvidence("education", "2016"),
-      datedEvidence("relationship", "2018"),
-      datedEvidence("career", "2015"),
-      datedEvidence("family", "2023"),
-    ],
+test("covered-domain existence probes are skipped; the highest remaining discriminator wins", () => {
+  const evidence = [
+    datedEvidence("education", "2016"),
+    datedEvidence("relationship", "2018"),
+    datedEvidence("career", "2015"),
+    datedEvidence("family", "2023"),
+  ];
+  const skipped = buildMethodFollowupPlan({
+    evidence,
     eventProbes: [{
       ...CAREER_CONFLICT_PROBE,
-      information_gain: 0.21,
+      information_gain: 0.56,
       semantic_key: "career.2018.dasha_activation",
     }],
   });
-  assert.equal(plan.next_followup?.source, "event_probe");
-  assert.equal(plan.next_followup?.domain, "career");
+  assert.notEqual(skipped.next_followup?.domain, "career");
+  assert.notEqual(skipped.next_followup?.source, "event_probe");
+
+  const d24 = {
+    probeId: "contrast:varga.d24.05:00|05:07|05:14",
+    candidateSetVersion: "05:00-05:14",
+    question: "当前几个候选在学业盘上还分得开。",
+    expectedOutcomes: [
+      { outcomeId: "yes", supportsCandidateIds: ["05:00"], conflictsCandidateIds: ["05:07", "05:14"] },
+      { outcomeId: "no", supportsCandidateIds: ["05:07", "05:14"], conflictsCandidateIds: ["05:00"] },
+    ],
+    candidateSplitHash: "varga.d24.05:00|05:07|05:14",
+    informationGain: 2.5,
+    sourceFeatures: [{ technique: "D24", calculationResultId: RESULT_ID }],
+    domain: "education",
+    year: null,
+    semanticKey: "varga.d24.05:00|05:07|05:14",
+    choiceKind: "event_quality" as const,
+  };
+  const d10 = {
+    probeId: "contrast:varga.d10.05:00|05:07|05:14",
+    candidateSetVersion: "05:00-05:14",
+    question: "当前几个候选在事业盘上还分得开。",
+    expectedOutcomes: [
+      { outcomeId: "yes", supportsCandidateIds: ["05:00"], conflictsCandidateIds: ["05:07", "05:14"] },
+      { outcomeId: "weak_yes", supportsCandidateIds: ["05:07"], conflictsCandidateIds: ["05:00", "05:14"] },
+      { outcomeId: "no", supportsCandidateIds: ["05:14"], conflictsCandidateIds: ["05:00", "05:07"] },
+    ],
+    candidateSplitHash: "varga.d10.05:00|05:07|05:14",
+    informationGain: 3.1,
+    sourceFeatures: [{ technique: "D10", calculationResultId: RESULT_ID }],
+    domain: "career",
+    year: null,
+    semanticKey: "varga.d10.05:00|05:07|05:14",
+    choiceKind: "varga_style" as const,
+    styleOptions: [
+      { label: "做事偏领导推进", answerClass: "yes" as const, sign: "白羊座" },
+      { label: "做事偏研究转化", answerClass: "weak_yes" as const, sign: "天蝎座" },
+    ],
+  };
+  const highest = buildMethodFollowupPlan({
+    evidence,
+    eventProbes: [{
+      ...CAREER_CONFLICT_PROBE,
+      information_gain: 0.56,
+      semantic_key: "career.2018.dasha_activation",
+    }],
+    contrastPacket: {
+      candidateSetVersion: "05:00-05:14",
+      vargaDifferences: [],
+      probes: [d24, d10],
+    },
+    candidatesSeparated: false,
+  });
+  assert.equal(highest.next_followup?.semantic_key, d10.semanticKey);
+  assert.doesNotMatch(highest.next_followup?.semantic_key ?? "", /career\.2018/);
+
+  const d24Wins = buildMethodFollowupPlan({
+    evidence,
+    contrastPacket: {
+      candidateSetVersion: "05:00-05:14",
+      vargaDifferences: [],
+      probes: [d24, { ...d10, informationGain: 1.1 }],
+    },
+    candidatesSeparated: false,
+  });
+  assert.equal(d24Wins.next_followup?.semantic_key, d24.semanticKey);
 });
 
 test("three dated events with one holdout keep collecting instead of discriminating", () => {