fix(runtime,kernel): HandTool 空壳修复 — 桥接到 HandRegistry 真实执行

B-HAND-1 修复: LLM 调用 hand_quiz/hand_researcher 等 Hand 工具后,
HandTool::execute() 原来返回假成功 JSON, 实际 Hand 并不执行.

修复方案 (沿用 SkillExecutor 模式):
- tool.rs: 新增 HandExecutor trait + ToolContext.hand_executor 字段
- hand_tool.rs: execute() 通过 context.hand_executor 分发到真实执行
- loop_runner.rs: AgentLoop 新增 hand_executor 字段 + builder + 3处 ToolContext 传递
- adapters.rs: 新增 KernelHandExecutor 桥接 HandRegistry.execute()
- kernel/mod.rs: 初始化 KernelHandExecutor + 注册到 AgentLoop
- messaging.rs: 两处 AgentLoop 构建添加 .with_hand_executor()

数据流: LLM tool call → HandTool::execute() → ToolContext.hand_executor
         → KernelHandExecutor → HandRegistry.execute() → Hand trait impl

809 tests passed, 0 failed.

crates/zclaw-runtime/src/tool/builtin/file_read.rs

@@ -139,6 +139,7 @@ mod tests {
             working_directory: None,
             session_id: None,
             skill_executor: None,
+            hand_executor: None,
             path_validator,
             event_sender: None,
         };

crates/zclaw-runtime/src/tool/builtin/file_write.rs

@@ -162,6 +162,7 @@ mod tests {
             working_directory: None,
             session_id: None,
             skill_executor: None,
+            hand_executor: None,
             path_validator,
             event_sender: None,
         }

crates/zclaw-runtime/src/tool/hand_tool.rs

@@ -78,21 +78,26 @@ impl Tool for HandTool {
         self.input_schema.clone()
     }
 
-    async fn execute(&self, input: Value, _context: &ToolContext) -> Result<Value> {
-        // Hand execution is delegated to HandRegistry via the kernel's
-        // hand execution path. This tool acts as the LLM-facing interface.
-        // The actual execution is handled by the HandRegistry when the
-        // kernel processes the tool call.
-
-        // For now, return a structured result that indicates the hand was invoked.
-        // The kernel's hand execution layer will handle the actual execution
-        // and emit HandStart/HandEnd events.
-        Ok(json!({
-            "hand_id": self.hand_id,
-            "status": "invoked",
-            "input": input,
-            "message": format!("Hand '{}' invoked successfully", self.hand_id)
-        }))
+    async fn execute(&self, input: Value, context: &ToolContext) -> Result<Value> {
+        // Delegate to the HandExecutor (bridged from HandRegistry via kernel).
+        // If no hand_executor is available (e.g., standalone runtime without kernel),
+        // return a descriptive error so the LLM knows the hand is unavailable.
+        match &context.hand_executor {
+            Some(executor) => {
+                executor.execute_hand(&self.hand_id, &context.agent_id, input).await
+            }
+            None => {
+                Ok(json!({
+                    "hand_id": self.hand_id,
+                    "status": "unavailable",
+                    "error": format!(
+                        "Hand '{}' cannot execute: no hand executor configured. \
+                         This usually means the kernel is not running or hands are not registered.",
+                        self.hand_id
+                    )
+                }))
+            }
+        }
     }
 }
 
@@ -130,13 +135,14 @@ mod tests {
     }
 
     #[tokio::test]
-    async fn test_hand_tool_execute() {
+    async fn test_hand_tool_execute_no_executor() {
         let tool = HandTool::from_config("quiz", "Generate quizzes", None);
         let ctx = ToolContext {
             agent_id: zclaw_types::AgentId::new(),
             working_directory: None,
             session_id: None,
             skill_executor: None,
+            hand_executor: None,
             path_validator: None,
             event_sender: None,
         };
@@ -144,6 +150,6 @@ mod tests {
         assert!(result.is_ok());
         let val = result.unwrap();
         assert_eq!(val["hand_id"], "quiz");
-        assert_eq!(val["status"], "invoked");
+        assert_eq!(val["status"], "unavailable");
     }
 }