Drop nothing running from dry fleet tallies

CHANGELOG.md

@@ -13,6 +13,17 @@ parallel copies under `docs/` or `scripts/notes/`. At cut time: rename
 
 ## [Unreleased]
 
+### Fixed
+
+- Dry-fleet transcript and `/status` report the outcome tally only
+  (`2 done, 1 failed`). They no longer claim `nothing running` when the
+  parent may still continue.
+- The prompt-box lockup stays on while a fleet is live or a dry-fleet
+  continuation is pending, even if the parent turn has settled. The word
+  cycles through a closed live-activity set (`working`, `warping`,
+  `buzzing`, `grinding`, `thinking`, `doing`, `cooking`, `creating`,
+  `imagining`, `inventing`) instead of going blank.
+
 ## [0.3.19] - 2026-09-10
 
 ### Security

docs/TUI.md

@@ -94,13 +94,16 @@ and `@mention` tokens anywhere paint `UI.action`. Bare skill or agent words
 (`implement`, `emil`, `brand review`) stay unstyled, as does a `/review`
 that appears mid-prose.
 
-While a turn is live the lockup slot swaps the wordmark for a semantic
-activity word — never the raw tool, MCP server, or plugin identifier that is
-actually executing. `resolveTurnLabel` (`src/tui/session-chrome.ts`)
-maps execution onto the closed set `ACTIVITY_STATES` exported from that
-module (`thinking`, `planning`, `researching`, `building`, `working`,
-`waiting`, `stalled`, `stopping`); that export is the source
-of truth for what the slot can say, not this list. It is led by a single density cell
+While a turn is live — or the session is still occupied by a live fleet
+or a pending dry-fleet continuation — the lockup slot swaps the wordmark
+for a semantic activity word — never the raw tool, MCP server, or plugin
+identifier that is actually executing. Live occupation cycles
+`LIVE_ACTIVITY_WORDS` (`working`, `warping`, `buzzing`, `grinding`,
+`thinking`, `doing`, `cooking`, `creating`, `imagining`, `inventing`)
+on `LIVE_WORD_MS`; gated turns still read `waiting` or `stopping`.
+`ACTIVITY_STATES` exported from the session chrome module is the source
+of truth for what the slot can say, not this list. It is led by a single
+density cell
 (`rampPulse`, `src/tui/ramp.ts`). The cell, not the word,
 is what says whether the session is healthy, and it carries four states:
 
@@ -120,16 +123,16 @@ printed identically, so the only way to tell them apart was to wait.
 waiting on something outside itself — and are told apart by motion: `blocked`
 holds perfectly still, which is the signal that the session is waiting on _you_.
 
-While fleet agents are running, the slot reports the _fleet_, not the parent.
-`resolveTurnLabel` and `resolveRampPhase` take a `FleetProgress` roll-up and
-rank it above the parent's own stall clock: with live lanes the parent is
-idle by design, so its silence says nothing about whether the session is
-progressing, and reporting it was how a session with every lane wedged still
-read as `working`. A fleet with no stalled lane reads `working`; one
-stalled lane makes the whole indicator read `stalled`, which is the state that
-should pull an operator's eye to the panel. A blocked gate and a stopping turn
-still outrank the fleet. With zero running fleet agents the roll-up is empty and
-every path through both functions behaves exactly as it does for a plain
+While fleet agents are running, the slot stays live even when the parent
+turn has settled (idle-with-fleet). `resolveTurnLabel` and
+`resolveRampPhase` take a `FleetProgress` roll-up plus session occupancy:
+with live lanes the parent is idle by design, so its silence says nothing
+about whether the session is progressing, and blanking the lockup made a
+busy fleet look hung. Occupied sessions keep cycling a live-activity
+word; recovery stays silent rather than painting `stalled`. A blocked
+gate and a stopping turn still outrank the fleet. With zero running fleet
+agents and no pending continuation the roll-up is empty and every path
+through both functions behaves exactly as it does for a plain
 single-agent turn.
 
 The stall phase is driven by the watchdog's own silence clock
@@ -255,8 +258,10 @@ Parent prose owns success narratives. Transcript fleet notices exist only for
 attention live spawn_agent rows cannot keep: a lane **failed** or **cancelled**
 while other work is still running, and **one** dry-fleet line when the last
 lane finishes
-(`N done · nothing running`; failed and cancelled counts appear only
-when non-zero, e.g. `N done, M failed, K cancelled · nothing running`).
+(`N done`; failed and cancelled counts appear only
+when non-zero, e.g. `N done, M failed, K cancelled`).
+The line does not claim the run is idle — the parent often continues.
+The prompt-box lockup is what names that occupation, not this tally.
 Per-lane `done — summary` walls and live `dispatched` re-announcements
 are never printed. That dry-fleet line stays operator-facing. If tasks
 are still todo/doing, the runtime re-enters the parent with collected

src/subagent/fleet-report.test.ts

@@ -89,7 +89,7 @@ describe("observeFleet", () => {
       ],
       T0 + 1000,
     );
-    expect(updates).toEqual(["2 done · nothing running"]);
+    expect(updates).toEqual(["2 done"]);
   });
 
   test("a failure names what went wrong while the fleet is still live", () => {
@@ -155,7 +155,7 @@ describe("observeFleet", () => {
         : { ...l, status: "failed" as const, error: "boom" },
     );
     const { updates } = observeFleet(seeded, after, T0 + 1000);
-    expect(updates).toEqual(["9 done, 3 failed · nothing running"]);
+    expect(updates).toEqual(["9 done, 3 failed"]);
   });
 
   test("a cancelled-only dry fleet counts cancelled, not failed", () => {
@@ -172,7 +172,7 @@ describe("observeFleet", () => {
       ],
       T0 + 1000,
     );
-    expect(updates).toEqual(["0 done, 2 cancelled · nothing running"]);
+    expect(updates).toEqual(["0 done, 2 cancelled"]);
   });
 
   test("a mixed dry fleet names done, failed, and cancelled separately", () => {
@@ -190,9 +190,7 @@ describe("observeFleet", () => {
       ],
       T0 + 1000,
     );
-    expect(updates).toEqual([
-      "1 done, 1 failed, 1 cancelled · nothing running",
-    ]);
+    expect(updates).toEqual(["1 done, 1 failed, 1 cancelled"]);
   });
 
   test("a burst of live cancels coalesces as cancelled, not failed", () => {
@@ -236,11 +234,11 @@ describe("fleetDigest", () => {
     expect(digest).toBe("2 running (api 1:20, docs 0:20) · 1 done · 1 failed");
   });
 
-  test("a fleet with nothing left running says so rather than going blank", () => {
+  test("a dry fleet is the outcome tally, not an idle claim", () => {
     expect(fleetDigest([lane({ id: "api", status: "done" })], T0)).toBe(
-      "nothing running · 1 done",
+      "1 done",
     );
-    expect(fleetDigest([], T0)).toBe("nothing running");
+    expect(fleetDigest([], T0)).toBe("");
   });
 
   test("cancelled lanes are named separately from failed", () => {
@@ -252,7 +250,7 @@ describe("fleetDigest", () => {
         ],
         T0,
       ),
-    ).toBe("nothing running · 1 failed · 1 cancelled");
+    ).toBe("1 failed · 1 cancelled");
   });
 });
 

src/subagent/fleet-report.ts

@@ -185,6 +185,10 @@ type Change =
   | { readonly kind: "cancelled"; readonly line: string }
   | { readonly kind: "stalled"; readonly line: string };
 
+export interface FleetReportOptions {
+  readonly stallMs?: number;
+}
+
 export interface FleetObservation {
   readonly watch: FleetWatch;
   /** Ready-to-print lines, already coalesced. Usually empty. */
@@ -195,8 +199,9 @@ export function observeFleet(
   previous: FleetWatch,
   lanes: readonly FleetLane[],
   nowMs: number,
-  stallMs: number = DEFAULT_STALL_MS,
+  options: FleetReportOptions = {},
 ): FleetObservation {
+  const stallMs = options.stallMs ?? DEFAULT_STALL_MS;
   const marks = new Map<string, LaneMark>();
   const changes: Change[] = [];
   let running = 0;
@@ -262,9 +267,7 @@ export function observeFleet(
   if (wentDry) {
     return {
       watch,
-      updates: [
-        clip(`${idleSummary(lanes)} · nothing running`, MAX_UPDATE_CHARS),
-      ],
+      updates: [clip(idleSummary(lanes), MAX_UPDATE_CHARS)],
     };
   }
 
@@ -355,14 +358,12 @@ function idleSummary(lanes: readonly FleetLane[]): string {
 export function fleetDigest(
   lanes: readonly FleetLane[],
   nowMs: number,
-  stallMs: number = DEFAULT_STALL_MS,
+  options: FleetReportOptions = {},
 ): string {
-  if (lanes.length === 0) return "nothing running";
+  const stallMs = options.stallMs ?? DEFAULT_STALL_MS;
   const running = lanes.filter((l) => l.status === "running");
   const parts: string[] = [];
-  if (running.length === 0) {
-    parts.push("nothing running");
-  } else {
+  if (running.length > 0) {
     const named = running
       .slice(0, DIGEST_NAMED_LANES)
       .map((lane) => {

src/subagent/index.ts

@@ -22,6 +22,7 @@ export {
   pendingAskWakeText,
   type FleetLane,
   type FleetObservation,
+  type FleetReportOptions,
   type FleetWatch,
   type PendingAskWake,
 } from "./fleet-report.js";

src/tui/commands/built-in.test.ts

@@ -79,6 +79,16 @@ describe("/status command", () => {
     });
   });
 
+  it("noops when the live fleet digest is empty so idle /status paints no blank row", () => {
+    const ctx: CommandContext = {
+      signalClear: () => undefined,
+      getFleetStatus: () => "",
+    };
+    expect(defined(getCommand("status"), "status").handler("", ctx)).toEqual({
+      type: "noop",
+    });
+  });
+
   it("says so rather than throwing when no fleet source is wired", () => {
     expect(
       defined(getCommand("status"), "status").handler("", makeCtx()),

src/tui/commands/built-in.ts

@@ -146,6 +146,9 @@ export function registerBuiltInCommands(): void {
           text: "Fleet status is not available in this session.",
         };
       }
+      if (status.length === 0) {
+        return { type: "noop" };
+      }
       return { type: "message", text: status };
     },
   });

src/tui/runtime-bridge.test.ts

@@ -14,6 +14,7 @@ import { streamRowCount } from "./shell/transcript";
 import { STEER_WAIT_NOTICE_MS } from "./notice-line";
 import { withTestRenderer } from "./harness";
 import { badgeCount } from "./session-queue";
+import { LIVE_ACTIVITY_WORDS } from "./session-chrome";
 
 describe("mapReactorLike", () => {
   test("message.received → user", () => {
@@ -1377,6 +1378,12 @@ describe("idle-with-fleet (CL-7057)", () => {
           // The parent turn settled but the fleet is live: the run stays
           // busy and the follow-up does not drain at mere parent-idle.
           expect(shell.session.run).toBe("busy");
+          expect(shell.lockupPhase).not.toBeNull();
+          expect(
+            (LIVE_ACTIVITY_WORDS as readonly string[]).includes(
+              shell.lockupPhase ?? "",
+            ),
+          ).toBe(true);
           expect(badgeCount(shell.session)).toBe(1);
           expect(port.calls).toEqual([]);
           await h.renderOnce();
@@ -1616,9 +1623,11 @@ describe("fleet-dry open-task drive (CL-7540)", () => {
           bridge.handle({ type: "fleet", running: 0 });
           expect(drives).toBe(0);
           expect(shell.session.run).toBe("busy");
+          expect(shell.lockupPhase).not.toBeNull();
           settleToollessTurn(bridge);
           expect(drives).toBe(1);
           expect(shell.session.run).toBe("busy");
+          expect(shell.lockupPhase).not.toBeNull();
           bridge.submit("when it finishes, summarize", "queue");
           expect(badgeCount(shell.session)).toBe(1);
           port.clear();

src/tui/runtime-bridge.ts

@@ -1011,6 +1011,11 @@ function drainLiveSteersAtBoundary(shell: AppShell, bag: BridgeBag): void {
   }
 }
 
+function occupancyHold(bag: BridgeBag, runBusy: boolean): boolean {
+  if (bag.liveFleet > 0 || bag.awaitingContinuationInference) return true;
+  return runBusy && bag.pendingDryOpenDrive;
+}
+
 /**
  * Release the run to idle and drain everything queued — but only at true
  * session-idle. A live fleet holds the run busy after the parent turn settles
@@ -1286,15 +1291,18 @@ export function attachSessionBridge(
       status: turn.status,
       currentToolName: turn.currentToolName,
       streamingType: turn.streamingType,
+      nowMs,
+      sessionActive: occupancyHold(bag, shell.session.run === "busy"),
     };
     const fleet = fleetProgress(bag.agentSessions, nowMs);
     const label = resolveTurnLabel(input, isStalled, fleet);
+    const sessionLive = label !== undefined;
     if (label === undefined) {
       // The bottom-left status slot rides the same re-entry as the landing
       // mark, so it crossfades between phases without a timer of its own.
       setLockupFrame(shell, {
         nowMs,
-        animating: turn.isProcessing,
+        animating: false,
         phase: null,
         rampPhase: null,
         stalledForMs: null,
@@ -1309,7 +1317,7 @@ export function attachSessionBridge(
     const stalledFor = stalledForMs(nowMs, rampPhase === "stalled");
     setLockupFrame(shell, {
       nowMs,
-      animating: turn.isProcessing,
+      animating: sessionLive,
       phase: label,
       rampPhase,
       stalledForMs: stalledFor,
@@ -1383,13 +1391,19 @@ export function attachSessionBridge(
       if (onTurnBoundary(event) && bag.turn.activeToolCalls.length > 0) {
         drainLiveSteersAtBoundary(shell, bag);
       }
-      if (settled) settleRun();
+      if (settled) {
+        settleRun();
+        paintPhase();
+      }
       return;
     }
     if (isBridgeInbound(event)) {
       applyInbound(shell, bag, event);
     }
-    if (settled) settleRun();
+    if (settled) {
+      settleRun();
+      paintPhase();
+    }
   };
 
   const recordLastSent = (