diff --git a/apps/desktop/src/main/services/ai/aiIntegrationService.ts b/apps/desktop/src/main/services/ai/aiIntegrationService.ts
index e5590830d..2351331e0 100644
--- a/apps/desktop/src/main/services/ai/aiIntegrationService.ts
+++ b/apps/desktop/src/main/services/ai/aiIntegrationService.ts
@@ -1505,7 +1505,7 @@ export function createAiIntegrationService(args: {
       throw new Error("No AI provider is available. Install and authenticate Claude Code and/or Codex CLI.");
     }
 
-    if (!getFeatureFlag(args.feature)) {
+    if (args.taskType !== "session_title" && !getFeatureFlag(args.feature)) {
       logger.warn("ai.task.skipped_feature_disabled", {
         requestId,
         taskType: args.taskType,
@@ -2007,6 +2007,7 @@ export function createAiIntegrationService(args: {
       prompt: string;
       timeoutMs?: number;
       model?: string;
+      reasoningEffort?: string | null;
       jsonSchema?: unknown;
       systemPrompt?: string;
       taskType?: Extract<AiTaskType, "terminal_summary" | "session_title" | "session_summary" | "handoff_summary" | "continuity_summary" | "context_compaction">;
@@ -2018,6 +2019,7 @@ export function createAiIntegrationService(args: {
         prompt: args.prompt,
         timeoutMs: args.timeoutMs,
         model: args.model,
+        reasoningEffort: args.reasoningEffort,
         jsonSchema: args.jsonSchema,
         systemPrompt: args.systemPrompt
       });
diff --git a/apps/desktop/src/main/services/chat/agentChatService.test.ts b/apps/desktop/src/main/services/chat/agentChatService.test.ts
index 3d2a12108..4dd2d0153 100644
--- a/apps/desktop/src/main/services/chat/agentChatService.test.ts
+++ b/apps/desktop/src/main/services/chat/agentChatService.test.ts
@@ -18822,11 +18822,11 @@ describe("createAgentChatService", () => {
 
 describe("suggestLaneNameFromPrompt", () => {
   function createProjectConfigServiceWithTitleOptions(
-    options: { titleGenerationEnabled?: boolean; titleModelId?: string } = {},
+    options: { titleGenerationEnabled?: boolean; titleModelId?: string | null; legacyTitleModelId?: string } = {},
   ) {
     const titleOptions: Record<string, unknown> = {};
     if (typeof options.titleGenerationEnabled === "boolean") titleOptions.enabled = options.titleGenerationEnabled;
-    if (options.titleModelId) titleOptions.modelId = options.titleModelId;
+    if (options.titleModelId !== undefined) titleOptions.modelId = options.titleModelId;
     const sessionIntelligence = Object.keys(titleOptions).length ? { titles: titleOptions } : {};
     return {
       get: vi.fn(() => ({
@@ -18836,7 +18836,9 @@ describe("suggestLaneNameFromPrompt", () => {
               cli: { mode: "edit" },
               inProcess: { mode: "edit" },
             },
-            chat: {},
+            chat: {
+              ...(options.legacyTitleModelId ? { autoTitleModelId: options.legacyTitleModelId } : {}),
+            },
             sessionIntelligence,
           },
         },
@@ -18846,7 +18848,7 @@ describe("suggestLaneNameFromPrompt", () => {
     } as any;
   }
 
-  function createSuggestService(options: { titleGenerationEnabled?: boolean; titleModelId?: string } = {}) {
+  function createSuggestService(options: { titleGenerationEnabled?: boolean; titleModelId?: string | null; legacyTitleModelId?: string } = {}) {
     return createService({
       projectConfigService: createProjectConfigServiceWithTitleOptions(options),
     });
@@ -18879,17 +18881,17 @@ describe("suggestLaneNameFromPrompt", () => {
       modelId: "anthropic/claude-haiku-4-5",
       laneId: "lane-1",
     });
-    expect(result).toBe("fix-the-login-bug");
+    expect(result).toBe("fix-login-bug");
   });
 
-  it("takes only first 4 words of a long prompt", async () => {
+  it("takes the first 5 meaningful words of a long prompt", async () => {
     const { service } = createSuggestService();
     const result = await service.suggestLaneNameFromPrompt({
       prompt: "Refactor the authentication service to use JWT tokens",
       modelId: "anthropic/claude-haiku-4-5",
       laneId: "lane-1",
     });
-    expect(result).toBe("refactor-the-authentication-service");
+    expect(result).toBe("refactor-authentication-service-jwt-tokens");
   });
 
   it("strips special characters from the prompt slug", async () => {
@@ -18899,7 +18901,7 @@ describe("suggestLaneNameFromPrompt", () => {
       modelId: "anthropic/claude-haiku-4-5",
       laneId: "lane-1",
     });
-    expect(result).toBe("fix-bug-123-in");
+    expect(result).toBe("fix-bug-123-module");
   });
 
   it("truncates the fallback slug to 48 characters", async () => {
@@ -18919,7 +18921,7 @@ describe("suggestLaneNameFromPrompt", () => {
       modelId: "anthropic/claude-haiku-4-5",
       laneId: "lane-1",
     });
-    expect(result).toBe("fix-the-bug-now");
+    expect(result).toBe("fix-bug-now");
   });
 
   it("falls back when the model runtime throws an error", async () => {
@@ -18935,14 +18937,14 @@ describe("suggestLaneNameFromPrompt", () => {
       laneId: "lane-1",
     });
 
-    expect(result).toBe("write-a-test-suite");
+    expect(result).toBe("write-test-suite");
     expect(logger.warn).toHaveBeenCalledWith(
       "agent_chat.suggest_lane_name_failed",
       expect.objectContaining({ error: "API rate limited" }),
     );
   });
 
-  it("keeps the prompt fallback readable while adding the temporary suffix when title generation is disabled", async () => {
+  it("uses the deterministic prompt fallback when title generation is disabled", async () => {
     vi.mocked(detectAllAuth).mockResolvedValue([
       { type: "cli-subscription" as any, cli: "claude", authenticated: true, path: "/usr/bin/claude", verified: true },
     ]);
@@ -18955,11 +18957,11 @@ describe("suggestLaneNameFromPrompt", () => {
       fallbackName: "chat-20260514-010203",
     });
 
-    expect(result).toBe("fix-the-authentication-login-20260514-010203");
+    expect(result).toBe("fix-authentication-login-failure-dashboard");
     expect(aiIntegrationService.summarizeTerminal).not.toHaveBeenCalled();
   });
 
-  it("uses the explicit fallback directly when the prompt fallback is generic", async () => {
+  it("preserves the generated suffix when the prompt fallback is generic", async () => {
     const { service } = createSuggestService();
     const result = await service.suggestLaneNameFromPrompt({
       prompt: "!!!",
@@ -18968,7 +18970,7 @@ describe("suggestLaneNameFromPrompt", () => {
       fallbackName: "chat-20260514-010203",
     });
 
-    expect(result).toBe("chat-20260514-010203");
+    expect(result).toBe("parallel-task-20260514-010203");
   });
 
   it("uses AI-generated name when the model runtime succeeds", async () => {
@@ -19022,6 +19024,35 @@ describe("suggestLaneNameFromPrompt", () => {
     }));
   });
 
+  it("does not fall back to a legacy title model when session intelligence model is explicitly cleared", async () => {
+    vi.mocked(detectAllAuth).mockResolvedValue([
+      { type: "cli-subscription" as any, cli: "claude", authenticated: true, path: "/usr/bin/claude", verified: true },
+    ]);
+    const { service, aiIntegrationService } = createSuggestService({
+      titleModelId: null,
+      legacyTitleModelId: "openai/gpt-5.4-mini",
+    });
+    vi.mocked(aiIntegrationService.summarizeTerminal).mockResolvedValueOnce({
+      text: "Fallback Title",
+      inputTokens: 10,
+      outputTokens: 5,
+    } as any);
+
+    const result = await service.suggestLaneNameFromPrompt({
+      prompt: "Fix null model clearing for background jobs",
+      modelId: "",
+      laneId: "lane-1",
+    });
+
+    expect(result).toBe("fallback-title");
+    expect(aiIntegrationService.summarizeTerminal).not.toHaveBeenCalledWith(expect.objectContaining({
+      model: "openai/gpt-5.4-mini",
+    }));
+    expect(aiIntegrationService.summarizeTerminal).toHaveBeenNthCalledWith(1, expect.objectContaining({
+      model: "anthropic/claude-haiku-4-5",
+    }));
+  });
+
   it("normalizes AI-generated name: strips special chars and lowercases", async () => {
     vi.mocked(detectAllAuth).mockResolvedValue([
       { type: "cli-subscription" as any, cli: "claude", authenticated: true, path: "/usr/bin/claude", verified: true },
diff --git a/apps/desktop/src/main/services/chat/agentChatService.ts b/apps/desktop/src/main/services/chat/agentChatService.ts
index 5a19dc9df..1f0213e4d 100644
--- a/apps/desktop/src/main/services/chat/agentChatService.ts
+++ b/apps/desktop/src/main/services/chat/agentChatService.ts
@@ -261,6 +261,12 @@ import { buildAdeCliAgentGuidance } from "../../../shared/adeCliGuidance";
 import { getAdeAgentSkillRootsForPrompt } from "../../../shared/agentSkillRoots";
 import { parseAgentChatTranscript } from "../../../shared/chatTranscript";
 import { extractLeadingSlashCommand, isProviderSlashCommandInput } from "../../../shared/chatSlashCommands";
+import {
+  deriveDeterministicLaneNameFromPrompt,
+  GENERIC_LANE_FALLBACK_NAME,
+  genericLaneFallbackName,
+  genericSuffixFromLaneFallbackName,
+} from "../../../shared/laneNameFallback";
 import { resolveSubagentCapability } from "../../../shared/subagentCapabilities";
 import { stripAnsi } from "../../utils/ansiStrip";
 import type { createCtoStateService } from "../cto/ctoStateService";
@@ -1833,9 +1839,11 @@ type ResolvedChatConfig = {
   sessionBudgetUsd: number | null;
   titleGenerationEnabled: boolean;
   titleModelId: string | null;
+  titleReasoningEffort: string | null;
   titleRefreshOnComplete: boolean;
   summaryEnabled: boolean;
   summaryModelId: string | null;
+  summaryReasoningEffort: string | null;
 };
 
 const MAX_PENDING_STEERS = 10;
@@ -3068,40 +3076,21 @@ function sanitizeAutoTitle(raw: string, maxChars = AUTO_TITLE_MAX_CHARS): string
   return normalized.length > maxChars ? normalized.slice(0, maxChars).trimEnd() : normalized;
 }
 
-const GENERIC_PROMPT_LANE_NAME = "parallel-task";
-
 function fallbackLaneNameFromPrompt(prompt: string): string {
-  const collapsed = prompt.replace(/\s+/g, " ");
-  if (!collapsed.length) return GENERIC_PROMPT_LANE_NAME;
-  const words = collapsed.split(/\s+/).filter(Boolean).slice(0, 4);
-  const slug = words
-    .join("-")
-    .toLowerCase()
-    .replace(/[^a-z0-9-]+/g, "-")
-    .replace(/-+/g, "-")
-    .replace(/^-|-$/g, "");
-  return slug.length ? slug.slice(0, 48) : GENERIC_PROMPT_LANE_NAME;
+  return deriveDeterministicLaneNameFromPrompt(prompt);
 }
 
 function uniquePromptFallbackLaneName(promptFallback: string, explicitFallback: string | null): string {
-  if (promptFallback === GENERIC_PROMPT_LANE_NAME) {
-    return explicitFallback ?? promptFallback;
+  if (promptFallback !== GENERIC_LANE_FALLBACK_NAME) {
+    return promptFallback;
   }
-  if (!explicitFallback) return promptFallback;
-
-  const uniqueSuffix = explicitFallback
-    .replace(/^chat-?/u, "")
-    .replace(/[^a-z0-9-]+/giu, "-")
-    .replace(/-+/g, "-")
-    .replace(/^-|-$/g, "")
-    .toLowerCase();
-  if (!uniqueSuffix.length) return promptFallback;
-
-  const maxPrefixLength = Math.max(1, 60 - uniqueSuffix.length - 1);
-  const prefix = promptFallback
-    .slice(0, maxPrefixLength)
-    .replace(/-+$/g, "");
-  return `${prefix}-${uniqueSuffix}`;
+  const suffix = genericSuffixFromLaneFallbackName(explicitFallback);
+  if (suffix) {
+    return genericLaneFallbackName(suffix);
+  }
+  return explicitFallback && !/^chat(?:-|$)/u.test(explicitFallback)
+    ? explicitFallback
+    : promptFallback;
 }
 
 function normalizeSuggestedLaneName(raw: string): string | null {
@@ -5357,14 +5346,23 @@ export function createAgentChatService(args: {
     prompt: string;
     systemPrompt?: string;
     timeoutMs?: number;
+    reasoningEffort?: string | null;
     taskType: "session_title" | "session_summary" | "handoff_summary" | "continuity_summary";
   }) => {
+    const config = resolveChatConfig();
+    const reasoningEffort = args.reasoningEffort
+      ?? (args.taskType === "session_title"
+        ? config.titleReasoningEffort
+        : args.taskType === "session_summary"
+          ? config.summaryReasoningEffort
+          : null);
     return await aiIntegrationService.summarizeTerminal({
       cwd: args.cwd,
       model: args.modelId,
       prompt: args.prompt,
       systemPrompt: args.systemPrompt,
       timeoutMs: args.timeoutMs,
+      reasoningEffort,
       taskType: args.taskType,
     });
   };
@@ -7660,10 +7658,18 @@ export function createAgentChatService(args: {
     const titleGenerationEnabled = si?.titles?.enabled
       ?? (typeof legacyChat.autoTitleEnabled === "boolean" ? legacyChat.autoTitleEnabled : undefined)
       ?? true;
-    const titleModelIdRaw = si?.titles?.modelId ?? legacyChat.autoTitleModelId;
+    const titleModelIdRaw = si?.titles && Object.prototype.hasOwnProperty.call(si.titles, "modelId")
+      ? si.titles.modelId
+      : legacyChat.autoTitleModelId;
     const titleModelId = typeof titleModelIdRaw === "string" && titleModelIdRaw.trim().length
       ? titleModelIdRaw.trim()
       : null;
+    const titleReasoningEffortRaw = si?.titles && Object.prototype.hasOwnProperty.call(si.titles, "reasoningEffort")
+      ? si.titles.reasoningEffort
+      : legacyChat.autoTitleReasoningEffort;
+    const titleReasoningEffort = typeof titleReasoningEffortRaw === "string" && titleReasoningEffortRaw.trim().length
+      ? titleReasoningEffortRaw.trim()
+      : null;
     const titleRefreshOnComplete = si?.titles?.refreshOnComplete
       ?? (typeof legacyChat.autoTitleRefreshOnComplete === "boolean" ? legacyChat.autoTitleRefreshOnComplete : undefined)
       ?? true;
@@ -7674,6 +7680,10 @@ export function createAgentChatService(args: {
     const summaryModelId = typeof summaryModelIdRaw === "string" && summaryModelIdRaw.trim().length
       ? summaryModelIdRaw.trim()
       : null;
+    const summaryReasoningEffortRaw = si?.summaries?.reasoningEffort;
+    const summaryReasoningEffort = typeof summaryReasoningEffortRaw === "string" && summaryReasoningEffortRaw.trim().length
+      ? summaryReasoningEffortRaw.trim()
+      : null;
 
     return {
       codexApprovalPolicy: approvalPolicy,
@@ -7683,9 +7693,11 @@ export function createAgentChatService(args: {
       sessionBudgetUsd,
       titleGenerationEnabled,
       titleModelId,
+      titleReasoningEffort,
       titleRefreshOnComplete,
       summaryEnabled,
       summaryModelId,
+      summaryReasoningEffort,
     };
   };
 
diff --git a/apps/desktop/src/main/services/config/projectConfigService.test.ts b/apps/desktop/src/main/services/config/projectConfigService.test.ts
index 3757bb16f..004806ab8 100644
--- a/apps/desktop/src/main/services/config/projectConfigService.test.ts
+++ b/apps/desktop/src/main/services/config/projectConfigService.test.ts
@@ -4,7 +4,7 @@ import path from "node:path";
 import YAML from "yaml";
 import { afterEach, describe, expect, it, vi } from "vitest";
 import { openKvDb } from "../state/kvDb";
-import { createProjectConfigService } from "./projectConfigService";
+import { createProjectConfigService, mergeAiConfig } from "./projectConfigService";
 
 function makeDb() {
   const store = new Map<string, unknown>();
@@ -528,10 +528,16 @@ describe("projectConfigService - AI mode migration", () => {
           },
           featureModelOverrides: {
             commit_messages: "openai/gpt-5.4-mini",
+            terminal_summaries: null,
+          },
+          featureReasoningOverrides: {
+            commit_messages: "minimal",
+            terminal_summaries: null,
           },
           chat: {
             autoTitleEnabled: true,
             autoTitleModelId: "openai/gpt-5.4-mini",
+            autoTitleReasoningEffort: "minimal",
             autoTitleRefreshOnComplete: false,
             autoAllowAskUser: false,
             codexSandbox: "workspace-write",
@@ -552,8 +558,12 @@ describe("projectConfigService - AI mode migration", () => {
     const snapshot = service.get();
     expect(snapshot.effective.ai?.features?.commit_messages).toBe(true);
     expect(snapshot.effective.ai?.featureModelOverrides?.commit_messages).toBe("openai/gpt-5.4-mini");
+    expect(snapshot.effective.ai?.featureModelOverrides?.terminal_summaries).toBeNull();
+    expect(snapshot.effective.ai?.featureReasoningOverrides?.commit_messages).toBe("minimal");
+    expect(snapshot.effective.ai?.featureReasoningOverrides?.terminal_summaries).toBeNull();
     expect(snapshot.effective.ai?.sessionIntelligence?.titles?.enabled).toBe(true);
     expect(snapshot.effective.ai?.sessionIntelligence?.titles?.modelId).toBe("openai/gpt-5.4-mini");
+    expect(snapshot.effective.ai?.sessionIntelligence?.titles?.reasoningEffort).toBe("minimal");
     expect(snapshot.effective.ai?.sessionIntelligence?.titles?.refreshOnComplete).toBe(false);
     expect(snapshot.effective.ai?.chat?.autoAllowAskUser).toBe(false);
     expect(snapshot.effective.ai?.chat?.codexSandbox).toBe("workspace-write");
@@ -566,15 +576,53 @@ describe("projectConfigService - AI mode migration", () => {
     const persisted = YAML.parse(fs.readFileSync(localPath, "utf8")) as Record<string, any>;
     expect(persisted.ai?.features?.commit_messages).toBe(true);
     expect(persisted.ai?.featureModelOverrides?.commit_messages).toBe("openai/gpt-5.4-mini");
+    expect(persisted.ai?.featureModelOverrides?.terminal_summaries).toBeNull();
+    expect(persisted.ai?.featureReasoningOverrides?.commit_messages).toBe("minimal");
+    expect(persisted.ai?.featureReasoningOverrides?.terminal_summaries).toBeNull();
     expect(persisted.ai?.chat?.autoTitleEnabled).toBeUndefined();
     expect(persisted.ai?.chat?.autoTitleModelId).toBeUndefined();
+    expect(persisted.ai?.chat?.autoTitleReasoningEffort).toBeUndefined();
     expect(persisted.ai?.chat?.autoTitleRefreshOnComplete).toBeUndefined();
     expect(persisted.ai?.sessionIntelligence?.titles?.enabled).toBe(true);
     expect(persisted.ai?.sessionIntelligence?.titles?.modelId).toBe("openai/gpt-5.4-mini");
+    expect(persisted.ai?.sessionIntelligence?.titles?.reasoningEffort).toBe("minimal");
     expect(persisted.ai?.sessionIntelligence?.titles?.refreshOnComplete).toBe(false);
     expect(persisted.ai?.chat?.autoAllowAskUser).toBe(false);
     expect(persisted.ai?.chat?.codexSandbox).toBe("workspace-write");
   });
+
+  it("clears session intelligence reasoning overrides with null", () => {
+    const merged = mergeAiConfig({
+      sessionIntelligence: {
+        titles: { modelId: "openai/gpt-5.4-mini", reasoningEffort: "minimal" },
+        summaries: { modelId: "openai/gpt-5.4-mini", reasoningEffort: "low" },
+      },
+      featureModelOverrides: {
+        terminal_summaries: "openai/gpt-5.4-mini",
+      },
+      featureReasoningOverrides: {
+        terminal_summaries: "low",
+      },
+    }, {
+      sessionIntelligence: {
+        titles: { modelId: null, reasoningEffort: null },
+        summaries: { modelId: null, reasoningEffort: null },
+      },
+      featureModelOverrides: {
+        terminal_summaries: null,
+      },
+      featureReasoningOverrides: {
+        terminal_summaries: null,
+      },
+    });
+
+    expect(merged?.sessionIntelligence?.titles?.modelId).toBeNull();
+    expect(merged?.sessionIntelligence?.summaries?.modelId).toBeNull();
+    expect(merged?.featureModelOverrides?.terminal_summaries).toBeNull();
+    expect(merged?.sessionIntelligence?.titles?.reasoningEffort).toBeNull();
+    expect(merged?.sessionIntelligence?.summaries?.reasoningEffort).toBeNull();
+    expect(merged?.featureReasoningOverrides?.terminal_summaries).toBeNull();
+  });
 });
 
 describe("projectConfigService - PR transcript gists", () => {
diff --git a/apps/desktop/src/main/services/config/projectConfigService.ts b/apps/desktop/src/main/services/config/projectConfigService.ts
index 85cff7d5d..e27e5ad76 100644
--- a/apps/desktop/src/main/services/config/projectConfigService.ts
+++ b/apps/desktop/src/main/services/config/projectConfigService.ts
@@ -365,6 +365,10 @@ function coerceSessionIntelligenceConfig(value: unknown): AiConfig["sessionIntel
     if (enabled != null) titles.enabled = enabled;
     const modelId = asString(titlesRaw.modelId)?.trim();
     if (modelId) titles.modelId = modelId;
+    else if (titlesRaw.modelId === null) titles.modelId = null;
+    const reasoningEffort = asString(titlesRaw.reasoningEffort)?.trim();
+    if (reasoningEffort) titles.reasoningEffort = reasoningEffort;
+    else if (titlesRaw.reasoningEffort === null) titles.reasoningEffort = null;
     const refreshOnComplete = asBool(titlesRaw.refreshOnComplete);
     if (refreshOnComplete != null) titles.refreshOnComplete = refreshOnComplete;
     if (Object.keys(titles).length) out.titles = titles;
@@ -377,6 +381,10 @@ function coerceSessionIntelligenceConfig(value: unknown): AiConfig["sessionIntel
     if (enabled != null) summaries.enabled = enabled;
     const modelId = asString(summariesRaw.modelId)?.trim();
     if (modelId) summaries.modelId = modelId;
+    else if (summariesRaw.modelId === null) summaries.modelId = null;
+    const reasoningEffort = asString(summariesRaw.reasoningEffort)?.trim();
+    if (reasoningEffort) summaries.reasoningEffort = reasoningEffort;
+    else if (summariesRaw.reasoningEffort === null) summaries.reasoningEffort = null;
     if (Object.keys(summaries).length) out.summaries = summaries;
   }
 
@@ -385,14 +393,31 @@ function coerceSessionIntelligenceConfig(value: unknown): AiConfig["sessionIntel
 
 function coerceFeatureModelOverrides(value: unknown): AiConfig["featureModelOverrides"] {
   if (!isRecord(value)) return undefined;
-  const featureModelOverrides: Partial<Record<AiFeatureKey, string>> = {};
+  const featureModelOverrides: Partial<Record<AiFeatureKey, string | null>> = {};
   for (const key of AI_FEATURE_KEYS) {
-    const modelId = asString(value[key])?.trim();
+    const rawValue = value[key];
+    const modelId = asString(rawValue)?.trim();
     if (modelId) featureModelOverrides[key] = modelId;
+    else if (rawValue === null) featureModelOverrides[key] = null;
   }
   return Object.keys(featureModelOverrides).length ? featureModelOverrides : undefined;
 }
 
+function coerceFeatureReasoningOverrides(value: unknown): AiConfig["featureReasoningOverrides"] {
+  if (!isRecord(value)) return undefined;
+  const featureReasoningOverrides: Partial<Record<AiFeatureKey, string | null>> = {};
+  for (const key of AI_FEATURE_KEYS) {
+    const rawValue = value[key];
+    const reasoningEffort = asString(rawValue)?.trim();
+    if (reasoningEffort) {
+      featureReasoningOverrides[key] = reasoningEffort;
+    } else if (rawValue === null) {
+      featureReasoningOverrides[key] = null;
+    }
+  }
+  return Object.keys(featureReasoningOverrides).length ? featureReasoningOverrides : undefined;
+}
+
 function parseReadiness(value: unknown): ConfigProcessReadiness | undefined {
   if (!isRecord(value)) return undefined;
   const type = asString(value.type);
@@ -1322,6 +1347,8 @@ function coerceAiConfig(value: unknown): AiConfig | undefined {
 
   const featureModelOverrides = coerceFeatureModelOverrides(value.featureModelOverrides);
   if (featureModelOverrides) out.featureModelOverrides = featureModelOverrides;
+  const featureReasoningOverrides = coerceFeatureReasoningOverrides(value.featureReasoningOverrides);
+  if (featureReasoningOverrides) out.featureReasoningOverrides = featureReasoningOverrides;
 
   const permissionsRaw = isRecord(value.permissions) ? value.permissions : null;
   if (permissionsRaw) {
@@ -1506,12 +1533,15 @@ function coerceAiConfig(value: unknown): AiConfig | undefined {
     if (autoTitleEnabled != null) titles.enabled = autoTitleEnabled;
     const autoTitleModelId = asString(chatRaw?.autoTitleModelId)?.trim();
     if (autoTitleModelId) titles.modelId = autoTitleModelId;
+    const autoTitleReasoningEffort = asString(chatRaw?.autoTitleReasoningEffort)?.trim();
+    if (autoTitleReasoningEffort) titles.reasoningEffort = autoTitleReasoningEffort;
     const autoTitleRefreshOnComplete = asBool(chatRaw?.autoTitleRefreshOnComplete);
     if (autoTitleRefreshOnComplete != null) titles.refreshOnComplete = autoTitleRefreshOnComplete;
 
     const summaries: NonNullable<NonNullable<AiConfig["sessionIntelligence"]>["summaries"]> = {};
-    if (featureModelOverrides?.terminal_summaries) {
-      summaries.modelId = featureModelOverrides.terminal_summaries;
+    const terminalSummariesModelId = featureModelOverrides?.terminal_summaries;
+    if (typeof terminalSummariesModelId === "string" && terminalSummariesModelId.trim().length) {
+      summaries.modelId = terminalSummariesModelId.trim();
     }
 
     const migrated: NonNullable<AiConfig["sessionIntelligence"]> = {
@@ -1882,6 +1912,10 @@ export function mergeAiConfig(sharedAi?: AiConfig, localAi?: Partial<AiConfig>):
     ...(sharedAi?.featureModelOverrides ?? {}),
     ...(localAi?.featureModelOverrides ?? {})
   };
+  const featureReasoningOverrides = {
+    ...(sharedAi?.featureReasoningOverrides ?? {}),
+    ...(localAi?.featureReasoningOverrides ?? {})
+  };
   const apiKeys = {
     ...(sharedAi?.apiKeys ?? {}),
     ...(localAi?.apiKeys ?? {})
@@ -1910,6 +1944,7 @@ export function mergeAiConfig(sharedAi?: AiConfig, localAi?: Partial<AiConfig>):
     ...(Object.keys(chat).length ? { chat } : {}),
     ...(sessionIntelligence ? { sessionIntelligence } : {}),
     ...(Object.keys(featureModelOverrides).length ? { featureModelOverrides } : {}),
+    ...(Object.keys(featureReasoningOverrides).length ? { featureReasoningOverrides } : {}),
     ...(Object.keys(apiKeys).length ? { apiKeys } : {}),
     ...(localProvidersEntries.length ? { localProviders } : {}),
     ...(workerSafety ? { workerSafety } : {}),
diff --git a/apps/desktop/src/main/services/pty/ptyService.test.ts b/apps/desktop/src/main/services/pty/ptyService.test.ts
index e3046d6a6..2cb1c1cc6 100644
--- a/apps/desktop/src/main/services/pty/ptyService.test.ts
+++ b/apps/desktop/src/main/services/pty/ptyService.test.ts
@@ -994,6 +994,46 @@ describe("ptyService", () => {
       }
     });
 
+    it("uses agent CLI initialInput as the first-prompt title seed", async () => {
+      vi.useFakeTimers();
+      try {
+        const aiIntegrationService = {
+          getMode: vi.fn(() => "subscription"),
+          summarizeTerminal: vi.fn(async () => ({ text: "Print cwd" })),
+        };
+        const { service, mockPty, sessionService } = createHarness({ aiIntegrationService });
+
+        await service.create({
+          laneId: "lane-1",
+          title: "Codex CLI",
+          cols: 80,
+          rows: 24,
+          toolType: "codex",
+          command: "codex",
+          args: ["--no-alt-screen"],
+          startupCommand: "codex --no-alt-screen",
+          initialInput: "print cwd",
+        });
+
+        const createdSessionId = (sessionService.create as ReturnType<typeof vi.fn>).mock.calls[0]?.[0]?.sessionId;
+        expect(createdSessionId).toBeTruthy();
+
+        mockPty._emitter.emit("data", "\x1b[2J\x1b[Hmodel: gpt-5.4 medium\nMCP startup incomplete (failed: linear)\n› ");
+        await vi.advanceTimersByTimeAsync(600);
+        await Promise.resolve();
+
+        expect(sessionService.get(createdSessionId!)?.goal).toBe("print cwd");
+        expect(aiIntegrationService.summarizeTerminal).toHaveBeenCalledWith(
+          expect.objectContaining({
+            prompt: expect.stringContaining("print cwd"),
+            taskType: "session_title",
+          }),
+        );
+      } finally {
+        vi.useRealTimers();
+      }
+    });
+
     it("does not send Codex initialInput into the update prompt", async () => {
       vi.useFakeTimers();
       try {
@@ -3118,6 +3158,7 @@ describe("ptyService", () => {
     });
 
     it.each([
+      ["claude", "Claude Code"],
       ["codex", "Codex session"],
       ["cursor-cli", "Cursor Agent CLI"],
       ["droid", "Factory Droid CLI"],
@@ -3289,7 +3330,7 @@ describe("ptyService", () => {
       }
     });
 
-    it("does not call ADE AI title generation for Claude CLI sessions", async () => {
+    it("uses ADE first-prompt AI title generation for Claude CLI sessions", async () => {
       const aiIntegrationService = {
         getMode: vi.fn(() => "subscription"),
         summarizeTerminal: vi.fn(async () => ({ text: "ADE generated title" })),
@@ -3310,7 +3351,12 @@ describe("ptyService", () => {
       await Promise.resolve();
 
       expect(sessionService.get(createdSessionId)?.title).toBe("Fix the flaky login tests");
-      expect(aiIntegrationService.summarizeTerminal).not.toHaveBeenCalled();
+      expect(aiIntegrationService.summarizeTerminal).toHaveBeenCalledWith(
+        expect.objectContaining({
+          prompt: expect.stringContaining("Fix the flaky login tests"),
+          taskType: "session_title",
+        }),
+      );
     });
 
     it("treats legacy slash-command CLI titles as placeholders", async () => {
diff --git a/apps/desktop/src/main/services/pty/ptyService.ts b/apps/desktop/src/main/services/pty/ptyService.ts
index 610161b1e..4e5543959 100644
--- a/apps/desktop/src/main/services/pty/ptyService.ts
+++ b/apps/desktop/src/main/services/pty/ptyService.ts
@@ -1376,6 +1376,12 @@ export function createPtyService({
     return typeof raw === "string" && raw.trim().length ? raw.trim() : undefined;
   };
 
+  const resolveTitleReasoningEffort = (): string | null => {
+    const si = getSessionIntelligence();
+    const raw = si?.titles?.reasoningEffort;
+    return typeof raw === "string" && raw.trim().length ? raw.trim() : null;
+  };
+
   const tryCliUserTitleFromWrite = (entry: PtyEntry, data: string): void => {
     if (!CLI_USER_TITLE_TOOL_TYPES.has(entry.toolTypeHint ?? "shell")) return;
     if (entry.cliUserTitleCommitted || entry.disposed) return;
@@ -1411,15 +1417,12 @@ export function createPtyService({
           sessionService.updateMeta({ sessionId: entry.sessionId, title: fallbackTitle, manuallyNamed: false });
         }
       }
-      // Claude Code writes its own generated `ai-title` into local session
-      // storage. Keep ADE's prompt summarizer out of this path so that native
-      // Claude names win when they arrive.
-      if (isClaudeTrackedCliToolType(entry.toolTypeHint)) return;
       if (!aiIntegrationService || aiIntegrationService.getMode() === "guest") return;
       if (!isTitleGenerationEnabled()) return;
 
       const laneName = session.laneName?.trim() || "Current lane";
       const titleModelId = resolveTitleModelId();
+      const titleReasoningEffort = resolveTitleReasoningEffort();
       const prompt = [
         "Write a concise title for this CLI coding session.",
         "Return only plain text, max 80 characters, no punctuation at the end.",
@@ -1438,6 +1441,7 @@ export function createPtyService({
           taskType: "session_title",
           timeoutMs: PTY_AI_TITLE_TIMEOUT_MS,
           ...(titleModelId ? { model: titleModelId } : {}),
+          ...(titleReasoningEffort ? { reasoningEffort: titleReasoningEffort } : {}),
         })
         .then((result) => {
           if (entry.disposed) return;
@@ -1624,11 +1628,15 @@ export function createPtyService({
             const summaryModelId = typeof si?.summaries?.modelId === "string" && si.summaries.modelId.trim().length
               ? si.summaries.modelId.trim()
               : undefined;
+            const summaryReasoningEffort = typeof si?.summaries?.reasoningEffort === "string" && si.summaries.reasoningEffort.trim().length
+              ? si.summaries.reasoningEffort.trim()
+              : undefined;
 
             const aiSummary = await aiIntegrationService!.summarizeTerminal({
               cwd: summaryCwd || laneService.getLaneBaseAndBranch(session.laneId).worktreePath,
               prompt,
               ...(summaryModelId ? { model: summaryModelId } : {}),
+              ...(summaryReasoningEffort ? { reasoningEffort: summaryReasoningEffort } : {}),
             });
             const text = aiSummary.text.trim();
             if (text.length) {
@@ -1664,12 +1672,14 @@ export function createPtyService({
               ].filter(Boolean).join("\n");
 
               const titleModelId = resolveTitleModelId();
+              const titleReasoningEffort = resolveTitleReasoningEffort();
               const titleResult = await aiIntegrationService!.summarizeTerminal({
                 cwd: summaryCwd || laneService.getLaneBaseAndBranch(session.laneId).worktreePath,
                 prompt: titlePrompt,
                 taskType: "session_title",
                 timeoutMs: PTY_AI_TITLE_TIMEOUT_MS,
                 ...(titleModelId ? { model: titleModelId } : {}),
+                ...(titleReasoningEffort ? { reasoningEffort: titleReasoningEffort } : {}),
               });
               const finalTitle = sanitizeGeneratedCliTitle(titleResult.text);
               if (finalTitle) {
@@ -3965,6 +3975,7 @@ export function createPtyService({
             if (provider) {
               const submittedInitialInput = normalizedInitialInput.trim();
               if (submittedInitialInput.length > 0) {
+                tryCliUserTitleFromWrite(entry, `${submittedInitialInput}\r`);
                 const wrote = await writeAgentCliInput((data) => {
                   pty.write(data);
                   return true;
@@ -4084,6 +4095,7 @@ export function createPtyService({
           ].join("\n");
 
           const titleModelId = resolveTitleModelId();
+          const titleReasoningEffort = resolveTitleReasoningEffort();
           capturedAi
             .summarizeTerminal({
               cwd: entry.boundCwd || entry.laneWorktreePath,
@@ -4091,6 +4103,7 @@ export function createPtyService({
               taskType: "session_title",
               timeoutMs: PTY_AI_TITLE_TIMEOUT_MS,
               ...(titleModelId ? { model: titleModelId } : {}),
+              ...(titleReasoningEffort ? { reasoningEffort: titleReasoningEffort } : {}),
             })
             .then((result) => {
               const title = sanitizeGeneratedCliTitle(result.text);
diff --git a/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.test.ts b/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.test.ts
index b651bba6f..7c118129e 100644
--- a/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.test.ts
+++ b/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.test.ts
@@ -636,7 +636,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
         command.includes("printf '%s\\n' '2.0.0' > $HOME/.ade/bin/ade.version")
       ) return ok("");
       if (command.includes("$HOME/.ade/bin/ade --version")) return ok("ade 2.0.0\n");
-      if (command.includes("$HOME/.ade/bin/ade runtime stop --text")) return ok("");
       return defaultRemoteBootstrapCommand(command);
     });
 
@@ -670,12 +669,10 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     expect(commands).toContain(
       'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade --version',
     );
-    expect(commands).toContain(
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade runtime stop --text >/dev/null 2>&1 || true',
-    );
+    expect(commands.some((command) => command.includes("runtime stop --text"))).toBe(false);
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
     expect(initializeMock).toHaveBeenCalledWith("ade-desktop-remote", APP_VERSION);
     expect(callMock).toHaveBeenCalledWith("projects.list", {});
@@ -773,7 +770,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     )).toBe(true);
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PTY_HOST_WORKER_PATH="$HOME/.ade/runtime/ptyHostWorker.cjs" ADE_PTY_HOST_WORKER_NODE=\'/usr/local/bin/node\' $HOME/.ade/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PTY_HOST_WORKER_PATH="$HOME/.ade/runtime/ptyHostWorker.cjs" ADE_PTY_HOST_WORKER_NODE=\'/usr/local/bin/node\' $HOME/.ade/bin/ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
     expect(connected.result).toMatchObject({
       arch: "linux-x64",
@@ -829,7 +826,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     expect(fakeSsh.sftpWrapper.fastPut).not.toHaveBeenCalled();
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PTY_HOST_WORKER_COMMAND="$HOME/.ade/bin/ade" $HOME/.ade/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PTY_HOST_WORKER_COMMAND="$HOME/.ade/bin/ade" $HOME/.ade/bin/ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
     expect(connected.result).toMatchObject({
       arch: "linux-x64",
@@ -948,7 +945,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
         command.includes(`printf '%s\\n' '${APP_VERSION}' > $HOME/.ade/bin/ade.version`)
       ) return ok("");
       if (command.includes("$HOME/.ade/bin/ade --version")) return ok(`ade ${APP_VERSION}\n`);
-      if (command.includes("$HOME/.ade/bin/ade runtime stop --text")) return ok("");
       return defaultRemoteBootstrapCommand(command);
     });
 
@@ -968,7 +964,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     )).toBe(true);
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
     expect(connected.result).toMatchObject({
       arch: "linux-x64",
@@ -1018,7 +1014,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
         command.includes(`printf '%s\\n' '${APP_VERSION}' > $HOME/.ade/bin/ade.version`)
       ) return ok("");
       if (command.includes("$HOME/.ade/bin/ade --version")) return ok(`ade ${APP_VERSION}\n`);
-      if (command.includes("$HOME/.ade/bin/ade runtime stop --text")) return ok("");
       return defaultRemoteBootstrapCommand(command);
     });
 
@@ -1037,7 +1032,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     )).toBe(true);
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
     expect(connected.result).toMatchObject({
       arch: "linux-x64",
@@ -1214,7 +1209,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
       ) return ok("");
       if (command === "codesign --force --sign - $HOME/.ade-alpha/bin/ade") return ok("");
       if (command.includes("$HOME/.ade-alpha/bin/ade --version")) return ok("ade 2.0.0\n");
-      if (command.includes("$HOME/.ade-alpha/bin/ade runtime stop --text")) return ok("");
       return defaultRemoteBootstrapCommand(command);
     });
 
@@ -1249,7 +1243,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     );
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade-alpha/runtime/darwin-arm64/node_modules${NODE_PATH:+:$NODE_PATH}" $HOME/.ade-alpha/bin/ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade-alpha/runtime/darwin-arm64/node_modules${NODE_PATH:+:$NODE_PATH}" $HOME/.ade-alpha/bin/ade --socket $HOME/.ade-alpha/sock/ade.sock rpc --stdio',
     );
   });
 
@@ -1302,7 +1296,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     expect(fakeSsh.exec).not.toHaveBeenCalled();
     expect(openSshRuntimeTransportMock).toHaveBeenCalledWith(
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 ade --socket $HOME/.ade-alpha/sock/ade.sock rpc --stdio',
     );
     expect(connected.result.version).toBe("1.9.0-alpha.4");
     expect(connected.result.compatibilityWarnings).toEqual([
@@ -1365,12 +1359,12 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     expect(openSshRuntimeTransportMock).toHaveBeenNthCalledWith(
       1,
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade-beta" PATH="$HOME/.ade-beta/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="beta" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade-beta" PATH="$HOME/.ade-beta/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="beta" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 ade --socket $HOME/.ade-beta/sock/ade.sock rpc --stdio',
     );
     expect(openSshRuntimeTransportMock).toHaveBeenNthCalledWith(
       2,
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade-alpha/runtime/linux-x64/node_modules${NODE_PATH:+:$NODE_PATH}" ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade-alpha" PATH="$HOME/.ade-alpha/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_PACKAGE_CHANNEL="alpha" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade-alpha/runtime/linux-x64/node_modules${NODE_PATH:+:$NODE_PATH}" ade --socket $HOME/.ade-alpha/sock/ade.sock rpc --stdio',
     );
     expect(connected.result).toMatchObject({
       version: "1.9.0-alpha.4",
@@ -1434,7 +1428,7 @@ describe("bootstrapRemoteRuntime upload flow", () => {
     expect(openSshRuntimeTransportMock).toHaveBeenNthCalledWith(
       2,
       fakeSsh.ssh,
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade/runtime/linux-x64/node_modules${NODE_PATH:+:$NODE_PATH}" ade rpc --stdio',
+      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" ADE_DISABLE_RUNTIME_SERVICE_INSTALL=1 NODE_PATH="$HOME/.ade/runtime/linux-x64/node_modules${NODE_PATH:+:$NODE_PATH}" ade --socket $HOME/.ade/sock/ade.sock rpc --stdio',
     );
   });
 
@@ -1463,7 +1457,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
         command.includes("shasum -a 256 $HOME/.ade/bin/ade") &&
         command.includes("echo ok")
       ) return ok("ok\n");
-      if (command.includes("$HOME/.ade/bin/ade runtime stop --text")) return ok("");
       return defaultRemoteBootstrapCommand(command);
     });
     initializeMock.mockResolvedValueOnce({
@@ -1494,8 +1487,6 @@ describe("bootstrapRemoteRuntime upload flow", () => {
 
     expect(fakeSsh.exec).not.toHaveBeenCalled();
     expect(openSshRuntimeTransportMock).toHaveBeenCalledTimes(1);
-    expect(commands).not.toContain(
-      'ADE_HOME="$HOME/.ade" PATH="$HOME/.ade/bin:$HOME/.local/bin:$HOME/.npm-global/bin${PATH:+:$PATH}" ADE_DEFAULT_ROLE="cto" $HOME/.ade/bin/ade runtime stop --text >/dev/null 2>&1 || true',
-    );
+    expect(commands.some((command) => command.includes("runtime stop --text"))).toBe(false);
   });
 });
diff --git a/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.ts b/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.ts
index f020a8fe4..c149ed28e 100644
--- a/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.ts
+++ b/apps/desktop/src/main/services/remoteRuntime/remoteBootstrap.ts
@@ -170,6 +170,7 @@ type RemoteRuntimeLayout = {
   binDirRelative: string;
   runtimeDirExpr: string;
   runtimeDirRelative: string;
+  socketExpr: string;
   binaryExpr: string;
   binaryRelative: string;
   versionExpr: string;
@@ -202,6 +203,7 @@ export function resolveRemoteRuntimeLayout(env: NodeJS.ProcessEnv = process.env)
     binDirRelative: `${homeDirName}/bin`,
     runtimeDirExpr,
     runtimeDirRelative: `${homeDirName}/runtime`,
+    socketExpr: `${homeDirExpr}/sock/ade.sock`,
     binaryExpr: `${binDirExpr}/ade`,
     binaryRelative: `${homeDirName}/bin/ade`,
     versionExpr: `${binDirExpr}/ade.version`,
@@ -1077,17 +1079,16 @@ async function signUploadedRuntimeBinaryIfNeeded(client: Client, layout: RemoteR
   }
 }
 
-async function stopRemoteRuntimeDaemon(client: Client, layout: RemoteRuntimeLayout, runtimeEnvPrefix: string): Promise<void> {
-  await execSsh(
-    client,
-    `${runtimeEnvPrefix}${layout.binaryExpr} runtime stop --text >/dev/null 2>&1 || true`,
-  );
-}
-
 function runtimeErrorMessage(error: unknown): string {
   return error instanceof Error ? error.message : String(error);
 }
 
+function remoteRuntimeRpcCommand(layout: RemoteRuntimeLayout, runtimeEnvPrefix: string, binaryExpr: string): string {
+  // Use the channel's explicit socket so a freshly uploaded helper CLI does not
+  // treat the desktop-owned brain as a build-hash mismatch and recycle it.
+  return `${runtimeEnvPrefix}${binaryExpr} --socket ${layout.socketExpr} rpc --stdio`;
+}
+
 async function openValidatedRuntimeClient(args: {
   ssh: Client;
   command: string;
@@ -1336,7 +1337,6 @@ export async function bootstrapRemoteRuntime(args: {
 
     if (runtimeUploaded) {
       await verifyUploadedRuntime();
-      await stopRemoteRuntimeDaemon(ssh, layout, runtimeEnvPrefix);
     }
 
     if (!runtimeVersion) {
@@ -1348,8 +1348,8 @@ export async function bootstrapRemoteRuntime(args: {
     }
 
     const command = localBinary || runtimeUploaded
-      ? `${runtimeEnvPrefix}${layout.binaryExpr} rpc --stdio`
-      : `${runtimeEnvPrefix}ade rpc --stdio`;
+      ? remoteRuntimeRpcCommand(layout, runtimeEnvPrefix, layout.binaryExpr)
+      : remoteRuntimeRpcCommand(layout, runtimeEnvPrefix, "ade");
     let openedRuntime: Awaited<ReturnType<typeof openValidatedRuntimeClient>> | null = null;
     const expectedVersion = localBinary || runtimeUploaded ? args.appVersion : null;
     try {
@@ -1382,7 +1382,7 @@ export async function bootstrapRemoteRuntime(args: {
           layout: candidateLayout,
           disableRuntimeServiceInstall: true,
         });
-        const candidateCommand = `${candidateRuntimeEnvPrefix}ade rpc --stdio`;
+        const candidateCommand = remoteRuntimeRpcCommand(candidateLayout, candidateRuntimeEnvPrefix, "ade");
         try {
           openedRuntime = await openValidatedRuntimeClient({
             ssh,
diff --git a/apps/desktop/src/main/services/sync/syncService.test.ts b/apps/desktop/src/main/services/sync/syncService.test.ts
index 686461094..f7b1042ad 100644
--- a/apps/desktop/src/main/services/sync/syncService.test.ts
+++ b/apps/desktop/src/main/services/sync/syncService.test.ts
@@ -6,10 +6,29 @@ import { isCrsqliteAvailable } from "../state/crsqliteExtension";
 import { openKvDb } from "../state/kvDb";
 import { createSyncService } from "./syncService";
 
-const { createDefaultSyncHostServiceMock, createSyncHostServiceMock, syncHostServiceMockState } = vi.hoisted(() => {
+const { acquireSyncHostSingletonMock, createDefaultSyncHostServiceMock, createSyncHostServiceMock, syncHostServiceMockState } = vi.hoisted(() => {
   const syncHostServiceMockState = {
     port: 8787,
   };
+  const acquireSyncHostSingletonMock = vi.fn(() => ({
+    owner: {
+      id: "test-sync-host",
+      pid: process.pid,
+      port: null,
+      appName: "ADE Test",
+      packageChannel: "test",
+      adeHome: null,
+      serviceName: null,
+      socketPath: null,
+      projectRoot: null,
+      commandLine: null,
+      quitCommand: "",
+      createdAt: "2026-03-15T00:00:00.000Z",
+      updatedAt: "2026-03-15T00:00:00.000Z",
+    },
+    updatePort: vi.fn(),
+    dispose: vi.fn(),
+  }));
   const createDefaultSyncHostServiceMock = () => ({
     async waitUntilListening() {
       return syncHostServiceMockState.port;
@@ -44,6 +63,7 @@ const { createDefaultSyncHostServiceMock, createSyncHostServiceMock, syncHostSer
     async dispose() {},
   });
   return {
+    acquireSyncHostSingletonMock,
     syncHostServiceMockState,
     createDefaultSyncHostServiceMock,
     createSyncHostServiceMock: vi.fn(createDefaultSyncHostServiceMock),
@@ -58,6 +78,10 @@ vi.mock("../../../../../ade-cli/src/services/sync/syncHostService", () => ({
   SYNC_TAILNET_DISCOVERY_SERVICE_PORT: 8787,
 }));
 
+vi.mock("../../../../../ade-cli/src/services/sync/syncHostSingleton", () => ({
+  acquireSyncHostSingleton: acquireSyncHostSingletonMock,
+}));
+
 function createLogger() {
   return {
     debug: () => {},
@@ -853,8 +877,18 @@ describe.skipIf(!isCrsqliteAvailable())("syncService", () => {
 
     await service.initialize();
 
-    expect(createSyncHostServiceMock.mock.calls.map((call: any[]) => call[0]?.port)).toEqual([8787, 8788]);
-    expect(disposeFirstAttempt).toHaveBeenCalledTimes(1);
+    expect(createSyncHostServiceMock.mock.calls.map((call: any[]) => call[0]?.port)).toEqual([
+      8787,
+      8787,
+      8787,
+      8787,
+      8787,
+      8787,
+      8787,
+      8787,
+      8788,
+    ]);
+    expect(disposeFirstAttempt).toHaveBeenCalledTimes(8);
     expect(service.getHostService()?.getPort()).toBe(8788);
   }, 30_000);
 
diff --git a/apps/desktop/src/renderer/components/chat/AgentChatPane.test.tsx b/apps/desktop/src/renderer/components/chat/AgentChatPane.test.tsx
index 1837efe0b..bbc6c8f4a 100644
--- a/apps/desktop/src/renderer/components/chat/AgentChatPane.test.tsx
+++ b/apps/desktop/src/renderer/components/chat/AgentChatPane.test.tsx
@@ -3435,7 +3435,7 @@ describe("AgentChatPane submit recovery", () => {
         laneId: "lane-primary",
         prompt: "Fix auto create lane routing.",
         modelId: "openai/gpt-5.4",
-        fallbackName: expect.stringMatching(/^chat-\d{8}-\d{6}$/),
+        fallbackName: "fix-auto-create-lane-routing",
       }));
       expect(createLane).toHaveBeenCalledWith({
         name: "fix-auto-create-flow",
@@ -3568,7 +3568,7 @@ describe("AgentChatPane submit recovery", () => {
 
       await waitFor(() => {
         expect(createLane).toHaveBeenCalledWith({
-          name: expect.stringMatching(/^chat-\d{8}-\d{6}$/),
+          name: "keep-going-even-if-naming",
           baseBranch: "origin/main",
         });
         expect(create).toHaveBeenCalledWith(expect.objectContaining({ laneId: "lane-created" }));
@@ -4127,7 +4127,7 @@ describe("AgentChatPane submit recovery", () => {
 
     await waitFor(() => {
       expect(suggestLaneName).toHaveBeenCalled();
-      expect(screen.getByText(/Choosing a branch name/i)).toBeTruthy();
+      expect(screen.getByText(/Naming lane with/i)).toBeTruthy();
       expect((screen.getByRole("button", { name: "Send" }) as HTMLButtonElement).disabled).toBe(true);
       expect((screen.getByRole("button", { name: "Auto-create in background" }) as HTMLButtonElement).disabled).toBe(true);
     });
@@ -4182,14 +4182,14 @@ describe("AgentChatPane submit recovery", () => {
 
     await waitFor(() => {
       expect(suggestLaneName).toHaveBeenCalledTimes(1);
-      expect(screen.getByText(/Choosing a branch name/i)).toBeTruthy();
+      expect(screen.getByText(/Naming lane with/i)).toBeTruthy();
     });
     expect(screen.queryByRole("button", { name: "Dismiss launch status" })).toBeNull();
 
     rendered.unmount();
     renderAutoCreateDraftPane();
 
-    expect(await screen.findByText(/Choosing a branch name/i)).toBeTruthy();
+    expect(await screen.findByText(/Naming lane with/i)).toBeTruthy();
     expect(screen.queryByRole("button", { name: "Dismiss launch status" })).toBeNull();
 
     await act(async () => {
@@ -4226,7 +4226,9 @@ describe("AgentChatPane submit recovery", () => {
           laneId: null,
           laneName: null,
           sessionId: null,
+          namingModelId: null,
           error: null,
+          warning: null,
           autoOpen: false,
           createdAtMs: Date.now() - DRAFT_LAUNCH_JOB_STALE_AFTER_MS - 1,
           snapshot: {
@@ -4289,7 +4291,7 @@ describe("AgentChatPane submit recovery", () => {
 
       await waitFor(() => {
         expect(suggestLaneName).toHaveBeenCalledTimes(1);
-        expect(screen.getByText(/Choosing a branch name/i)).toBeTruthy();
+        expect(screen.getByText(/Naming lane with/i)).toBeTruthy();
       });
 
       const scopeKey = draftLaunchJobsScopeKeyForTest({
@@ -4395,7 +4397,7 @@ describe("AgentChatPane submit recovery", () => {
 
     await waitFor(() => {
       expect(suggestLaneName).toHaveBeenCalledTimes(1);
-      expect(screen.getAllByText(/Choosing a branch name/i)).toHaveLength(1);
+      expect(screen.getAllByText(/Naming lane with/i)).toHaveLength(1);
     });
   });
 
@@ -4476,9 +4478,9 @@ describe("AgentChatPane submit recovery", () => {
 
     await waitFor(() => {
       expect(suggestLaneName).toHaveBeenCalledTimes(1);
-      expect(within(paneOne).getByText(/Choosing a branch name/i)).toBeTruthy();
+      expect(within(paneOne).getByText(/Naming lane with/i)).toBeTruthy();
     });
-    expect(within(paneTwo).queryByText(/Choosing a branch name/i)).toBeNull();
+    expect(within(paneTwo).queryByText(/Naming lane with/i)).toBeNull();
     expect(within(paneTwo).queryByTestId("draft-launch-job")).toBeNull();
   });
 
@@ -4510,7 +4512,7 @@ describe("AgentChatPane submit recovery", () => {
     }
 
     expect(screen.getAllByTestId("draft-launch-job")).toHaveLength(9);
-    expect(screen.getAllByText(/Choosing a branch name/i)).toHaveLength(9);
+    expect(screen.getAllByText(/Naming lane with/i)).toHaveLength(9);
   });
 
   it("allows multiple background auto-create launches to stay pending at the same time", async () => {
@@ -4552,7 +4554,7 @@ describe("AgentChatPane submit recovery", () => {
     fireEvent.click(await screen.findByRole("button", { name: "Auto-create in background" }));
     await waitFor(() => {
       expect(suggestLaneName).toHaveBeenCalledTimes(2);
-      expect(screen.getAllByText(/Choosing a branch name/i)).toHaveLength(2);
+      expect(screen.getAllByText(/Naming lane with/i)).toHaveLength(2);
     });
 
     await act(async () => {
@@ -5785,7 +5787,7 @@ describe("AgentChatPane submit recovery", () => {
         laneId: "lane-1",
         prompt: "Fix the login bug",
         modelId: "openai/gpt-5.4",
-        fallbackName: expect.stringMatching(/^chat-\d{8}-\d{6}$/),
+        fallbackName: "fix-login-bug",
       }));
       expect(createChild).toHaveBeenCalledTimes(2);
     });
diff --git a/apps/desktop/src/renderer/components/chat/AgentChatPane.tsx b/apps/desktop/src/renderer/components/chat/AgentChatPane.tsx
index 1c12df43e..c62a32bee 100644
--- a/apps/desktop/src/renderer/components/chat/AgentChatPane.tsx
+++ b/apps/desktop/src/renderer/components/chat/AgentChatPane.tsx
@@ -61,6 +61,7 @@ import type {
 } from "../../../shared/types/orchestration";
 import { parseAgentChatTranscript } from "../../../shared/chatTranscript";
 import { isProviderSlashCommandInput } from "../../../shared/chatSlashCommands";
+import { deriveDeterministicLaneNameFromPrompt } from "../../../shared/laneNameFallback";
 import {
   LOCAL_PROVIDER_LABELS,
   MODEL_REGISTRY,
@@ -650,23 +651,33 @@ function draftLaunchRequestKey(args: {
   });
 }
 
-function createTemporaryAutoLaneName(date = new Date()): string {
+function autoLaneGenericSuffix(date = new Date()): string {
   const pad = (value: number) => String(value).padStart(2, "0");
   return [
-    "chat",
-    `${date.getFullYear()}${pad(date.getMonth() + 1)}${pad(date.getDate())}`,
-    `${pad(date.getHours())}${pad(date.getMinutes())}${pad(date.getSeconds())}`,
-  ].join("-");
+    date.getFullYear(),
+    pad(date.getMonth() + 1),
+    pad(date.getDate()),
+  ].join("") + "-" + [
+    pad(date.getHours()),
+    pad(date.getMinutes()),
+    pad(date.getSeconds()),
+  ].join("");
 }
 
-const AUTO_LANE_NAME_SUGGEST_TIMEOUT_MS = 3_500;
+function createDeterministicAutoLaneName(prompt: string, options: { genericSuffix?: string | null } = {}): string {
+  return deriveDeterministicLaneNameFromPrompt(prompt, options);
+}
+
+const AUTO_LANE_NAME_SUGGEST_TIMEOUT_MS = 10_000;
 
 async function suggestAutoLaneName(args: {
   laneId: string;
   prompt: string;
   modelId: string;
+  genericSuffix?: string | null;
+  onFallback?: (message: string) => void;
 }): Promise<string> {
-  const fallbackName = createTemporaryAutoLaneName();
+  const fallbackName = createDeterministicAutoLaneName(args.prompt, { genericSuffix: args.genericSuffix });
   let timeoutId: ReturnType<typeof setTimeout> | null = null;
   try {
     const suggested = await Promise.race([
@@ -677,12 +688,16 @@ async function suggestAutoLaneName(args: {
         fallbackName,
       }),
       new Promise<string>((resolve) => {
-        timeoutId = setTimeout(() => resolve(fallbackName), AUTO_LANE_NAME_SUGGEST_TIMEOUT_MS);
+        timeoutId = setTimeout(() => {
+          args.onFallback?.(`Lane naming with ${formatLocalModelLabel(args.modelId)} took longer than 10s; using a deterministic name.`);
+          resolve(fallbackName);
+        }, AUTO_LANE_NAME_SUGGEST_TIMEOUT_MS);
       }),
     ]);
     return suggested.trim() || fallbackName;
   } catch (error) {
     console.warn("draft launch lane name suggestion failed; using fallback name", error);
+    args.onFallback?.("Lane naming failed; using a deterministic name.");
     return fallbackName;
   } finally {
     if (timeoutId) clearTimeout(timeoutId);
@@ -737,10 +752,11 @@ function draftLaunchPromptSnippet(job: DraftLaunchJob): string {
 
 function draftLaunchJobMessage(job: DraftLaunchJob): string {
   const laneSuffix = job.laneName ? ` in ${job.laneName}` : "";
-  if (job.status === "naming-lane") return `Creating lane for ${draftLaunchKindLabel(job.draftKind)}... Choosing a branch name.`;
-  if (job.status === "creating-lane") return `Creating lane for ${draftLaunchKindLabel(job.draftKind)}...`;
-  if (job.status === "starting-session") return `Starting ${draftLaunchKindLabel(job.draftKind)}${laneSuffix}...`;
-  if (job.status === "sending-prompt") return `Sending prompt to ${draftLaunchKindLabel(job.draftKind)}${laneSuffix}...`;
+  const warningSuffix = job.warning ? ` ${job.warning}` : "";
+  if (job.status === "naming-lane") return `Naming lane with ${formatLocalModelLabel(job.namingModelId ?? job.snapshot.modelId)}...${warningSuffix}`;
+  if (job.status === "creating-lane") return `Creating lane for ${draftLaunchKindLabel(job.draftKind)}...${warningSuffix}`;
+  if (job.status === "starting-session") return `Starting ${draftLaunchKindLabel(job.draftKind)}${laneSuffix}...${warningSuffix}`;
+  if (job.status === "sending-prompt") return `Sending prompt to ${draftLaunchKindLabel(job.draftKind)}${laneSuffix}...${warningSuffix}`;
   if (job.status === "failed") return job.error ? `Launch failed: ${job.error}` : "Launch failed.";
   return job.mode === "background"
     ? `Launched ${draftLaunchKindLabel(job.draftKind)}${laneSuffix}.`
@@ -6744,6 +6760,8 @@ export function AgentChatPane({
   const resolveDraftLaunchLane = useCallback(async (
     snapshot: DraftLaunchSnapshot,
     onAutoCreateNameResolved?: () => void,
+    onAutoCreateNameFallback?: (message: string) => void,
+    onAutoCreateNameModelResolved?: (modelId: string) => void,
   ): Promise<DraftLaunchLaneTarget> => {
     if (draftLaunchTargetIsAutoCreate) {
       if (!laneId) throw new Error("Select a lane before auto-creating a new lane.");
@@ -6751,13 +6769,23 @@ export function AgentChatPane({
         ?? availableLanes?.find((candidate) => candidate.name.trim().toLowerCase() === "primary")
         ?? null;
       if (!primaryLane) throw new Error("Auto-create requires a primary lane.");
-      const laneName = await suggestAutoLaneName({
-        laneId: primaryLane.id,
-        prompt: buildDraftLaunchNamingSeed(snapshot),
-        modelId: snapshot.modelId,
-      });
-      onAutoCreateNameResolved?.();
+      const namingSeed = buildDraftLaunchNamingSeed(snapshot);
       const projectConfigSnapshot = await getProjectConfigCached({ projectRoot, force: true }).catch(() => null);
+      const titleSettings = projectConfigSnapshot?.effective?.ai?.sessionIntelligence?.titles;
+      const titleModelId = typeof titleSettings?.modelId === "string" ? titleSettings.modelId.trim() : "";
+      const namingModelId = titleModelId || snapshot.modelId;
+      onAutoCreateNameModelResolved?.(namingModelId);
+      const genericSuffix = autoLaneGenericSuffix();
+      const laneName = titleSettings?.enabled === false
+        ? createDeterministicAutoLaneName(namingSeed, { genericSuffix })
+        : await suggestAutoLaneName({
+            laneId: primaryLane.id,
+            prompt: namingSeed,
+            modelId: namingModelId,
+            genericSuffix,
+            onFallback: onAutoCreateNameFallback,
+          });
+      onAutoCreateNameResolved?.();
       const baseSource = effectiveNewLaneBaseSource(projectConfigSnapshot);
       const branches = await fetchNewLaneBaseBranches({
         source: baseSource,
@@ -6990,7 +7018,9 @@ export function AgentChatPane({
       laneId: null,
       laneName: null,
       sessionId: null,
+      namingModelId: null,
       error: null,
+      warning: null,
       autoOpen: mode === "foreground",
       createdAtMs: Date.now(),
       snapshot,
@@ -7012,6 +7042,10 @@ export function AgentChatPane({
     try {
       targetLane = await resolveDraftLaunchLane(snapshot, () => {
         patchDraftLaunchJob(jobId, { status: "creating-lane" });
+      }, (message) => {
+        patchDraftLaunchJob(jobId, { warning: message });
+      }, (modelId) => {
+        patchDraftLaunchJob(jobId, { namingModelId: modelId });
       });
       patchDraftLaunchJob(jobId, {
         status: "starting-session",
@@ -7391,11 +7425,23 @@ export function AgentChatPane({
             issueCount ? `${issueCount} issue${issueCount === 1 ? "" : "s"}` : null,
           ].filter(Boolean).join(" · ");
         }
-        const baseName = await suggestAutoLaneName({
-          laneId,
-          prompt: namingSeed,
-          modelId: parallelModelSlots[0]!.modelId,
-        });
+        const projectConfigSnapshot = await getProjectConfigCached({ projectRoot, force: false }).catch(() => null);
+        const titleSettings = projectConfigSnapshot?.effective?.ai?.sessionIntelligence?.titles;
+        const titleModelId = typeof titleSettings?.modelId === "string" ? titleSettings.modelId.trim() : "";
+        const namingModelId = titleModelId || parallelModelSlots[0]!.modelId;
+        if (titleSettings?.enabled !== false) {
+          setParallelLaunchStatus(`Naming lanes with ${formatLocalModelLabel(namingModelId)}...`);
+        }
+        const genericSuffix = autoLaneGenericSuffix();
+        const baseName = titleSettings?.enabled === false
+          ? createDeterministicAutoLaneName(namingSeed, { genericSuffix })
+          : await suggestAutoLaneName({
+              laneId,
+              prompt: namingSeed,
+              modelId: namingModelId,
+              genericSuffix,
+              onFallback: setParallelLaunchStatus,
+            });
         setParallelLaunchStatus(`Creating ${parallelModelSlots.length} child lanes…`);
 
         for (const slot of parallelModelSlots) {
@@ -9511,6 +9557,15 @@ export function AgentChatPane({
                   Restore
                 </button>
               ) : null}
+              {isActiveJob && job.status === "naming-lane" ? (
+                <button
+                  type="button"
+                  className="rounded-md px-2 py-0.5 text-[length:calc(var(--chat-font-size)*10.5/14)] font-medium text-fg/65 transition-colors hover:bg-white/10 hover:text-fg/85"
+                  onClick={() => navigate("/settings?tab=background-jobs")}
+                >
+                  Settings
+                </button>
+              ) : null}
               {canOpen ? (
                 <button
                   type="button"
diff --git a/apps/desktop/src/renderer/components/lanes/laneAgents.ts b/apps/desktop/src/renderer/components/lanes/laneAgents.ts
index 03ffce8f2..66c1c905a 100644
--- a/apps/desktop/src/renderer/components/lanes/laneAgents.ts
+++ b/apps/desktop/src/renderer/components/lanes/laneAgents.ts
@@ -72,7 +72,7 @@ function chatAgentFrom(summary: AgentChatSessionSummary): LaneAgent {
     activity: chatActivity(summary),
     lastHint: summary.awaitingInput
       ? "Awaiting your input"
-      : summary.lastOutputPreview?.trim() || null,
+      : summary.summary?.trim() || summary.lastOutputPreview?.trim() || null,
     lastActivityAt: summary.lastActivityAt ?? summary.startedAt,
   };
 }
@@ -88,7 +88,7 @@ function cliAgentFrom(summary: TerminalSessionSummary): LaneAgent {
     activity: cliActivity(summary),
     lastHint: summary.runtimeState === "waiting-input"
       ? "Awaiting your input"
-      : summary.lastOutputPreview?.trim() || null,
+      : summary.summary?.trim() || summary.lastOutputPreview?.trim() || null,
     lastActivityAt: summary.endedAt ?? summary.startedAt,
   };
 }
diff --git a/apps/desktop/src/renderer/components/onboarding/AiRuntimesBand.tsx b/apps/desktop/src/renderer/components/onboarding/AiRuntimesBand.tsx
index e9fc5bcb0..fbf0fbda6 100644
--- a/apps/desktop/src/renderer/components/onboarding/AiRuntimesBand.tsx
+++ b/apps/desktop/src/renderer/components/onboarding/AiRuntimesBand.tsx
@@ -111,19 +111,19 @@ export function AiRuntimesBand() {
     try {
       if (key === "auto_title") {
         await window.ade.ai.updateConfig({
-          sessionIntelligence: { titles: { modelId: modelId || undefined } } as AiConfig["sessionIntelligence"],
+          sessionIntelligence: { titles: { modelId: modelId || null } } as AiConfig["sessionIntelligence"],
         });
       } else {
-        const overrides: Record<string, string> = {};
+        const overrides: Record<string, string | null> = {};
         for (const [k, v] of Object.entries(aiConfig?.featureModelOverrides ?? {})) {
           if (typeof v === "string" && v) overrides[k] = v;
         }
         if (modelId) overrides[key] = modelId;
-        else delete overrides[key];
+        else overrides[key] = null;
         await window.ade.ai.updateConfig({
           featureModelOverrides: overrides as AiConfig["featureModelOverrides"],
           ...(key === "terminal_summaries"
-            ? { sessionIntelligence: { summaries: { modelId: modelId || undefined } } as AiConfig["sessionIntelligence"] }
+            ? { sessionIntelligence: { summaries: { modelId: modelId || null } } as AiConfig["sessionIntelligence"] }
             : {}),
         });
       }
diff --git a/apps/desktop/src/renderer/components/settings/AiFeaturesSection.tsx b/apps/desktop/src/renderer/components/settings/AiFeaturesSection.tsx
index f667bbf3f..375a03312 100644
--- a/apps/desktop/src/renderer/components/settings/AiFeaturesSection.tsx
+++ b/apps/desktop/src/renderer/components/settings/AiFeaturesSection.tsx
@@ -34,7 +34,7 @@ type ChatTitleSettingsPatch = {
 };
 
 const FEATURES: FeatureInfo[] = [
-  { key: "terminal_summaries", label: "Chat & terminal summaries", description: "Summarize closed terminal sessions and keep chat session summaries updated", subtitle: "Never lose track of what happened in closed sessions", icon: ChatCircleDots },
+  { key: "terminal_summaries", label: "Summarize completed chats and terminals", description: "Replace raw last output with a concise session summary when work completes", subtitle: "Show what happened instead of the last terminal line", icon: ChatCircleDots },
   { key: "pr_descriptions", label: "PR description drafting", description: "Draft PR descriptions when you trigger the action in the PR flows", subtitle: "Get a head start on PR descriptions when you're ready to merge", icon: GitPullRequest },
   { key: "commit_messages", label: "Commit messages", description: "Generate a brief git commit subject when the field is empty", subtitle: "Meaningful commit messages generated from your staged changes", icon: GitCommit },
 ];
@@ -161,13 +161,20 @@ export function AiFeaturesSection() {
       );
       setChatAutoTitleEnabled(effectiveAi?.sessionIntelligence?.titles?.enabled ?? true);
       setChatAutoTitleRefresh(effectiveAi?.sessionIntelligence?.titles?.refreshOnComplete ?? true);
-      setChatAutoTitleReasoning(effectiveAi?.chat?.autoTitleReasoningEffort ?? null);
+      setChatAutoTitleReasoning(
+        effectiveAi?.sessionIntelligence?.titles?.reasoningEffort
+        ?? effectiveAi?.chat?.autoTitleReasoningEffort
+        ?? null
+      );
 
       const persistedReasoning = effectiveAi?.featureReasoningOverrides ?? {};
       const nextReasoning: Record<string, string | null> = {};
       for (const key of Object.keys(persistedReasoning)) {
         nextReasoning[key] = persistedReasoning[key as AiFeatureKey] ?? null;
       }
+      nextReasoning.terminal_summaries = effectiveAi?.sessionIntelligence?.summaries?.reasoningEffort
+        ?? nextReasoning.terminal_summaries
+        ?? null;
       setFeatureReasoning(nextReasoning);
     } finally {
       setLoading(false);
@@ -205,13 +212,11 @@ export function AiFeaturesSection() {
         sessionIntelligence: {
           titles: {
             enabled: nextEnabled,
-            modelId: nextModelId || undefined,
+            modelId: nextModelId || null,
+            reasoningEffort: nextReasoning,
             refreshOnComplete: nextRefresh,
           },
         },
-        chat: {
-          autoTitleReasoningEffort: nextReasoning,
-        },
       });
 
       setChatAutoTitleEnabled(nextEnabled);
@@ -261,13 +266,17 @@ export function AiFeaturesSection() {
     try {
       const nextFeatureModels = { ...featureModels, [key]: modelId };
       setFeatureModels(nextFeatureModels);
+      const featureModelOverrides = toFeatureModelOverrides(nextFeatureModels) ?? {};
+      if (!modelId) {
+        featureModelOverrides[key] = null;
+      }
       await window.ade.ai.updateConfig({
-        featureModelOverrides: toFeatureModelOverrides(nextFeatureModels),
+        featureModelOverrides,
         ...(key === "terminal_summaries"
           ? {
               sessionIntelligence: {
                 summaries: {
-                  modelId: modelId || undefined,
+                  modelId: modelId || null,
                 },
               } as AiConfig["sessionIntelligence"],
             }
@@ -288,8 +297,20 @@ export function AiFeaturesSection() {
       for (const [k, v] of Object.entries(nextReasoning)) {
         if (v != null) overrides[k] = v;
       }
+      if (effort == null) {
+        overrides[key] = null;
+      }
       await window.ade.ai.updateConfig({
         featureReasoningOverrides: overrides as AiConfig["featureReasoningOverrides"],
+        ...(key === "terminal_summaries"
+          ? {
+              sessionIntelligence: {
+                summaries: {
+                  reasoningEffort: effort,
+                },
+              } as AiConfig["sessionIntelligence"],
+            }
+          : {}),
       });
     } finally {
       setSaving(false);
@@ -447,7 +468,7 @@ export function AiFeaturesSection() {
             );
           })}
 
-          {/* Auto-name chat tabs */}
+          {/* Auto-name chats, CLI sessions, and lanes */}
           <div
             className="ai-feature-row"
             style={{
@@ -486,7 +507,7 @@ export function AiFeaturesSection() {
                       color: chatAutoTitleEnabled ? COLORS.textPrimary : COLORS.textMuted,
                     }}
                   >
-                    Auto-name chat tabs
+                    Auto-name chats, CLI sessions, and lanes
                   </div>
                   <div
                     style={{
@@ -499,7 +520,7 @@ export function AiFeaturesSection() {
                       wordBreak: "break-word",
                     }}
                   >
-                    Tabs automatically get descriptive names based on conversation content
+                    Names come from the first prompt, with a deterministic fallback when the model is off or too slow
                   </div>
                   <label style={{ display: "flex", alignItems: "center", gap: 6, marginTop: 6, cursor: "pointer" }}>
                     <input
diff --git a/apps/desktop/src/renderer/components/terminals/SessionCard.tsx b/apps/desktop/src/renderer/components/terminals/SessionCard.tsx
index 1ed66cf09..8e9dc2a38 100644
--- a/apps/desktop/src/renderer/components/terminals/SessionCard.tsx
+++ b/apps/desktop/src/renderer/components/terminals/SessionCard.tsx
@@ -111,9 +111,13 @@ function orchestrationRoleA11yLabel(role: OrchestrationRole, tag?: string | null
 function getPreviewLine(session: TerminalSessionSummary, primaryText: string): string | null {
   const summary = preferredSessionLabel(session.summary);
   if (summary && summary !== primaryText) return summary;
+  const goal = preferredSessionLabel(session.goal);
+  if (session.status !== "running") {
+    if (goal && goal !== primaryText) return goal;
+    return null;
+  }
   const preview = sanitizeTerminalInlineText(session.lastOutputPreview, 120);
   if (preview && preview !== primaryText) return preview;
-  const goal = preferredSessionLabel(session.goal);
   if (goal && goal !== primaryText) return goal;
   return null;
 }
diff --git a/apps/desktop/src/renderer/lib/draftLaunchJobs.ts b/apps/desktop/src/renderer/lib/draftLaunchJobs.ts
index e5ba7af72..ed36cd1d3 100644
--- a/apps/desktop/src/renderer/lib/draftLaunchJobs.ts
+++ b/apps/desktop/src/renderer/lib/draftLaunchJobs.ts
@@ -79,7 +79,9 @@ export type DraftLaunchJob = {
   laneId: string | null;
   laneName: string | null;
   sessionId: string | null;
+  namingModelId: string | null;
   error: string | null;
+  warning: string | null;
   autoOpen: boolean;
   createdAtMs: number;
   snapshot: DraftLaunchSnapshot;
diff --git a/apps/desktop/src/shared/laneNameFallback.test.ts b/apps/desktop/src/shared/laneNameFallback.test.ts
new file mode 100644
index 000000000..6a9b2f6d6
--- /dev/null
+++ b/apps/desktop/src/shared/laneNameFallback.test.ts
@@ -0,0 +1,27 @@
+import { describe, expect, it } from "vitest";
+
+import {
+  deriveDeterministicLaneNameFromPrompt,
+  genericLaneFallbackName,
+  genericSuffixFromLaneFallbackName,
+} from "./laneNameFallback";
+
+describe("lane name fallback", () => {
+  it("derives compact task slugs from prompts", () => {
+    expect(deriveDeterministicLaneNameFromPrompt("Can you please fix the login bug?")).toBe("fix-login-bug");
+  });
+
+  it("uses the generic suffix only when the prompt has no meaningful slug", () => {
+    expect(deriveDeterministicLaneNameFromPrompt("!!!", { genericSuffix: "20260610-142233" })).toBe(
+      "parallel-task-20260610-142233",
+    );
+    expect(deriveDeterministicLaneNameFromPrompt("Fix the login bug", { genericSuffix: "20260610-142233" })).toBe(
+      "fix-login-bug",
+    );
+  });
+
+  it("extracts old chat timestamp fallback suffixes", () => {
+    expect(genericSuffixFromLaneFallbackName("chat-20260514-010203")).toBe("20260514-010203");
+    expect(genericLaneFallbackName("20260514-010203")).toBe("parallel-task-20260514-010203");
+  });
+});
diff --git a/apps/desktop/src/shared/laneNameFallback.ts b/apps/desktop/src/shared/laneNameFallback.ts
new file mode 100644
index 000000000..128bb8d77
--- /dev/null
+++ b/apps/desktop/src/shared/laneNameFallback.ts
@@ -0,0 +1,90 @@
+export const GENERIC_LANE_FALLBACK_NAME = "parallel-task";
+
+export const LANE_FALLBACK_STOPWORDS = new Set([
+  "a",
+  "an",
+  "and",
+  "are",
+  "as",
+  "can",
+  "could",
+  "for",
+  "from",
+  "have",
+  "help",
+  "how",
+  "i",
+  "in",
+  "into",
+  "is",
+  "it",
+  "just",
+  "let",
+  "make",
+  "me",
+  "my",
+  "of",
+  "on",
+  "please",
+  "pls",
+  "the",
+  "this",
+  "to",
+  "use",
+  "we",
+  "with",
+  "you",
+]);
+
+function normalizeGenericSuffix(raw: string | null | undefined): string | null {
+  const normalized = String(raw ?? "")
+    .toLowerCase()
+    .replace(/[^a-z0-9-]+/g, "-")
+    .replace(/-+/g, "-")
+    .replace(/^-|-$/g, "")
+    .slice(0, 24)
+    .replace(/^-|-$/g, "");
+  return normalized.length ? normalized : null;
+}
+
+export function genericLaneFallbackName(genericSuffix?: string | null): string {
+  const suffix = normalizeGenericSuffix(genericSuffix);
+  return suffix ? `${GENERIC_LANE_FALLBACK_NAME}-${suffix}` : GENERIC_LANE_FALLBACK_NAME;
+}
+
+export function deriveDeterministicLaneNameFromPrompt(
+  prompt: string,
+  options: { genericSuffix?: string | null } = {},
+): string {
+  const collapsed = prompt
+    .replace(/```[\s\S]*?```/g, " ")
+    .replace(/`([^`]+)`/g, "$1")
+    .replace(/\b(?:please|pls|can you|could you|help me|i need(?: you)? to|let'?s|we need to)\b/giu, " ")
+    .replace(/\s+/g, " ")
+    .trim();
+  if (!collapsed.length) return genericLaneFallbackName(options.genericSuffix);
+  const tokens = collapsed.toLowerCase().match(/[a-z0-9]+/g) ?? [];
+  const meaningfulWords = tokens
+    .filter((token) => token.length > 1 && !LANE_FALLBACK_STOPWORDS.has(token))
+    .slice(0, 5);
+  const fallbackWords = tokens
+    .filter((token) => token.length > 1)
+    .slice(0, 4);
+  const words = meaningfulWords.length ? meaningfulWords : fallbackWords;
+  const slug = words
+    .join("-")
+    .replace(/[^a-z0-9-]+/g, "-")
+    .replace(/-+/g, "-")
+    .replace(/^-|-$/g, "");
+  return slug.length ? slug.slice(0, 48) : genericLaneFallbackName(options.genericSuffix);
+}
+
+export function genericSuffixFromLaneFallbackName(fallbackName: string | null | undefined): string | null {
+  const normalized = normalizeGenericSuffix(fallbackName);
+  if (!normalized) return null;
+  if (normalized.startsWith(`${GENERIC_LANE_FALLBACK_NAME}-`)) {
+    return normalized.slice(GENERIC_LANE_FALLBACK_NAME.length + 1) || null;
+  }
+  const chatMatch = normalized.match(/^chat-(\d{8}-\d{6}(?:-[a-z0-9]+)?)/u);
+  return chatMatch?.[1] ?? null;
+}
diff --git a/apps/desktop/src/shared/types/config.ts b/apps/desktop/src/shared/types/config.ts
index c7873e741..2a2886d12 100644
--- a/apps/desktop/src/shared/types/config.ts
+++ b/apps/desktop/src/shared/types/config.ts
@@ -1301,13 +1301,15 @@ export type AiOrchestratorConfig = {
 export type SessionIntelligenceConfig = {
   titles?: {
     enabled?: boolean;
-    modelId?: ModelId;
+    modelId?: ModelId | null;
+    reasoningEffort?: string | null;
     /** Whether to regenerate the title when the session completes */
     refreshOnComplete?: boolean;
   };
   summaries?: {
     enabled?: boolean;
-    modelId?: ModelId;
+    modelId?: ModelId | null;
+    reasoningEffort?: string | null;
   };
 };
 
@@ -1345,7 +1347,7 @@ export type AiConfig = {
   localProviders?: AiLocalProviderConfigs;
   workerSafety?: WorkerSafetyPolicy;
   /** Per-feature model overrides, e.g. { pr_descriptions: "claude-sonnet-4-6" } */
-  featureModelOverrides?: Partial<Record<AiFeatureKey, string>>;
+  featureModelOverrides?: Partial<Record<AiFeatureKey, string | null>>;
   /** Per-feature reasoning effort overrides */
   featureReasoningOverrides?: Partial<Record<AiFeatureKey, string | null>>;
   /** Unified title + summary intelligence config for all session types */
diff --git a/docs/features/chat/README.md b/docs/features/chat/README.md
index ac12585f2..fcfd1809a 100644
--- a/docs/features/chat/README.md
+++ b/docs/features/chat/README.md
@@ -11,7 +11,7 @@ machinery layered on top.
 
 | Path | Role |
 |---|---|
-| `apps/desktop/src/main/services/chat/agentChatService.ts` | Main service: session lifecycle, turn dispatch, event emission, provider adapters, steer queue, handoff, auto-title, prompt-derived lane-name suggestions for auto-created / parallel lanes, event-history snapshots, slash-command discovery/merge (delegates to per-provider discovery modules and `slashCommandPromptExpansion` for unified prompt expansion), and active-workload detection used by project/window close guards. Lane naming runs through the session-intelligence prompt path, retries the requested/configured/default title models, then falls back to a prompt slug with an optional temporary suffix for uniqueness. Tracks Fast Mode with the legacy `codexFastMode: boolean` session field for every provider whose descriptor advertises `serviceTiers: ["fast"]`; Codex forwards it as `serviceTier: "fast" \| null` on every `thread/start` and `turn/start` JSON-RPC call, while Cursor SDK sessions resolve it through discovered model parameters (see [Agent Routing](agent-routing.md#provider-service-tiers-fast-mode)). Codex chat goals are managed through the app-server `thread/goal/get` / `set` / `clear` RPCs, persisted in session summaries, validated to the provider's 4,000-character objective limit, and normalized to ADE's unlimited-budget policy by sending `tokenBudget: null` and clearing provider-reported budgets. `applyCodexEffectiveThreadState` accepts a `requestedCodexPolicy` option and uses `shouldPreserveRequestedCodexPolicy` to keep ADE-controlled picker selections authoritative when the lifecycle response echoes an older thread policy (prevents a manual Plan→Edit switch from snapping back); it also syncs the abstract `permissionMode` via `syncLegacyPermissionMode` after every policy application. Builds ADE guidance from the active lane worktree so Agent Skill roots are lane-scoped in persistent system/developer prompts and provider fallback injection. Spawns Claude/Codex agent runtimes with `buildAgentRuntimeEnv(managed)` so every agent process inherits `ADE_CHAT_SESSION_ID`, `ADE_LANE_ID`, `ADE_PROJECT_ROOT`, and `ADE_WORKSPACE_ROOT` (used by the agent guidance to call `ade --socket app-control logs` / `terminal read --chat-session "$ADE_CHAT_SESSION_ID"` without resolving the chat ID itself). When the session has Linear issues attached (`session_linear_issues`), `buildAgentRuntimeEnv` also materializes them into a per-session context file via `writeSessionLinearIssueContextFile` (`<contextDir>/<sessionId>/linear-issues.json`, written atomically; stale files cleared when nothing is attached) and sets `ADE_LINEAR_ISSUE_IDS` (comma-joined identifiers) + `ADE_LINEAR_CONTEXT_FILE` so the agent reads its issue context without Linear credentials. Attaching a `linear_issue` context attachment at run time calls `laneService.attachLinearIssueToSession({ chatSessionId, issues, role: "worked", source: "chat_attach", includeInPr: true })` so the link is persisted even for standalone (laneless) chats; when the session has a lane it additionally runs `laneService.linkLinearIssues` for the lane/PR-card semantics. See [Linear integration](../linear-integration/README.md#session-scoped-issue-attachment-and-cli-context-injection). Claude SDK sessions also resolve the executable through `claudeCodeExecutable.ts` and pass `pathToClaudeCodeExecutable` so packaged builds can prefer the bundled native binary before PATH/auth fallbacks; interrupted Claude turns call `stopTask` for active subagents before emitting stopped subagent results. Claude resume paths run `claudeThinkingTranscriptRepair` before loading a transcript, and the runtime self-heals the same corruption after the Anthropic thinking-block 400 error. Full-auto plan acceptance emits the same plan-mode exit notice as the manual approval path so the renderer composer chip can update even when the session refresh races with compaction. Cursor SDK setup records interrupts that arrive while the worker is still being acquired, releases the acquired generation if setup loses the race, and suppresses false provider-health failures for user-initiated setup interrupts. Cursor provider slash commands use a dedicated discovery path (`cursorSlashCommandDiscovery`) instead of falling through to the generic filesystem-backed list. Large service file. |
+| `apps/desktop/src/main/services/chat/agentChatService.ts` | Main service: session lifecycle, turn dispatch, event emission, provider adapters, steer queue, handoff, auto-title, prompt-derived lane-name suggestions for auto-created / parallel lanes, event-history snapshots, slash-command discovery/merge (delegates to per-provider discovery modules and `slashCommandPromptExpansion` for unified prompt expansion), and active-workload detection used by project/window close guards. Lane naming runs through the session-intelligence prompt path, retries the requested/configured/default title models, then falls back to a deterministic prompt slug; branch uniqueness is handled by the lane id suffix added by lane creation. Tracks Fast Mode with the legacy `codexFastMode: boolean` session field for every provider whose descriptor advertises `serviceTiers: ["fast"]`; Codex forwards it as `serviceTier: "fast" \| null` on every `thread/start` and `turn/start` JSON-RPC call, while Cursor SDK sessions resolve it through discovered model parameters (see [Agent Routing](agent-routing.md#provider-service-tiers-fast-mode)). Codex chat goals are managed through the app-server `thread/goal/get` / `set` / `clear` RPCs, persisted in session summaries, validated to the provider's 4,000-character objective limit, and normalized to ADE's unlimited-budget policy by sending `tokenBudget: null` and clearing provider-reported budgets. `applyCodexEffectiveThreadState` accepts a `requestedCodexPolicy` option and uses `shouldPreserveRequestedCodexPolicy` to keep ADE-controlled picker selections authoritative when the lifecycle response echoes an older thread policy (prevents a manual Plan→Edit switch from snapping back); it also syncs the abstract `permissionMode` via `syncLegacyPermissionMode` after every policy application. Builds ADE guidance from the active lane worktree so Agent Skill roots are lane-scoped in persistent system/developer prompts and provider fallback injection. Spawns Claude/Codex agent runtimes with `buildAgentRuntimeEnv(managed)` so every agent process inherits `ADE_CHAT_SESSION_ID`, `ADE_LANE_ID`, `ADE_PROJECT_ROOT`, and `ADE_WORKSPACE_ROOT` (used by the agent guidance to call `ade --socket app-control logs` / `terminal read --chat-session "$ADE_CHAT_SESSION_ID"` without resolving the chat ID itself). When the session has Linear issues attached (`session_linear_issues`), `buildAgentRuntimeEnv` also materializes them into a per-session context file via `writeSessionLinearIssueContextFile` (`<contextDir>/<sessionId>/linear-issues.json`, written atomically; stale files cleared when nothing is attached) and sets `ADE_LINEAR_ISSUE_IDS` (comma-joined identifiers) + `ADE_LINEAR_CONTEXT_FILE` so the agent reads its issue context without Linear credentials. Attaching a `linear_issue` context attachment at run time calls `laneService.attachLinearIssueToSession({ chatSessionId, issues, role: "worked", source: "chat_attach", includeInPr: true })` so the link is persisted even for standalone (laneless) chats; when the session has a lane it additionally runs `laneService.linkLinearIssues` for the lane/PR-card semantics. See [Linear integration](../linear-integration/README.md#session-scoped-issue-attachment-and-cli-context-injection). Claude SDK sessions also resolve the executable through `claudeCodeExecutable.ts` and pass `pathToClaudeCodeExecutable` so packaged builds can prefer the bundled native binary before PATH/auth fallbacks; interrupted Claude turns call `stopTask` for active subagents before emitting stopped subagent results. Claude resume paths run `claudeThinkingTranscriptRepair` before loading a transcript, and the runtime self-heals the same corruption after the Anthropic thinking-block 400 error. Full-auto plan acceptance emits the same plan-mode exit notice as the manual approval path so the renderer composer chip can update even when the session refresh races with compaction. Cursor SDK setup records interrupts that arrive while the worker is still being acquired, releases the acquired generation if setup loses the race, and suppresses false provider-health failures for user-initiated setup interrupts. Cursor provider slash commands use a dedicated discovery path (`cursorSlashCommandDiscovery`) instead of falling through to the generic filesystem-backed list. Large service file. |
 | `apps/desktop/src/main/services/chat/runtimeEvents.ts` | Canonical cross-runtime event vocabulary (`turn.*`, `content.delta`, `tool.*`, `subagent.*`, teammate/task events, compaction boundaries) plus shims between legacy `AgentChatEvent` rows and the canonical runtime envelope. Claude emits canonical subagent events alongside the legacy rows while the other adapters migrate. |
 | `apps/ade-cli/src/tuiClient/` | Terminal **Work** chat TUI (Ink + React): same action/RPC contracts as desktop, **attached** (socket) or **embedded** (headless runtime via `ade-cli`). See [ADE Code](../ade-code/README.md). |
 | `apps/desktop/src/main/services/builtInBrowser/builtInBrowserService.ts` | Main-process broker for the in-app web browser. Owns persistent project-profile partitions derived from the active project root (fallback `persist:ade-browser`) and one window/project browser service per ADE `BrowserWindow`, so each project keeps isolated cookies/storage while its tabs share that project's authenticated browser profile. Each window service manages multiple `WebContentsView` tabs (cap 10), active tab, per-tab lane/chat owner and lease metadata, lightweight browser agent sessions, bounds, visibility, inspect state, targeted status events, screenshot capture, scratch browser-agent observations, diagnostics, per-tab action traces, and emission of `BuiltInBrowserContextItem`s for selected page elements. Observe/click/type/key/scroll/fill/clear/wait/screenshot/select/reload/back/forward/stop can target a hidden or non-active tab by `tabId` or `sessionId`; inspect mode remains a visible-tab interaction. Sessions bind an agent workflow to one tab, remember owner plus last observation/trace ids, and have `ade browser session <action> <id>` CLI aliases. Scratch observations live under `.ade/cache/browser-observations/`, include a bounded DOM element list plus console/network diagnostics by default, can render a numbered element-map screenshot with `includeElementMap`, and prune to the latest 3 observations per tab by default; click/fill/clear/press/wait can target viewport coordinates or resolve `selector`/`text`/`testId`/`elementIndex`/saved observation `handle` before dispatching CDP input. Waits wake from browser/network/page events with a timeout fallback, and `network-idle` requires complete ready state, no pending browser requests, and a configurable idle window. Handles preserve same-origin iframe/open-shadow-root context when available, and tab traces record action target metadata, duration, before/after URL, session id, observation id, and errors without storing typed fill/type text. `ade browser proof` promotes a fresh scratch observation to durable proof through the proof broker. Window-open requests from a page are handled via `setWindowOpenHandler` returning `action: "allow"` + a `createWindow` factory: a new internal tab is created and its `webContents` is returned to Chromium so the popup keeps its real `window.opener` relationship with the opener tab (important for OAuth flows that postMessage back to the parent). Navigation normalization/protocol policy lives in `builtInBrowserNavigation.ts`; Google sign-in permission policy lives in `builtInBrowserPermissions.ts`. Backs the `ade.builtInBrowser.*` IPC surface and is consumed by both `ChatBuiltInBrowserPanel` (sidebar Browser tab) and `openExternal.ts` (links inside the renderer route through the built-in browser when the protocol is `http`/`https`/`about:blank`). |
@@ -533,8 +533,8 @@ config service):
 
 - `ai.mode` -- `subscription` vs `guest`; gates auto-title, tool
   availability, and provider selection.
-- `ai.sessionIntelligence.titles.*` and
-  `ai.chat.autoTitleReasoningEffort` -- AI title generation.
+- `ai.sessionIntelligence.titles.*` -- AI title generation. Legacy
+  `ai.chat.autoTitleReasoningEffort` is migrated into this tree.
 - `ai.permissions.*` -- per-provider permission defaults
   (`claudePermissionMode`, Codex approval/sandbox defaults, OpenCode
   permission).
diff --git a/docs/features/chat/composer-and-ui.md b/docs/features/chat/composer-and-ui.md
index ece4007ad..2269fd6f4 100644
--- a/docs/features/chat/composer-and-ui.md
+++ b/docs/features/chat/composer-and-ui.md
@@ -215,12 +215,15 @@ and a footer that contains the composer.
 
 - **Work auto-create launch behavior.** The embedded draft composer can
   ask the main process for a lane name before creating a new Work lane.
-  The request includes a temporary `chat-YYYYMMDD-HHMMSS` fallback so
-  prompt-derived fallback names remain unique when model naming is
-  unavailable. Each launch creates a `DraftLaunchJob` that tracks
-  progress through `naming-lane` / `creating-lane` /
+  The request includes a deterministic prompt-derived fallback, so
+  model timeouts or errors still create readable lane and branch names.
+  Branch uniqueness is handled by the lane id suffix added by the lane
+  service. Each launch creates a `DraftLaunchJob` that tracks progress
+  through `naming-lane` / `creating-lane` /
   `starting-session` / `sending-prompt` / `ready` / `failed` states.
-  The lane-name suggestion phase is visible as `naming-lane`; once the
+  The lane-name suggestion phase is visible as `naming-lane` and names
+  the model being used. If naming takes longer than 10 seconds or fails,
+  the status row says ADE is using the deterministic fallback. Once the
   name is resolved the job advances to `creating-lane`. The composer is
   cleared optimistically when the job starts so the user can begin
   composing the next prompt immediately; the `DraftLaunchSnapshot`
@@ -558,8 +561,9 @@ These modules are pure and unit-testable:
   async launches and is stored in `appStore.draftLaunchJobsByScope`
   rather than local pane state. The composer is cleared immediately when
   the job starts, not when it finishes. Auto-created lanes begin at
-  `naming-lane`, switch to `creating-lane` after the suggested branch
-  name resolves, then move through session start and prompt send. If
+  `naming-lane`, show the naming model, switch to `creating-lane` after
+  the suggested branch name resolves or after the deterministic fallback
+  wins, then move through session start and prompt send. If
   the launch fails, the Restore action merges the snapshot back via
   `restoreDraftLaunchSnapshot`, which appends rather than replaces
   existing draft text and merges context items by id.
diff --git a/docs/features/onboarding-and-settings/README.md b/docs/features/onboarding-and-settings/README.md
index 73ce9dcb0..adc3c85c7 100644
--- a/docs/features/onboarding-and-settings/README.md
+++ b/docs/features/onboarding-and-settings/README.md
@@ -215,8 +215,9 @@ Renderer — settings:
   get the bundled CLI automatically; this surface is what makes
   `ade` available to the user's own terminals.
 - `apps/desktop/src/renderer/components/settings/AiFeaturesSection.tsx`
-  — Background Jobs settings for AI-powered helpers such as terminal
-  summaries, PR descriptions, commit messages, and chat titles.
+  — Background Jobs settings for AI-powered helpers: auto-naming chats,
+  CLI sessions, and lanes; summarizing completed chats and terminals;
+  PR description drafting; and commit message drafting.
 - `apps/desktop/src/renderer/components/settings/ProvidersSection.tsx`
   — AI Connections settings for provider CLIs, authentication, API keys,
   and model availability.
diff --git a/docs/features/onboarding-and-settings/configuration-schema.md b/docs/features/onboarding-and-settings/configuration-schema.md
index 42b5110a9..ff7ee5066 100644
--- a/docs/features/onboarding-and-settings/configuration-schema.md
+++ b/docs/features/onboarding-and-settings/configuration-schema.md
@@ -297,7 +297,7 @@ type AiConfig = {
   apiKeys?: Record<string, string>;       // stored encrypted per provider
   localProviders?: AiLocalProviderConfigs;
   workerSafety?: WorkerSafetyPolicy;
-  featureModelOverrides?: Partial<Record<AiFeatureKey, string>>;
+  featureModelOverrides?: Partial<Record<AiFeatureKey, string | null>>;
   featureReasoningOverrides?: Partial<Record<AiFeatureKey, string | null>>;
   sessionIntelligence?: SessionIntelligenceConfig;
 };
@@ -307,13 +307,16 @@ type AiConfig = {
 behavior. Legacy `providers.mode` migration is still in the service
 but idempotent.
 
-`sessionIntelligence` controls auto-titles and end-of-session
-summaries:
+`sessionIntelligence` controls background session naming and
+end-of-session summaries:
 
 - `titles.enabled`
 - `titles.refreshOnComplete`
-- `titles.modelId`
-- `summaries.enabled` and similar
+- `titles.modelId` (`null` clears a project override)
+- `titles.reasoningEffort`
+- `summaries.enabled`
+- `summaries.modelId` (`null` clears a project override)
+- `summaries.reasoningEffort`
 
 Legacy `ai.chat.autoTitleEnabled`, `ai.chat.autoTitleModelId`, and
 `ai.chat.autoTitleRefreshOnComplete` are read on load and migrated
diff --git a/docs/features/terminals-and-sessions/ui-surfaces.md b/docs/features/terminals-and-sessions/ui-surfaces.md
index 09dca30a7..c22a68d78 100644
--- a/docs/features/terminals-and-sessions/ui-surfaces.md
+++ b/docs/features/terminals-and-sessions/ui-surfaces.md
@@ -125,9 +125,10 @@ Three rows:
 1. **Status dot + title + relative time** — `sessionStatusDot()` and
    `primarySessionLabel()` drive these. The relative time comes from
    `relativeTimeCompact`.
-2. **Preview line** (conditional) — `session.summary`, then sanitized
-   `session.lastOutputPreview`, then `session.goal`, whichever differs
-   from the title. Sanitization strips ANSI and control chars via
+2. **Preview line** (conditional) — `session.summary` first. Running
+   sessions may then show sanitized `session.lastOutputPreview`; completed,
+   failed, disposed, and detached sessions fall back to `session.goal`
+   instead of raw last output. Sanitization strips ANSI and control chars via
    `sanitizeTerminalInlineText`.
 3. **Tool type + lane + badges** — `ToolLogo`, `shortToolTypeLabel`,
    lane icon/name, `ClaudeCacheTtlBadge` (Claude chat only), delta chips