Keep Assistant session transitions success-bound

docs/release-control/v6/internal/subsystems/ai-runtime.md

@@ -289,7 +289,11 @@ runtime cost control, and shared AI transport surfaces.
     safe visible briefing: the next chat turn must carry
     `autonomous_mode:false` even when the summary is context-only and has no
     queued action, while the visible badge/action copy must still reflect the
-    actual last-known action state instead of inventing a pending approval.
+    actual last-known action state instead of inventing a pending approval. That
+    restoration is success-bound: if the underlying session message load fails,
+    the drawer must leave the current context untouched instead of applying
+    summary-derived Patrol or approval state for a session the operator is not
+    actually viewing.
     Before `/api/ai/sessions` returns summaries with stored handoff action
     references, the chat runtime must refresh their safe approval/action status
     from the canonical approval store and action-audit store. Session listing is

docs/release-control/v6/internal/subsystems/frontend-primitives.md

@@ -189,7 +189,9 @@ work extends shared components instead of creating new local variants.
    context. The drawer must treat `handoff_summary.requires_approval` as a
    current pending-decision flag, not a historical action marker, so completed
    or rejected handoff actions render as action context rather than pending
-   approval.
+   approval. Session-load and new-conversation transitions must be
+   success-bound: if the underlying session operation fails, the shared drawer
+   store must not clear or replace the current scoped handoff context.
 9. `frontend-modern/src/utils/platformSupportManifest.generated.ts` shared with `unified-resources`: the generated platform support projection is both a canonical unified-resource platform union boundary and a shared frontend source/platform vocabulary boundary.
 10. `frontend-modern/src/utils/sourcePlatforms.ts` shared with `unified-resources`: the source platform normalizer is both a canonical unified-resource source adapter boundary and a shared frontend source/platform vocabulary boundary.
     That shared boundary must preserve `availability` as the agentless

frontend-modern/src/components/AI/Chat/__tests__/AIChat.test.tsx

@@ -23,7 +23,7 @@ const {
     sendMessage: vi.fn().mockResolvedValue(true),
     stop: vi.fn(),
     clearMessages: vi.fn(),
-    loadSession: vi.fn().mockResolvedValue(undefined),
+    loadSession: vi.fn().mockResolvedValue(true),
     newSession: vi.fn().mockResolvedValue({
       id: 'new-sess',
       title: '',
@@ -569,6 +569,27 @@ describe('AIChat', () => {
       });
     });
 
+    it('keeps scoped handoff context when starting a new session fails', async () => {
+      mockChat.newSession.mockResolvedValueOnce(null);
+      mockAiChatStore.context = {
+        findingId: 'finding-old',
+        autonomousMode: false,
+        briefing: {
+          sourceLabel: 'Pulse Patrol',
+          title: 'Old finding handoff',
+        },
+      };
+
+      renderChat();
+      fireEvent.click(screen.getByText('New'));
+
+      await waitFor(() => {
+        expect(mockChat.newSession).toHaveBeenCalledTimes(1);
+      });
+      expect(mockAiChatStore.clearContext).not.toHaveBeenCalled();
+      expect(mockAiChatStore.context.findingId).toBe('finding-old');
+    });
+
     it('opens session picker on click', async () => {
       renderChat();
       fireEvent.click(screen.getByTitle('Pulse Assistant sessions'));
@@ -700,6 +721,50 @@ describe('AIChat', () => {
       expect(restoredContext.handoffActions).toBeUndefined();
     });
 
+    it('does not restore handoff context when loading a session fails', async () => {
+      mockChat.loadSession.mockResolvedValueOnce(false);
+      mockAIChatAPI.listSessions.mockResolvedValue([
+        {
+          id: 's-patrol',
+          title: 'High CPU follow-up',
+          created_at: '',
+          updated_at: '',
+          message_count: 4,
+          handoff_summary: {
+            kind: 'patrol_finding',
+            finding_id: 'finding-operator-briefing',
+            has_model_context: true,
+            resource_count: 1,
+            primary_resource: {
+              id: 'host:web-server',
+              name: 'web-server',
+              type: 'host',
+              node: 'pve-1',
+            },
+            action_count: 1,
+            requires_approval: true,
+          },
+        },
+      ]);
+
+      renderChat();
+      await waitFor(() => {
+        expect(mockAIChatAPI.listSessions).toHaveBeenCalled();
+      });
+      fireEvent.click(screen.getByTitle('Pulse Assistant sessions'));
+      await waitFor(() => {
+        expect(screen.getByText('High CPU follow-up')).toBeInTheDocument();
+      });
+
+      fireEvent.click(screen.getByText('High CPU follow-up'));
+
+      await waitFor(() => {
+        expect(mockChat.loadSession).toHaveBeenCalledWith('s-patrol');
+      });
+      expect(mockAiChatStore.setContext).not.toHaveBeenCalled();
+      expect(mockAiChatStore.clearContext).not.toHaveBeenCalled();
+    });
+
     it('shows completed Patrol actions as action context instead of pending approval', async () => {
       mockAIChatAPI.listSessions.mockResolvedValue([
         {

frontend-modern/src/components/AI/Chat/__tests__/useChat.test.ts

@@ -1010,8 +1010,9 @@ describe('useChat', () => {
       ]);
 
       const { value: chat, dispose } = withRoot(() => useChat());
-      await chat.loadSession('sess-42');
+      const loaded = await chat.loadSession('sess-42');
 
+      expect(loaded).toBe(true);
       expect(chat.sessionId()).toBe('sess-42');
       const msgs = chat.messages();
       expect(msgs).toHaveLength(2);
@@ -1026,8 +1027,9 @@ describe('useChat', () => {
       mockGetMessages.mockRejectedValue(new Error('not found'));
 
       const { value: chat, dispose } = withRoot(() => useChat());
-      await chat.loadSession('bad-id');
+      const loaded = await chat.loadSession('bad-id');
 
+      expect(loaded).toBe(false);
       expect(mockNotifyError).toHaveBeenCalledWith('Failed to load session');
       expect(chat.messages()).toEqual([]);
       dispose();

frontend-modern/src/components/AI/Chat/hooks/useChat.ts

@@ -607,7 +607,7 @@ export function useChat(options: UseChatOptions = {}) {
   };
 
   // Load session messages
-  const loadSession = async (id: string) => {
+  const loadSession = async (id: string): Promise<boolean> => {
     try {
       const msgs = await AIChatAPI.getMessages(id);
       setMessages(
@@ -620,9 +620,11 @@ export function useChat(options: UseChatOptions = {}) {
         })),
       );
       setSessionId(id);
+      return true;
     } catch (error) {
       logger.error('[useChat] Failed to load session:', error);
       notificationStore.error('Failed to load session');
+      return false;
     }
   };
 

frontend-modern/src/components/AI/Chat/index.tsx

@@ -913,7 +913,8 @@ export const AIChat: Component<AIChatProps> = (props) => {
 
   // New conversation
   const handleNewConversation = async () => {
-    await chat.newSession();
+    const session = await chat.newSession();
+    if (!session) return;
     aiChatStore.clearContext?.();
     setShowSessions(false);
   };
@@ -947,7 +948,8 @@ export const AIChat: Component<AIChatProps> = (props) => {
   // Load session
   const handleLoadSession = async (sessionId: string) => {
     const session = sessions().find((candidate) => candidate.id === sessionId);
-    await chat.loadSession(sessionId);
+    const loaded = await chat.loadSession(sessionId);
+    if (!loaded) return;
     const restoredContext = buildSessionHandoffContext(session);
     if (restoredContext) {
       aiChatStore.setContext(restoredContext);