fix(coding-agent): surfaced title generation failures

Logged structured session-title generation skip and failure outcomes with session/model context, and prevented credential lookup errors from escaping into the caller's swallowed promise path.

Added regression coverage for missing and failing title credentials.

Fixes #1892

packages/coding-agent/CHANGELOG.md

@@ -19,6 +19,8 @@
 
 ### Fixed
 
+- Fixed auto session-title generation failures being swallowed without an actionable diagnostic. Title generation now logs structured start, missing-model/API-key, provider-error, empty-result, and exception outcomes with the session id and resolved title model; the interactive auto-title caller also logs uncaught persistence/generation errors instead of dropping them. ([#1892](https://github.com/can1357/oh-my-pi/issues/1892))
+
 - Fixed a streamed assistant message freezing at a partial prefix (e.g. only "Nat" of "Natives built, now…") on ED3-risk terminals (Ghostty/kitty/iTerm2/Alacritty), with the final text appearing only after a resize. `TranscriptContainer` freezes each non-live block by replaying its last live render, but render coalescing can finalize a block's content and append the next block within the same throttled frame — so the block was sealed at its stale mid-stream snapshot and never repainted until the next `thaw`. The block that was live on the previous render is now recomputed once on the live→frozen transition, sealing it at its final content.
 
 - Fixed ACP/RPC stdio startup so protocol frames are no longer consumed as one-shot piped prompt input before the JSON-RPC transport starts.

packages/coding-agent/src/modes/controllers/input-controller.ts

@@ -1,6 +1,6 @@
 import * as fs from "node:fs/promises";
 import type { AutocompleteProvider, SlashCommand } from "@oh-my-pi/pi-tui";
-import { $env, sanitizeText } from "@oh-my-pi/pi-utils";
+import { $env, logger, sanitizeText } from "@oh-my-pi/pi-utils";
 import { getRoleInfo } from "../../config/model-registry";
 import { isSettingsInitialized, settings } from "../../config/settings";
 import { renderSegmentTrack } from "../../modes/components/segment-track";
@@ -406,7 +406,13 @@ export class InputController {
 							}
 						}
 					})
-					.catch(() => {});
+					.catch(err => {
+						logger.warn("title-generator: uncaught auto-title error", {
+							sessionId: this.ctx.session.sessionId,
+							reason: "uncaught-auto-title-error",
+							error: err instanceof Error ? err.message : String(err),
+						});
+					});
 			}
 
 			if (this.ctx.onInputCallback) {

packages/coding-agent/src/utils/title-generator.ts

@@ -149,7 +149,10 @@ export async function generateSessionTitle(
 	// tiny title model can't reliably decline trivial input, so this happens
 	// deterministically before any model is invoked; the caller retries on the
 	// next user message while the session stays unnamed.
-	if (isLowSignalTitleInput(firstMessage)) return null;
+	if (isLowSignalTitleInput(firstMessage)) {
+		logger.debug("title-generator: skipped low-signal input", { sessionId, reason: "low-signal" });
+		return null;
+	}
 
 	const tinyModel = settings.get("providers.tinyModel");
 	if (tinyModel === ONLINE_TINY_TITLE_MODEL_KEY) {
@@ -159,7 +162,14 @@ export async function generateSessionTitle(
 	const onlineAbortController = new AbortController();
 	const localTitle = tinyTitleClient.generate(tinyModel, firstMessage).then(
 		title => title || null,
-		() => null,
+		err => {
+			logger.warn("title-generator: local model error", {
+				sessionId,
+				model: tinyModel,
+				error: err instanceof Error ? err.message : String(err),
+			});
+			return null;
+		},
 	);
 	const startOnline = (): Promise<string | null> =>
 		generateTitleOnline(
@@ -188,49 +198,48 @@ export async function generateTitleOnline(
 ): Promise<string | null> {
 	const model = getTitleModel(registry, settings, currentModel);
 	if (!model) {
-		logger.debug("title-generator: no title model found");
+		logger.warn("title-generator: no title model found", { sessionId, reason: "no-title-model" });
 		return null;
 	}
 
 	const userMessage = formatTitleUserMessage(firstMessage);
-
-	const apiKey = await registry.getApiKey(model, sessionId);
-	if (!apiKey) {
-		logger.debug("title-generator: no API key for smol model", {
-			provider: model.provider,
-			id: model.id,
-		});
-		return null;
-	}
-	// Resolve metadata after getApiKey so the session-sticky credential for this
-	// request is already recorded; metadataResolver can then return the correct
-	// account_uuid rather than the snapshot-at-call-site value.
-	const metadata = metadataResolver?.(model.provider);
-
-	// Title generation is a 3-6 word task, but some reasoning backends ignore
-	// disableReasoning. Keep the normal cheap budget for non-reasoning models
-	// while reserving enough output room for reasoning models to still emit
-	// the forced tool call after any unavoidable thinking tokens.
-	const maxTokens = model.reasoning ? Math.max(TITLE_MAX_TOKENS, REASONING_SAFE_MAX_TOKENS) : TITLE_MAX_TOKENS;
-	const request = {
-		model: `${model.provider}/${model.id}`,
-		systemPrompt: TITLE_SYSTEM_PROMPT,
-		userMessage,
-		maxTokens,
+	const modelName = `${model.provider}/${model.id}`;
+	const modelContext = {
+		sessionId,
+		provider: model.provider,
+		id: model.id,
+		model: modelName,
 	};
-	logger.debug("title-generator: request", request);
+	logger.debug("title-generator: start", modelContext);
 
 	try {
+		const apiKey = await registry.getApiKey(model, sessionId);
+		if (!apiKey) {
+			logger.warn("title-generator: no API key", { ...modelContext, reason: "missing-api-key" });
+			return null;
+		}
+		// Resolve metadata after getApiKey so the session-sticky credential for this
+		// request is already recorded; metadataResolver can then return the correct
+		// account_uuid rather than the snapshot-at-call-site value.
+		const metadata = metadataResolver?.(model.provider);
+
+		// Title generation is a 3-6 word task, but some reasoning backends ignore
+		// disableReasoning. Keep the normal cheap budget for non-reasoning models
+		// while reserving enough output room for reasoning models to still emit
+		// the forced tool call after any unavoidable thinking tokens.
+		const maxTokens = model.reasoning ? Math.max(TITLE_MAX_TOKENS, REASONING_SAFE_MAX_TOKENS) : TITLE_MAX_TOKENS;
+		logger.debug("title-generator: request", { ...modelContext, maxTokens });
+
 		const response = await completeSimple(
 			model,
 			{
-				systemPrompt: [request.systemPrompt],
-				messages: [{ role: "user", content: request.userMessage, timestamp: Date.now() }],
+				systemPrompt: [TITLE_SYSTEM_PROMPT],
+				messages: [{ role: "user", content: userMessage, timestamp: Date.now() }],
 				tools: [setTitleTool],
 			},
 			{
 				apiKey,
-				maxTokens: request.maxTokens,
+				maxTokens,
 				disableReasoning: true,
 				toolChoice: { type: "tool", name: SET_TITLE_TOOL_NAME },
 				metadata,
@@ -239,8 +248,9 @@ export async function generateTitleOnline(
 		);
 
 		if (response.stopReason === "error") {
-			logger.debug("title-generator: response error", {
-				model: request.model,
+			logger.warn("title-generator: response error", {
+				...modelContext,
+				reason: "provider-response-error",
 				stopReason: response.stopReason,
 				errorMessage: response.errorMessage,
 			});
@@ -249,8 +259,18 @@ export async function generateTitleOnline(
 
 		const title = normalizeGeneratedTitle(extractGeneratedTitle(response.content));
 
-		logger.debug("title-generator: response", {
-			model: request.model,
+		if (!title) {
+			logger.debug("title-generator: no title returned", {
+				...modelContext,
+				reason: "model-returned-none",
+				usage: response.usage,
+				stopReason: response.stopReason,
+			});
+			return null;
+		}
+
+		logger.debug("title-generator: success", {
+			...modelContext,
 			title,
 			usage: response.usage,
 			stopReason: response.stopReason,
@@ -258,8 +278,9 @@ export async function generateTitleOnline(
 
 		return title;
 	} catch (err) {
-		logger.debug("title-generator: error", {
-			model: request.model,
+		logger.warn("title-generator: error", {
+			...modelContext,
+			reason: "exception",
 			error: err instanceof Error ? err.message : String(err),
 		});
 		return null;

packages/coding-agent/test/title-generator.test.ts

@@ -1,6 +1,7 @@
 import { afterEach, describe, expect, it, vi } from "bun:test";
 import * as ai from "@oh-my-pi/pi-ai";
 import { type Api, getBundledModel, type Model } from "@oh-my-pi/pi-ai";
+import { logger } from "@oh-my-pi/pi-utils";
 import { generateSessionTitle } from "../src/utils/title-generator";
 
 function getModelOrThrow(id: string): Model<Api> {
@@ -116,6 +117,65 @@ describe("title generator", () => {
 		expect(completeSimpleMock).toHaveBeenCalledTimes(1);
 	});
 
+	it("logs and returns null when title credentials are missing", async () => {
+		const model = getModelOrThrow("claude-sonnet-4-5");
+		const completeSimpleMock = vi.spyOn(ai, "completeSimple");
+		const warnSpy = vi.spyOn(logger, "warn").mockImplementation(() => {});
+
+		const title = await generateSessionTitle(
+			"Investigate the resolver",
+			{
+				getAvailable: () => [model],
+				getApiKey: async () => undefined,
+			} as never,
+			createSettings(model),
+			"session-1",
+		);
+
+		expect(title).toBeNull();
+		expect(completeSimpleMock).not.toHaveBeenCalled();
+		expect(warnSpy).toHaveBeenCalledWith(
+			"title-generator: no API key",
+			expect.objectContaining({
+				sessionId: "session-1",
+				provider: model.provider,
+				id: model.id,
+				reason: "missing-api-key",
+			}),
+		);
+	});
+
+	it("logs and returns null when title credential lookup throws", async () => {
+		const model = getModelOrThrow("claude-sonnet-4-5");
+		const completeSimpleMock = vi.spyOn(ai, "completeSimple");
+		const warnSpy = vi.spyOn(logger, "warn").mockImplementation(() => {});
+
+		const title = await generateSessionTitle(
+			"Investigate the resolver",
+			{
+				getAvailable: () => [model],
+				getApiKey: async () => {
+					throw new Error("credential lookup failed");
+				},
+			} as never,
+			createSettings(model),
+			"session-2",
+		);
+
+		expect(title).toBeNull();
+		expect(completeSimpleMock).not.toHaveBeenCalled();
+		expect(warnSpy).toHaveBeenCalledWith(
+			"title-generator: error",
+			expect.objectContaining({
+				sessionId: "session-2",
+				provider: model.provider,
+				id: model.id,
+				reason: "exception",
+				error: "credential lookup failed",
+			}),
+		);
+	});
+
 	it("uses a reasoning-safe output budget for reasoning models", async () => {
 		const model = getModelOrThrow("claude-sonnet-4-5");
 		const completeSimpleMock = vi.spyOn(ai, "completeSimple").mockResolvedValue({