stopping a job that is already gone is a success, not an error

The chip could not be cleared by any means, and the toast could not be silenced
from the call site: useClient raises a GLOBAL onError before it throws, so
handling the failure where Stop is pressed was always going to be too late.

DELETE is idempotent. Asking for a job to stop when it is already gone has
achieved what was asked, so "no such task" and "already stopped" now answer 200
with `gone: true` instead of 400. No error, no toast, and the client clears the
chip on the flag.

A genuine refusal — a malformed id, a subagent with no fd to trace — is still
400.

The previous client-side check stays as a fallback, for a browser talking to a
server that has not been updated yet.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

src/servers/api/chat/chat.ts

@@ -357,7 +357,17 @@ chatRouter.delete('/tasks/:id', async (ctx) => {
   // the same thing, differing only in how hard it asks.
   const force = ctx.req.query('force') === '1';
   const result = await stopBackgroundTask(who, ctx.req.param('id'), force);
-  return ctx.json(result, result.stopped ? 200 : 400);
+  // "no such task" and "already stopped" are 200, not 400.
+  //
+  // DELETE is idempotent: asking for a job to stop when it is already gone has achieved what was asked.
+  // Reporting it as an error made the tray shout `{"stopped":false,"reason":"no such task"}` at somebody
+  // pressing Stop on a chip that could not be stopped, killed or dismissed — and because `useClient`
+  // raises a GLOBAL toast before the caller ever sees the failure, no amount of handling at the call
+  // site could quiet it.
+  //
+  // A genuine refusal — a malformed id, a subagent that has no fd to trace — is still 400.
+  const gone = result.reason === 'no such task' || result.reason === 'already stopped';
+  return ctx.json({ ...result, gone }, result.stopped || gone ? 200 : 400);
 });
 
 // POST /chat/tasks/:id/restart — run this job's command again, without involving the agent.

src/workspaces/officerdev/src/apps/Chat/components/BackgroundTaskTray.tsx

@@ -80,16 +80,17 @@ export const BackgroundTaskTray = ({ messages, sessionId }: BackgroundTaskTrayPr
     markStopRequested(taskId);
     setStopAttempted((prev) => ({ ...prev, [taskId]: Date.now() }));
     try {
-      await client.delete(`/chat/tasks/${taskId}${force ? '?force=1' : ''}`);
+      // `gone` means the server could not find it: no output file, so no pid, so nothing running. That is
+      // authoritative rather than a failure — the chip was the last thing claiming otherwise, so it goes.
+      const result = await client.delete<{ stopped: boolean; gone?: boolean }>(
+        `/chat/tasks/${taskId}${force ? '?force=1' : ''}`,
+      );
+      if (result?.gone) dismiss(taskId);
     } catch (err) {
+      // Kept as a fallback for a server that predates the idempotent DELETE: it answers 400 for a job it
+      // cannot find, and `useClient` has already raised a global toast by the time this runs — but the
+      // chip can at least still be cleared.
       const message = err instanceof Error ? err.message : '';
-      // The server saying it cannot FIND the job is authoritative, not an error to report: there is no
-      // output file, therefore no pid, therefore nothing running. The chip was the last thing claiming
-      // otherwise.
-      //
-      // Without this the tray showed a job that could not be stopped, killed or dismissed — a chip with
-      // no way out, because dismiss is only offered for a task the tray believes has finished. Pressing
-      // Stop then Kill both answered {"stopped":false,"reason":"no such task"} and changed nothing.
       if (message.includes('no such task') || message.includes('already stopped')) {
         dismiss(taskId);
         return;