diff --git a/src/services/ai-chat-qa.ts b/src/services/ai-chat-qa.ts
index ad2def9d6b..9311b51a80 100644
--- a/src/services/ai-chat-qa.ts
+++ b/src/services/ai-chat-qa.ts
@@ -57,11 +57,13 @@ const PRIVATE_DECISION_BLOCKER_PATTERN =
   /\b(?:open_pr_pressure|closed_pr_credibility|low_credibility|maintainer_lane|inactive_or_unknown_lane|issue_discovery_only|merged_pr_history_floor|issue_discovery_validity_floor)\b/gi;
 const PRIVATE_BOUNDARY_TERM_PATTERN =
   /\b(?:wallets?|hotkeys?|coldkeys?|seed phrases?|mnemonics?|raw trust scores?|trust scores?|scoreability|reviewability|payouts?|rewards?|reward estimates?|farming|rankings?)\b/gi;
+const PRIVATE_LANE_SIGNAL_PATTERN =
+  /\b(?:maintainer cut|direct PR lane share|direct PR share|direct PR|issue-discovery|split lane)(?:\s*[:(,;-]?\s*\d+(?:\.\d+)?%?)?/gi;
 
 // Public-safe forbidden-term guard on the MODEL's OWN output, mirroring ai-summaries' containsPublicForbiddenText:
 // near-miss phrasings the throwing word-list validator narrows (e.g. bare "estimated score") are also caught.
 const PUBLIC_FORBIDDEN_TEXT_PATTERN =
-  /\b(wallets?|hotkeys?|coldkeys?|seed phrases?|mnemonics?|raw trust scores?|trust scores?|estimated scores?|score estimates?|scoreability|score preview|public score estimates?|estimated rewards?|rewards?|reward estimates?|payouts?|farming|reviewability(?: internals?)?|private reviewability|private scoreability|private rankings?|rankings?|reward optimization)\b/i;
+  /\b(wallets?|hotkeys?|coldkeys?|seed phrases?|mnemonics?|raw trust scores?|trust scores?|estimated scores?|score estimates?|scoreability|score preview|public score estimates?|estimated rewards?|rewards?|reward estimates?|payouts?|farming|reviewability(?: internals?)?|private reviewability|private scoreability|private rankings?|rankings?|reward optimization|maintainer cut|direct PR lane share|direct PR share|issue-discovery|split lane)\b/i;
 
 type ChatGroundingAction = {
   actionType: string;
@@ -192,6 +194,7 @@ function redactGroundingText(value: string): string {
   return value
     .replace(/\blikely_duplicate\b/gi, "possible overlap with existing work")
     .replace(PRIVATE_DECISION_BLOCKER_PATTERN, "private readiness context")
+    .replace(PRIVATE_LANE_SIGNAL_PATTERN, "private readiness context")
     .replace(PRIVATE_BOUNDARY_TERM_PATTERN, "private context")
     .trim();
 }
diff --git a/test/unit/ai-chat-qa.test.ts b/test/unit/ai-chat-qa.test.ts
index 7a56b1cfe1..259236665a 100644
--- a/test/unit/ai-chat-qa.test.ts
+++ b/test/unit/ai-chat-qa.test.ts
@@ -256,6 +256,29 @@ describe("generateChatQaAnswer", () => {
     expect(userMessage).not.toMatch(/\blikely_duplicate\b/);
   });
 
+  it("redacts private lane signals from cached rationale before prompting the provider", async () => {
+    const run = vi.fn(async () => ({ response: "Public-safe readiness answer." }));
+    const env = createTestEnv({ AI_ADVISORY: { run } as unknown as Ai, AI_DAILY_NEURON_BUDGET: "10000" });
+    const result = await generateChatQaAnswer(env, {
+      bundle: bundleFixture(undefined, {
+        why: [
+          "owner/repo: Maintainer cut: 1.",
+          "owner/repo: split lane (direct PR 1, issue-discovery 1); both lanes are useful here.",
+          "owner/repo: direct PR lane share 1 with no hard personal blocker.",
+        ],
+      }),
+      question: "what should I know?",
+      advisoryAiRouting: ADVISORY_ON,
+      repoFullName: "owner/repo",
+      issueNumber: 42,
+    });
+    expect(result).toMatchObject({ status: "ok" });
+    const call = run.mock.calls[0] as unknown as [string, { messages: Array<{ content: string }> }];
+    const userMessage = call[1].messages[1]?.content ?? "";
+    expect(userMessage).not.toMatch(/Maintainer cut|split lane|direct PR lane share|issue-discovery/i);
+    expect(userMessage).toContain("private readiness context");
+  });
+
   it("honors a custom model override and clamps output tokens", async () => {
     const run = vi.fn(async () => ({ response: "Custom-model answer." }));
     const env = createTestEnv({
@@ -284,6 +307,13 @@ describe("generateChatQaAnswer", () => {
     expect(result).toMatchObject({ status: "unsafe" });
   });
 
+  it("withholds private lane signals if the provider repeats them", async () => {
+    const run = vi.fn(async () => ({ response: "This mentions direct PR lane share and issue-discovery details." }));
+    const env = createTestEnv({ AI_ADVISORY: { run } as unknown as Ai, AI_DAILY_NEURON_BUDGET: "10000" });
+    const result = await generateChatQaAnswer(env, { bundle: bundleFixture(), question: "why?", advisoryAiRouting: ADVISORY_ON, repoFullName: "owner/repo", issueNumber: 1 });
+    expect(result).toMatchObject({ status: "unsafe" });
+  });
+
   it("reports an error status with the underlying message when the provider throws an Error", async () => {
     const run = vi.fn(async () => {
       throw new Error("provider_down");
@@ -317,6 +347,9 @@ describe("__chatQaInternals", () => {
     expect(redactGroundingText("blocked by open_pr_pressure")).toBe("blocked by private readiness context");
     expect(redactGroundingText("do not mention a wallet or hotkey")).toBe("do not mention a private context or private context");
     expect(redactGroundingText("likely_duplicate of #123")).toBe("possible overlap with existing work of #123");
+    expect(redactGroundingText("Maintainer cut: 1; direct PR lane share 1; issue-discovery 1")).toBe(
+      "private readiness context; private readiness context; private readiness context",
+    );
     expect(redactGroundingText("perfectly safe text")).toBe("perfectly safe text");
   });
 
@@ -342,6 +375,7 @@ describe("__chatQaInternals", () => {
 
   it("flags forbidden public terms via the shared sanitizer and the local near-miss pattern", () => {
     expect(containsPublicForbiddenText("mentions a wallet")).toBe(true);
+    expect(containsPublicForbiddenText("mentions direct PR lane share details")).toBe(true);
     expect(containsPublicForbiddenText("perfectly safe prose")).toBe(false);
   });