fix(#234 review): wire Stop to server run + real autonomousRuns toggle (F1/F2/F3)

F1: the Stop button was wired only to useChat.stop() (a local SSE abort the server
ignores for a detached run), and dropping isStreaming re-armed observer-polling so the
still-running run streamed back into view. Add a client stopRun(chatId) -> POST
/ai-chat/stop; in autonomous mode Stop now also calls it (the authoritative stop). A
stoppingRun latch suppresses the observer MERGE (observedRow requires !stoppingRun) so
the stopped run's newly-persisted partial steps don't re-stream between Stop-press and
terminal settle. The latch clears only once this tab is no longer the streamer
(!localStreaming) — while streaming, the disabled run query holds the PREVIOUS turn's
terminal run in cache, which would otherwise clear the latch early and re-open the
flash on turn 2+. Latch releases on stopRun failure (view resumes) and on chat switch.
F2: give autonomousRuns a real enable path through the standard settings.ai.* toggle
(update-workspace.dto + workspace.service persist/audit + ai-provider-settings switch +
client workspace type), mirroring aiDictation. Persists to settings.ai.autonomousRuns
— the exact key the controller and ai-chat-window read.
F3: correct the pending->running comments (beginRun inserts 'running' directly;
'pending' is a reserved default never written in phase 1).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

apps/server/src/core/ai-chat/ai-chat-run.service.ts

@@ -69,8 +69,10 @@ export interface RunHandle {
  * server-side lifecycle object detached from the HTTP request / browser window.
  *
  * Responsibilities:
- *  - create a run row when a turn starts (pending -> running) and register an
- *    in-memory AbortController for it (the explicit-stop lever);
+ *  - create a run row when a turn starts (inserted directly as 'running'; the
+ *    'pending' status is only the column default + a reserved value, never
+ *    written by code in phase 1) and register an in-memory AbortController for it
+ *    (the explicit-stop lever);
  *  - finalize the run row (succeeded / failed / aborted) and unregister it;
  *  - service an EXPLICIT user stop (`requestStop`) — the ONLY thing that aborts a
  *    run; a browser disconnect deliberately does NOT;

apps/server/src/core/workspace/dto/update-workspace.dto.ts

@@ -55,6 +55,14 @@ export class UpdateWorkspaceDto extends PartialType(CreateWorkspaceDto) {
   @IsBoolean()
   aiDictationStreaming: boolean;
 
+  // #184: detached/autonomous agent runs (settings.ai.autonomousRuns). When on, a
+  // chat turn becomes a server-side RUN that survives a browser disconnect; only
+  // an explicit /ai-chat/stop ends it. Off by default; single-instance-only in
+  // phase 1 (see AiChatRunService.warnIfMultiInstance / AGENTS.md).
+  @IsOptional()
+  @IsBoolean()
+  autonomousRuns: boolean;
+
   // Workspace master toggle that enables/disables the HTML embed block type.
   // Persisted at settings.htmlEmbed. ABSENT/false => OFF (default). The block
   // itself renders in a sandboxed iframe, so this is a feature switch, not a

apps/server/src/core/workspace/services/workspace.service.ts

@@ -526,6 +526,20 @@ export class WorkspaceService {
         );
       }
 
+      if (typeof updateWorkspaceDto.autonomousRuns !== 'undefined') {
+        const prev = settingsBefore?.ai?.autonomousRuns ?? false;
+        if (prev !== updateWorkspaceDto.autonomousRuns) {
+          before.autonomousRuns = prev;
+          after.autonomousRuns = updateWorkspaceDto.autonomousRuns;
+        }
+        await this.workspaceRepo.updateAiSettings(
+          workspaceId,
+          'autonomousRuns',
+          updateWorkspaceDto.autonomousRuns,
+          trx,
+        );
+      }
+
       if (typeof updateWorkspaceDto.htmlEmbed !== 'undefined') {
         const prev = settingsBefore?.htmlEmbed ?? false;
         if (prev !== updateWorkspaceDto.htmlEmbed) {
@@ -579,6 +593,7 @@ export class WorkspaceService {
       delete updateWorkspaceDto.aiChat;
       delete updateWorkspaceDto.aiDictation;
       delete updateWorkspaceDto.aiDictationStreaming;
+      delete updateWorkspaceDto.autonomousRuns;
       delete updateWorkspaceDto.htmlEmbed;
       delete updateWorkspaceDto.trackerHead;
       delete updateWorkspaceDto.aiPublicShareAssistant;

apps/server/src/database/migrations/20260627T130000-ai-chat-runs.ts

@@ -7,8 +7,10 @@ import { type Kysely, sql } from 'kysely';
  * Until now an agent turn lived ONLY as long as the HTTP request was open
  * (`res.hijack()` in ai-chat.controller.ts); a browser disconnect aborted it.
  * This table makes a turn a persistent object the server owns: it is created
- * when a run starts, transitions pending -> running -> succeeded|failed|aborted,
- * and survives the subscriber (browser) going away. The DB is the source of
+ * when a run starts (inserted directly as 'running' in phase 1 — 'pending' is
+ * only this column's default + a reserved value, never written by code yet) and
+ * advances to succeeded|failed|aborted, surviving the subscriber (browser) going
+ * away when it settles. The DB is the source of
  * truth — a later client reconnects/sees the result by reading this row plus the
  * assistant message it projects (`assistant_message_id`).
  *