fix: steer-delivery starvation — continuation coalescing (claimSteers)

The bug (seen live): a teacher that ends every turn with a new pending quiz
keeps the wake pipeline's idle gates permanently closed — the steer sits in
'delivering result' forever while the learner answers check after check.

- claimSteers (server/jobs.ts): the chat-handler claims every queued,
  undelivered steer the instant a client turn wins the turn registry, and
  appends it after the learner's answer — conversion-safe, the answered
  quiz carries its output. Every answer flushes pending steers immediately.
- The pump keeps its post-idle duplicate-guard (re-check steerDelivered
  before firing), so a claim racing the pump converges to exactly one
  delivery either way.
- The stale-job sweep now skips chats with a turn in flight (a healthy but
  slow wake no longer gets falsely re-queued as 'recovering').
- Out-of-lane surgical touch (flagged for master review): the 3-line
  handleChatRequest seam (claim after beginTurn, persist+stream 'outgoing').
- Verified live end-to-end on the stuck session: answering the pending quiz
  fired the steer + wake turn within a second; job flipped done/delivered.

src/server/chat-handler.ts

@@ -3,7 +3,7 @@ import { createTeacherAgent, type TeacherAgent } from "../agent/teacher-agent.ts
 import { createJobRunner } from "../agent/subagents/runner.ts";
 import { createExtractionPipeline } from "../agent/subagents/extraction.ts";
 import { createTeacherTools } from "../agent/tools.ts";
-import { deliverJobResult, setJobRespawn, setTurnRunner } from "./jobs.ts";
+import { claimSteers, deliverJobResult, setJobRespawn, setTurnRunner } from "./jobs.ts";
 import { stores } from "./stores.ts";
 import { beginTurn, endTurn } from "./turn-liveness.ts";
 import { buildTeacherContext, persistTurn, triggerExtraction } from "./turn-service.ts";
@@ -71,6 +71,16 @@ export async function handleChatRequest(req: Request, agent?: TeacherAgent): Pro
 		return Response.json({ error: "A turn is already running for this session" }, { status: 409 });
 	}
 
+	// CONTINUATION COALESCING (async protocol §4, the delivery-starvation fix):
+	// steers finished while a quiz chain held the wake gates ride THIS turn —
+	// appended after the learner's answer (conversion-safe: the answer carries
+	// its output). Claiming happens after beginTurn, so this turn owns the
+	// registry and the write-ahead persist below makes the claim durable.
+	const steers = claimSteers(chatId);
+	const outgoing = (
+		steers.length > 0 ? [...(body.messages as UIMessage[]), ...steers] : body.messages
+	) as UIMessage[];
+
 	// Write-ahead persistence: the request body IS the client's canonical
 	// state, so it goes on disk BEFORE the stream starts. The learner's answer
 	// is durable the instant the request arrives — any remount, second tab, or
@@ -78,7 +88,7 @@ export async function handleChatRequest(req: Request, agent?: TeacherAgent): Pro
 	// returns mid-turn reads turnInFlight from the registry and re-attaches by
 	// polling instead of resurrecting the stale transcript.)
 	try {
-		stores.chatStore.saveMessages(chatId, body.messages as UIMessage[]);
+		stores.chatStore.saveMessages(chatId, outgoing);
 	} catch (err) {
 		console.error(`[chat ${chatId}] write-ahead persist failed:`, err);
 	}
@@ -97,7 +107,7 @@ export async function handleChatRequest(req: Request, agent?: TeacherAgent): Pro
 				}),
 				buildTeacherContext(stores.graphStore, chatId),
 			),
-		uiMessages: body.messages as UIMessage[],
+		uiMessages: outgoing,
 		// Tool-call validation failures are messages OUR schemas authored (the
 		// quiz authoring contract, unknown tools) — safe and useful in the
 		// transcript: they are exactly what the teacher saw and retried on. The

src/server/jobs.ts

@@ -128,6 +128,45 @@ export function deliverJobResult(steer: SubagentSteerMessage): Promise<void> {
 	});
 }
 
+/**
+ * CONTINUATION COALESCING ("THE ASYNC PROTOCOL" §4): the chat-handler calls
+ * this the instant a client turn wins the turn registry, BEFORE its
+ * write-ahead persist. Undelivered steers for the chat ride THAT turn —
+ * appended after the learner's answer, which is conversion-safe (the
+ * answered quiz carries its output) — instead of waiting for the pump's
+ * idle window, which a quiz-chaining teacher may never open (the delivery
+ * starvation bug: every turn ends with a new pending quiz, so the gates
+ * never stay open long enough to fire a wake).
+ *
+ * Claimed steers resolve their deliver() promise and flip the job done
+ * here: the caller holds the turn registry, so the write-ahead persist is
+ * imminent and durable. Steers already in the transcript (a wake raced in)
+ * are finalized, not returned. The pump re-checks steerDelivered right
+ * before firing, so a claim racing the pump converges to exactly one
+ * delivery either way.
+ */
+export function claimSteers(chatId: string): SubagentSteerMessage[] {
+	const s = state();
+	const queue = s.queues.get(chatId);
+	if (!queue || queue.length === 0) return [];
+	const record = stores.chatStore.getChat(chatId);
+	if (!record) return []; // the handler validated the chat; defensive only
+	const claimed: SubagentSteerMessage[] = [];
+	for (const entry of queue) {
+		const jobId = jobIdOf(entry.steer);
+		if (steerDelivered(record.messages, jobId)) {
+			// A wake already landed it — just finish the bookkeeping.
+			finalizeJob(jobId);
+		} else {
+			claimed.push(entry.steer);
+			finalizeJob(jobId); // delivery is now the client turn's job
+		}
+		entry.resolve();
+	}
+	s.queues.delete(chatId);
+	return claimed;
+}
+
 function chatIdOf(steer: SubagentSteerMessage): string {
 	// The chat is resolvable from the job record; the steer's metadata
 	// carries jobId, which indexes the store.
@@ -194,6 +233,15 @@ async function pump(chatId: string): Promise<void> {
 				continue;
 			}
 
+			// A client continuation may have COALESCED this steer (claimSteers) while
+			// the gates were closed — re-check before firing: exactly one delivery.
+			if (steerDelivered(fresh.messages, jobId)) {
+				queue.shift();
+				finalizeJob(jobId);
+				entry.resolve();
+				continue;
+			}
+
 			// Fire the server-side teacher turn. A 409 = a client turn raced
 			// in between idle and firing — back off and retry the same steer.
 			const status = await s.turnRunner(chatId, [...fresh.messages, entry.steer]).catch((err: unknown) => {
@@ -288,6 +336,7 @@ function recoverStaleJobs(): void {
 				if (job.status !== "running") continue;
 				if (now - Date.parse(job.updatedAt) < STALE_JOB_MS) continue;
 				if (!stores.chatStore.getChat(job.chatId)) continue;
+				if (isTurnRunning(job.chatId)) continue; // a wake turn may be delivering it right now
 				if (job.result) {
 					// The wake never fired (or crashed mid-flight): re-deliver.
 					const steer = buildSubagentSteer(job, job.result as never);