loopx-agent · Nov 2, 2026 · Nov 2, 2026
@@ -2772,6 +2772,7 @@ const privateConversationSchema = z.object({
   ]).optional(),
   managed_steward: z.discriminatedUnion("available", [
     z.object({available: z.literal(true), prepared: z.boolean(), scheduling_enabled: z.boolean(),
+      execution_available: z.boolean().optional(), execution_error_code: z.string().optional(),
       goal_active: z.boolean().optional(), monitor: z.object({todo_id: z.string(), status: z.string(),
         cadence_minutes: z.number().int().nullable(), next_due_at: z.string().nullable(),
         last_checked_at: z.string().nullable()}).nullable().optional(),
@@ -2784,7 +2785,8 @@ const privateConversationSchema = z.object({
         z.object({available: z.literal(false), error_code: z.string()}),
       ]).optional(),
       goal_id: z.string().optional(), agent_id: z.string().optional(), session_id: z.string().optional()}),
-    z.object({available: z.literal(false), error_code: z.string()}),
+    z.object({available: z.literal(false), error_code: z.string(),
+      execution_available: z.boolean().optional(), execution_error_code: z.string().optional()}),
   ]).optional(),
   agent_candidates: z.array(privateAgentSessionSchema).default([]), agent_targets: z.array(privateAgentTargetSchema).default([]),
 });

@@ -193,7 +193,7 @@ export function PrivateConversationPanel({apps, onOpenOriginal}: {apps: LarkApp[
       {row.recovery_count > 0 ? <p role="status">{zh ? "存在尚未确认的发送回执。服务会读取原回执恢复；不要重新发送同一任务。检查 App 登录、权限和原会话后刷新状态。" : "A send receipt is unconfirmed. The service reads the original receipt to recover; avoid resending the same task. Check this App login, permissions and original Session, then refresh status."}</p> : null}
       {!row.context_available ? <p role="alert">{zh ? "工作区授权已失效；请恢复原工作区或重新选择。旧会话不会移到其它工作区。" : "The workspace grant is unavailable. Restore the original workspace or select a new one; the old Session will not move."}</p> : null}
       {row.context_kind === "project" && row.audience !== "group" ? <PrivateAgentTargets row={row} revision={revision} zh={zh} busy={busy} act={act}/> : null}
-      {row.context_kind === "steward" && row.audience === "owner" && row.executor_endpoint_id === "codex"
+      {row.context_kind === "steward" && row.audience === "owner" && row.managed_steward !== undefined
         ? <StewardManagedMode state={row.managed_steward} zh={zh} busy={busy} available={row.context_available}
             onPrepare={() => void act(() => prepareManagedSteward(row.binding_id))}
             onMonitor={(operation, cadence) => void act(() => configureStewardMonitor(row.binding_id, operation, cadence))}

@@ -15,31 +15,37 @@ export function StewardManagedMode({state, zh, busy, available, onPrepare, onMon
   const [cadence, setCadence] = useState(initial ?? 30);
   useEffect(() => setCadence(initial ?? 30), [initial, monitorId]);
   if (state === undefined) return null;
+  const executionAvailable = state.execution_available !== false;
+  const canConfigure = available && executionAvailable;
   return <details className="personal-cadence-details">
     <summary>{zh ? "原生周期监控" : "Native periodic monitoring"}</summary>
+    {!executionAvailable ? <p role="status">{zh
+      ? "当前执行器不支持此原生周期任务；原计划、窗口和回执均保留，执行与修改暂不可用。"
+      : "The current executor does not support this native scheduled work. Original schedules, windows and receipts are retained; execution and changes are unavailable."}</p> : null}
     {!state.available ? <p role="status">{zh ? "管家工作身份暂不可用；请检查原会话和连接后重试。" : "The steward identity is unavailable. Check the original conversation and connection, then retry."}</p>
       : state.prepared ? <>
-        <p role="status">{state.scheduling_enabled
+        <p role="status">{!executionAvailable ? (zh ? "原监控配置已保留。" : "Original monitoring configuration retained.") : state.scheduling_enabled
           ? (zh ? "授权 Goal 状态监控已启用。" : "Authorized Goal monitoring is enabled.")
           : state.monitor ? (zh ? "状态监控已暂停。" : "Goal monitoring is paused.")
           : (zh ? "管家身份已准备，状态监控尚未启用。" : "The steward identity is prepared; monitoring is not enabled.")}</p>
         <ContextHelp label={zh ? "状态检查范围" : "Status inspection scope"}><p>{zh ? "状态检查复用原生监控，不调用模型。" : "State checks use the native monitor without model calls."}</p></ContextHelp>
         {state.goal_active === false ? <p role="status">{zh ? "管家的工作 Goal 已停止；恢复原 Goal 后才能继续监控。" : "The steward Goal is stopped. Resume that original Goal before monitoring."}</p> : null}
-        {state.monitor ? <p>{state.scheduling_enabled ? (zh ? "下次检查：" : "Next check: ") : (zh ? "保留的到期时间：" : "Retained deadline: ")}{state.monitor.next_due_at ? new Date(state.monitor.next_due_at).toLocaleString(zh ? "zh-CN" : "en-US") : (zh ? "未排期" : "Unscheduled")}
+        {state.monitor ? <p>{executionAvailable && state.scheduling_enabled ? (zh ? "下次检查：" : "Next check: ") : (zh ? "保留的到期时间：" : "Retained deadline: ")}{state.monitor.next_due_at ? new Date(state.monitor.next_due_at).toLocaleString(zh ? "zh-CN" : "en-US") : (zh ? "未排期" : "Unscheduled")}
           <br/>{zh ? "上次检查：" : "Last check: "}{state.monitor.last_checked_at ? new Date(state.monitor.last_checked_at).toLocaleString() : (zh ? "尚未完成首次检查" : "No completed check yet")}</p> : null}
         <div className="personal-cadence-fields personal-cadence-target"><label>{zh ? "检查周期（分钟）" : "Check interval (minutes)"}
-          <input type="number" min={30} max={1440} step={1} value={cadence} disabled={busy || !available || state.goal_active === false}
+          <input type="number" min={30} max={1440} step={1} value={cadence} disabled={busy || !canConfigure || state.goal_active === false}
             aria-invalid={!Number.isInteger(cadence) || cadence < 30 || cadence > 1440}
             onChange={event => setCadence(Number(event.target.value))}/>
         </label></div>
-        <div className="personal-cadence-actions"><button className="is-primary" style={{minHeight: 44}} type="button" disabled={busy || !available || state.goal_active === false || !Number.isInteger(cadence) || cadence < 30 || cadence > 1440}
+        <div className="personal-cadence-actions"><button className="is-primary" style={{minHeight: 44}} type="button" disabled={busy || !canConfigure || state.goal_active === false || !Number.isInteger(cadence) || cadence < 30 || cadence > 1440}
           onClick={() => onMonitor("enable", cadence)}>{state.scheduling_enabled ? (zh ? "更新周期" : "Update interval") : (zh ? "启用状态监控" : "Enable monitoring")}</button>
-        {state.monitor?.status === "open" ? <button style={{minHeight: 44}} type="button" disabled={busy || !available} onClick={() => onMonitor("pause", initial ?? 30)}>{zh ? "暂停状态监控" : "Pause monitoring"}</button> : null}</div>
-        <StewardReport state={state.report} zh={zh} busy={busy} available={available} goalActive={state.goal_active !== false} onChange={onReport}/>
+        {state.monitor?.status === "open" ? <button style={{minHeight: 44}} type="button" disabled={busy || !canConfigure} onClick={() => onMonitor("pause", initial ?? 30)}>{zh ? "暂停状态监控" : "Pause monitoring"}</button> : null}</div>
+        <StewardReport state={state.report} zh={zh} busy={busy} available={canConfigure}
+          executionAvailable={executionAvailable} goalActive={state.goal_active !== false} onChange={onReport}/>
       </>
       : <>
         <p>{zh ? "为当前管家准备一个可见的 LoopX 工作身份，保留原对话和权限。此步骤不会启动定时任务或模型。" : "Prepare one visible LoopX identity for this steward, retaining its original conversation and permissions. This step starts neither scheduled tasks nor a model."}</p>
-        <button type="button" disabled={busy || !available} onClick={onPrepare}>
+        <button type="button" disabled={busy || !canConfigure} onClick={onPrepare}>
           {zh ? "准备管家身份" : "Prepare steward identity"}
         </button>
       </>}

@@ -5,8 +5,9 @@ import "./steward-report.css";
 
 type Managed = Extract<NonNullable<PrivateConversation["managed_steward"]>, {available: true}>;
 
-export function StewardReport({state, zh, busy, available, goalActive, onChange}: {
+export function StewardReport({state, zh, busy, available, executionAvailable = true, goalActive, onChange}: {
   state: Managed["report"]; zh: boolean; busy: boolean; available: boolean; goalActive: boolean;
+  executionAvailable?: boolean;
   onChange: (operation: "enable" | "pause" | "cancel", schedule?: Record<string, unknown>) => void;
 }) {
   const original = state?.available ? state.schedule : null;
@@ -21,17 +22,18 @@ export function StewardReport({state, zh, busy, available, goalActive, onChange}
   const signature = JSON.stringify(original);
   useEffect(() => {if (original) {setSchedule(original); setZone(String(original.timezone));}}, [signature]);
   if (!state) return null;
-  const disabled = busy || !available;
+  const disabled = busy || !available || !executionAvailable;
   let validZone = true;
   try {new Intl.DateTimeFormat("en", {timeZone: zone}).format();} catch {validZone = false;}
   const stamp = (value?: string | null) => value ? new Date(value).toLocaleString(zh ? "zh-CN" : "en-US") : (zh ? "暂无" : "None");
   return <section className="personal-steward-report">
     <h4>{zh ? "定期详细汇报" : "Detailed periodic reports"}</h4>
     {!state.available ? <p role="status">{zh ? "汇报状态暂不可读；原计划和回执已保留。刷新连接后重试。" : "Report status is unavailable. The original schedule and receipts are retained; refresh the connection."}</p> : <>
-      <p role="status">{state.enabled ? (zh ? "已启用 · 完整中文正文回此私聊" : "Enabled · Full Chinese report returns to this private Chat")
+      <p role="status">{!executionAvailable ? (zh ? "原汇报计划已保留，当前无法运行。" : "Original report configuration retained; execution unavailable.")
+        : state.enabled ? (zh ? "已启用 · 完整中文正文回此私聊" : "Enabled · Full Chinese report returns to this private Chat")
         : state.configured ? (zh ? "已暂停 · 原窗口和回执保留" : "Paused · Original windows and receipts retained")
         : (zh ? "尚未启用" : "Not enabled")}</p>
-      <p>{zh ? "汇总全部获准本机 Goal，包括停止项和读取缺口。每天汇报前24小时，生成与实际送达分别核验；状态监控照常运行。" : "Covers every authorized local Goal, including stopped Goals and read gaps. Daily reports cover the previous 24 hours; generation and delivery are verified separately. Monitoring continues."}</p>
+      <p>{zh ? "汇总全部获准本机 Goal，包括停止项和读取缺口。每天汇报前24小时，生成与实际送达分别核验；监控与汇报分别配置。" : "Covers every authorized local Goal, including stopped Goals and read gaps. Daily reports cover the previous 24 hours; generation and delivery are verified separately. Monitoring and reports are configured separately."}</p>
       {state.configured ? <p>{zh ? "下次日历计划：" : "Next calendar deadline: "}{stamp(state.next_due_at)}<br/>
         {zh ? "最近窗口：" : "Latest window: "}{stamp(state.retained_window_due_at)} · {state.delivery_status === "delivered"
           ? (zh ? "原私聊全文已核验送达" : "Full private reply verified")

@@ -6,6 +6,29 @@ Agent Session calendar reports are implemented. Installation,
 an enabled subscription and a natural scheduled delivery require their own
 readback evidence. Preparing an identity enables neither monitoring nor reports.
 
+Changing a previously prepared private steward to an executor without native
+scheduled-work support keeps its original configuration readable in connection
+settings. The monitor deadline, report schedule, retained window and delivery
+state remain visible with an explicit execution-unavailable notice. Configuration
+controls are disabled; inspecting retained state neither re-enables the old host
+nor retries a failed window. Native admission still owns every execution and
+configuration change. Restoring a supported executor restores the existing
+configuration view rather than creating a replacement identity or subscription.
+
+The disposable `steward_report_fixture.py --retained-executor` journey prepares
+one original synthetic steward and calendar window. With baseline assets the
+section is absent; current packaged assets expose the retained state at 1200
+and 390 pixels. Its `restore-supported-executor` control restores the same
+configuration. No wake owner is started, no model Turn is requested and no
+provider write is added during inspection or restoration. This qualifies UI
+readback and executor compatibility, not scheduled generation or live delivery.
+
+| Baseline: retained section hidden | Current: original schedule remains readable |
+| --- | --- |
+| ![Baseline packaged settings hide the retained schedule](../assets/personal-workspace/steward-reports/retained-executor-before.png) | ![Current packaged settings retain the schedule with unavailable execution](../assets/personal-workspace/steward-reports/retained-executor-desktop.png) |
+
+![The same retained schedule at 390 pixels](../assets/personal-workspace/steward-reports/retained-executor-narrow.png)
+
 Markdown, the HTML report and its copied full text use the same source-link
 rule. An unchanged, well-formed HTTP(S) address is a link. Opaque evidence
 digests, native Task/artifact references and private relative paths stay source

@@ -23,11 +23,14 @@ def main() -> None:
     parser.add_argument("--baseline-assets-dir", type=Path)
     parser.add_argument("--invalid-editorial", action="store_true", help="Return a known invalid synthetic report to inspect owner cancellation and retained failure.")
     parser.add_argument("--host-failure", action="store_true", help="Fail the synthetic native model before an editorial result; retain the original window without retry authority.")
+    parser.add_argument("--retained-executor", action="store_true", help="Keep an original configured window after selecting an executor without native scheduled-work support; readback only.")
     journey = parser.add_mutually_exclusive_group()
     journey.add_argument("--managed-session", action="store_true", help="Exercise ordinary Goal Session native work with the existing task validator, instead of a scheduled report.")
     journey.add_argument("--managed-report", action="store_true", help="Configure a calendar report for an existing local Agent Session through the packaged Goal settings.")
     journey.add_argument("--managed-calendar-task", choices=("file", "sqlite"), help="Repeat a completed, independently validated Task through the real calendar and packaged Goal settings.")
     args = parser.parse_args()
+    if args.retained_executor and (args.host_failure or args.invalid_editorial or args.managed_session or args.managed_report or args.managed_calendar_task):
+        parser.error("retained executor readback uses the original steward journey alone")
     if args.host_failure and (args.invalid_editorial or args.managed_session or args.managed_calendar_task):
         parser.error("select the pre-editorial report failure journey alone")
     if args.invalid_editorial and (args.managed_session or args.managed_calendar_task):
@@ -117,6 +120,23 @@ def fail_report(_adapter, _prompt, event_sink):
         if args.invalid_editorial:
             helper = root / "report_model.py"
             helper.write_text(helper.read_text().replace('"request_digest": request["request_digest"]', '"request_digest": "known-invalid"'))
+        if args.retained_executor:
+            from datetime import datetime, timezone
+            from loopx.capabilities.periodic_report.cadence_journal import admit_cadence_window
+            reports = StewardReports(runtime)
+            schedule = {"schema_version": "periodic_report_schedule_v0", "schedule_id": "retained-report",
+                "timezone": "UTC", "rrule": "FREQ=DAILY;BYHOUR=9;BYMINUTE=0", "start_date": "2001-01-01"}
+            plan = reports.configure(binding["binding_id"], operation="enable", schedule=schedule)["plan"]
+            reports.configure(binding["binding_id"], operation="enable", schedule=schedule,
+                execute=True, expected_plan_id=plan["plan_id"])
+            subscription = reports.subscription(binding["binding_id"])
+            admit_cadence_window(runtime_root=runtime.coordination_runtime_root, goal_id=actor["goal_id"],
+                agent_id=actor["agent_id"], subscription=subscription, now=datetime.now(timezone.utc),
+                subscription_resolver=lambda: subscription)
+            current = transport.bindings.read()
+            selected = next(row for row in current["bindings"] if row["binding_id"] == binding["binding_id"])
+            selected["executor_endpoint_id"] = "antigravity"
+            transport.bindings.path.write_text(json.dumps(current))
         before_calls, before_messages = capture.read_bytes(), store.messages(sid)
         server = ChatHTTPServer(("127.0.0.1", 0), ChatRequestHandler)
         server.chat_store, server.runtime_controller = store, runtime
@@ -153,9 +173,16 @@ def fail_report(_adapter, _prompt, event_sink):
         owner_start_count = 0
         restart_tokens = set()
         restart_refusals = set()
+        executor_restored = False
         try:
             while not stop.wait(.2):
                 server.assets_dir = current_assets if (output / "use-current-assets").exists() else baseline_assets
+                if args.retained_executor and not executor_restored and (output / "restore-supported-executor").exists():
+                    current = transport.bindings.read()
+                    selected = next(row for row in current["bindings"] if row["binding_id"] == binding["binding_id"])
+                    selected["executor_endpoint_id"] = "codex"
+                    transport.bindings.path.write_text(json.dumps(current))
+                    executor_restored = True
                 provider.owner_ready = not (output / "provider-unavailable").exists()
                 owners_stopped = (output / "stop-existing-wake").exists()
                 if owners_stopped and wake:
@@ -235,6 +262,7 @@ def fail_report(_adapter, _prompt, event_sink):
                     "provider_history_pages": sum("+chat-messages-list" in call for call in provider.calls),
                     "existing_wake_owner_running": wake is not None,
                     "existing_wake_owner_starts": owner_start_count,
+                    "retained_executor_restored": executor_restored,
                     "runtime_restarts": len(restart_tokens),
                     "runtime_restart_rejections": len(restart_refusals),
                     **({"managed_source_present": artifact.exists(),