fix: deep audit round 2 — non-streaming mode config + ClarificationCard + settings restructure

HIGH fixes:
- H-NS-1: Non-streaming agent_chat now builds ChatModeConfig and calls
  send_message_with_chat_mode(), matching the streaming path
- H-FE-1: ClarificationCard component renders structured clarification
  questions with type badge, question text, and numbered options
- H-SEM-1: SemanticSkillRouter annotated as @reserved (Phase 3 wiring)

MEDIUM fixes:
- M-SETTINGS-1: Settings menu restructured with "高级" section separator;
  skills/audit/tasks/heartbeat/semantic-memory grouped under advanced
- M-MAN-1: LoopEvent→StreamChatEvent mapping completeness checklist
  added as documentation comment in agent_chat_stream loop
- M-ORPH-1: Deleted orphaned Automation/ and SkillMarket/ files,
  plus transitively orphaned types, hooks, and adapters

desktop/src-tauri/src/kernel_commands/chat.rs

@@ -18,6 +18,18 @@ use crate::intelligence::validation::validate_string_length;
 pub struct ChatRequest {
     pub agent_id: String,
     pub message: String,
+    /// Enable extended thinking/reasoning
+    #[serde(default)]
+    pub thinking_enabled: Option<bool>,
+    /// Reasoning effort level (low/medium/high)
+    #[serde(default)]
+    pub reasoning_effort: Option<String>,
+    /// Enable plan mode
+    #[serde(default)]
+    pub plan_mode: Option<bool>,
+    /// Enable sub-agent delegation (Ultra mode only)
+    #[serde(default)]
+    pub subagent_enabled: Option<bool>,
 }
 
 /// Chat response
@@ -88,7 +100,23 @@ pub async fn agent_chat(
     let id: AgentId = request.agent_id.parse()
         .map_err(|_| "Invalid agent ID format".to_string())?;
 
-    let response = kernel.send_message(&id, request.message)
+    // Build chat mode config from request fields
+    let chat_mode = if request.thinking_enabled.is_some()
+        || request.reasoning_effort.is_some()
+        || request.plan_mode.is_some()
+        || request.subagent_enabled.is_some()
+    {
+        Some(zclaw_kernel::ChatModeConfig {
+            thinking_enabled: request.thinking_enabled,
+            reasoning_effort: request.reasoning_effort.clone(),
+            plan_mode: request.plan_mode,
+            subagent_enabled: request.subagent_enabled,
+        })
+    } else {
+        None
+    };
+
+    let response = kernel.send_message_with_chat_mode(&id, request.message, chat_mode)
         .await
         .map_err(|e| format!("Chat failed: {}", e))?;
 
@@ -258,6 +286,24 @@ pub async fn agent_chat_stream(
         let stream_timeout = tokio::time::Duration::from_secs(300);
 
         loop {
+            // === LoopEvent → StreamChatEvent mapping ===
+            //
+            // COMPLETENESS CHECKLIST: When adding a new LoopEvent variant, you MUST:
+            // 1. Add a match arm below
+            // 2. Add the corresponding StreamChatEvent variant (see struct defs above)
+            // 3. Add the TypeScript type in desktop/src/lib/kernel-types.ts
+            // 4. Add a handler in desktop/src/lib/kernel-chat.ts
+            //
+            // Current mapping (LoopEvent → StreamChatEvent):
+            //   Delta             → Delta
+            //   ThinkingDelta     → ThinkingDelta
+            //   ToolStart         → ToolStart / HandStart (if name starts with "hand_")
+            //   ToolEnd           → ToolEnd / HandEnd (if name starts with "hand_")
+            //   SubtaskStatus     → SubtaskStatus
+            //   IterationStart    → IterationStart
+            //   Complete          → Complete
+            //   Error             → Error
+            // ============================================
             // Check cancellation flag before each recv
             if cancel_clone.load(std::sync::atomic::Ordering::SeqCst) {
                 tracing::info!("[agent_chat_stream] Stream cancelled for session: {}", session_id);