fix(desktop): DeerFlow UI — ChatArea refactor + ai-elements + dead CSS cleanup

ChatArea retry button uses setInput instead of direct sendToGateway,
fix bootstrap spinner stuck for non-logged-in users,
remove dead CSS (aurora-title/sidebar-open/quick-action-chips),
add ai components (ReasoningBlock/StreamingText/ChatMode/ModelSelector/TaskProgress),
add ClassroomPlayer + ResizableChatLayout + artifact panel

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

desktop/src/store/chatStore.ts

@@ -103,10 +103,6 @@ interface ChatState {
   chatMode: ChatModeType;
   // Follow-up suggestions
   suggestions: string[];
-  // Artifacts (DeerFlow-inspired)
-  artifacts: import('../components/ai/ArtifactPanel').ArtifactFile[];
-  selectedArtifactId: string | null;
-  artifactPanelOpen: boolean;
 
   addMessage: (message: Message) => void;
   updateMessage: (id: string, updates: Partial<Message>) => void;
@@ -128,11 +124,6 @@ interface ChatState {
   setSuggestions: (suggestions: string[]) => void;
   addSubtask: (messageId: string, task: Subtask) => void;
   updateSubtask: (messageId: string, taskId: string, updates: Partial<Subtask>) => void;
-  // Artifact management (DeerFlow-inspired)
-  addArtifact: (artifact: import('../components/ai/ArtifactPanel').ArtifactFile) => void;
-  selectArtifact: (id: string | null) => void;
-  setArtifactPanelOpen: (open: boolean) => void;
-  clearArtifacts: () => void;
 }
 
 function generateConvId(): string {
@@ -271,10 +262,6 @@ export const useChatStore = create<ChatState>()(
   totalOutputTokens: 0,
   chatMode: 'thinking' as ChatModeType,
   suggestions: [],
-  artifacts: [],
-  selectedArtifactId: null,
-  artifactPanelOpen: false,
-
   addMessage: (message: Message) =>
     set((state) => ({ messages: [...state.messages, message] })),
 
@@ -401,6 +388,10 @@ export const useChatStore = create<ChatState>()(
   },
 
   sendMessage: async (content: string) => {
+    // Concurrency guard: prevent rapid double-click bypassing UI-level isStreaming check.
+    // React re-render is async — two clicks within the same frame both read isStreaming=false.
+    if (get().isStreaming) return;
+
     const { addMessage, currentAgent, sessionKey } = get();
     // Clear stale suggestions when user sends a new message
     set({ suggestions: [] });
@@ -436,27 +427,10 @@ export const useChatStore = create<ChatState>()(
 
     // Context compaction is handled by the kernel (AgentLoop with_compaction_threshold).
     // Frontend no longer performs duplicate compaction — see crates/zclaw-runtime/src/compaction.rs.
-
-    // Build memory-enhanced content using layered context (L0/L1/L2)
-    let enhancedContent = content;
-    try {
-      const contextResult = await intelligenceClient.memory.buildContext(
-        agentId,
-        content,
-        500, // token budget for memory context
-      );
-      if (contextResult.systemPromptAddition) {
-        const systemPrompt = await intelligenceClient.identity.buildPrompt(
-          agentId,
-          contextResult.systemPromptAddition,
-        );
-        if (systemPrompt) {
-          enhancedContent = `<context>\n${systemPrompt}\n</context>\n\n${content}`;
-        }
-      }
-    } catch (err) {
-      log.warn('Memory context build failed, proceeding without:', err);
-    }
+    // Memory context injection is handled by backend MemoryMiddleware (before_completion),
+    // which injects relevant memories into the system prompt. Frontend must NOT duplicate
+    // this by embedding old conversation memories into the user message content — that causes
+    // context leaking (old conversations appearing in new chat thinking/output).
 
     // Add user message (original content for display)
     // Mark as optimistic -- will be cleared when server confirms via onComplete
@@ -504,7 +478,7 @@ export const useChatStore = create<ChatState>()(
 
       // Try streaming first (ZCLAW WebSocket)
       const result = await client.chatStream(
-        enhancedContent,
+        content,
         {
           onDelta: (delta: string) => {
             // Update message content directly (works for both KernelClient and GatewayClient)
@@ -516,6 +490,15 @@ export const useChatStore = create<ChatState>()(
               ),
             }));
           },
+          onThinkingDelta: (delta: string) => {
+            set((s) => ({
+              messages: s.messages.map((m) =>
+                m.id === assistantId
+                  ? { ...m, thinkingContent: (m.thinkingContent || '') + delta }
+                  : m
+              ),
+            }));
+          },
           onTool: (tool: string, input: string, output: string) => {
             const step: ToolCallStep = {
               id: `step_${Date.now()}_${generateRandomString(4)}`,
@@ -732,20 +715,6 @@ export const useChatStore = create<ChatState>()(
       ),
     })),
 
-  // Artifact management (DeerFlow-inspired)
-  addArtifact: (artifact) =>
-    set((state) => ({
-      artifacts: [...state.artifacts, artifact],
-      selectedArtifactId: artifact.id,
-      artifactPanelOpen: true,
-    })),
-
-  selectArtifact: (id) => set({ selectedArtifactId: id }),
-
-  setArtifactPanelOpen: (open) => set({ artifactPanelOpen: open }),
-
-  clearArtifacts: () => set({ artifacts: [], selectedArtifactId: null, artifactPanelOpen: false }),
-
   initStreamListener: () => {
     const client = getClient();