Stale chat wakes expire instead of resuming days later

Live incident: the founder DMed the CEO and got silence — the reply run
sat 13th in a FIFO queue behind chat wakes from four days ago, which the
reopened app was faithfully executing (necro-replies that woke yet more
agents). A wake run is a conversation reply; past one hour the moment is
gone. expireStaleChatWakes (chat-agent-wake-service) fails queued
chat_mention runs older than the TTL and closes their vehicle tasks;
the worker tick runs it before claiming, with one log line. Real work
(heartbeat/kickoff triggers) never expires. Red-green tested.

tests/unit/chat-wake-expiry.test.ts

@@ -0,0 +1,99 @@
+import { randomUUID } from "node:crypto";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+import { describe, expect, test } from "vitest";
+import { DatabaseConnection } from "../../src/main/database/connection";
+import { OrganizationService } from "../../src/main/services/organization-service";
+import { ChatService } from "../../src/main/services/chat-service";
+import { CHAT_WAKE_TTL_MS, expireStaleChatWakes } from "../../src/main/services/chat-agent-wake-service";
+
+interface Harness {
+  connection: DatabaseConnection;
+  chat: ChatService;
+  orgId: string;
+  employeeId: string;
+  ceoId: string;
+  danaId: string;
+}
+
+function harness(): Harness {
+  const connection = new DatabaseConnection(join(tmpdir(), `ac-wake-expiry-${randomUUID()}.db`));
+  const organization = new OrganizationService(connection.db);
+  let orgId = "";
+  let ceoId = "";
+  let danaId = "";
+  connection.transaction(() => {
+    orgId = organization.createOrganization({ organizationName: "OrbitOps", employeeName: "Yang", mission: "m", goal: "g" });
+    const providerId = (connection.db.prepare("select id from providers where org_id = ? and provider_key = 'openai'").get(orgId) as { id: string }).id;
+    connection.db.prepare("update providers set enabled = 1 where id = ? and org_id = ?").run(providerId, orgId);
+    ceoId = organization.createAgent({
+      orgId, name: "CEO", role: "CEO", department: "Executive",
+      modelProviderId: providerId, modelId: "gpt-5.4", budgetCents: 25000, managerAgentId: null, roleLevel: "c_suite",
+    });
+    danaId = organization.createAgent({
+      orgId, name: "Dana", role: "Engineer", department: "Engineering",
+      modelProviderId: providerId, modelId: "gpt-5.4", budgetCents: 1000, managerAgentId: ceoId,
+    });
+  });
+  const employeeId = (connection.db.prepare("select id from employees where org_id = ?").get(orgId) as { id: string }).id;
+  return { connection, chat: new ChatService(connection.db), orgId, employeeId, ceoId, danaId };
+}
+
+/** DM an agent as the founder, returning the woken run id. */
+function wake(h: Harness, agentId: string): string {
+  const spaceId = h.chat.ensureDirectSpace(h.orgId, { type: "employee", id: h.employeeId }, { type: "agent", id: agentId });
+  h.chat.sendMessage({
+    orgId: h.orgId, spaceId, threadId: null, senderType: "employee", senderId: h.employeeId,
+    bodyMd: "ping", attachmentFileIds: [],
+  });
+  const run = h.connection.db.prepare(
+    "select id from task_runs where org_id = ? and agent_id = ? and trigger = 'chat_mention' and state = 'queued'",
+  ).get(h.orgId, agentId) as { id: string };
+  return run.id;
+}
+
+function runRow(h: Harness, runId: string): { state: string; error_message: string | null; task_status: string } {
+  return h.connection.db.prepare(`
+    select r.state, r.error_message, t.status as task_status
+    from task_runs r join tasks t on t.id = r.task_id
+    where r.id = ?
+  `).get(runId) as { state: string; error_message: string | null; task_status: string };
+}
+
+describe("stale chat wakes expire", () => {
+  test("a wake queued past the TTL is expired with its vehicle task; fresh wakes survive", () => {
+    const h = harness();
+    const staleRunId = wake(h, h.danaId);
+    const freshRunId = wake(h, h.ceoId);
+    const now = new Date();
+    const beyondTtl = new Date(now.getTime() - CHAT_WAKE_TTL_MS - 60_000).toISOString();
+    h.connection.db.prepare("update task_runs set created_at = ? where id = ?").run(beyondTtl, staleRunId);
+
+    const expired = h.connection.transaction(() => expireStaleChatWakes(h.connection.db, now.toISOString()));
+    expect(expired).toBe(1);
+
+    const stale = runRow(h, staleRunId);
+    expect(stale.state).toBe("failed");
+    expect(stale.error_message).toContain("Expired");
+    expect(stale.task_status).toBe("done");
+
+    const fresh = runRow(h, freshRunId);
+    expect(fresh.state).toBe("queued");
+    expect(fresh.task_status).toBe("backlog");
+
+    // Idempotent: nothing left to expire.
+    expect(h.connection.transaction(() => expireStaleChatWakes(h.connection.db, now.toISOString()))).toBe(0);
+    h.connection.close();
+  });
+
+  test("real work is never expired: heartbeat and kickoff runs outlive any queue age", () => {
+    const h = harness();
+    // The bootstrap-style run: a plain queued run with a non-chat trigger.
+    const staleRunId = wake(h, h.danaId);
+    h.connection.db.prepare("update task_runs set trigger = 'heartbeat', created_at = '2026-01-01T00:00:00Z' where id = ?").run(staleRunId);
+
+    expect(h.connection.transaction(() => expireStaleChatWakes(h.connection.db, new Date().toISOString()))).toBe(0);
+    expect((h.connection.db.prepare("select state from task_runs where id = ?").get(staleRunId) as { state: string }).state).toBe("queued");
+    h.connection.close();
+  });
+});