diff --git a/agent/deep_assistant.py b/agent/deep_assistant.py
index ac6688b..205ee64 100644
--- a/agent/deep_assistant.py
+++ b/agent/deep_assistant.py
@@ -251,15 +251,15 @@ async def init_agent(config: AgentConfig):
     middleware = []
     middleware.append(EmptyResponseRetryMiddleware())
     middleware.append(ToolUseCleanupMiddleware())
-    tool_output_middleware = ToolOutputLengthMiddleware(
-        max_length=(getattr(config.generate_cfg, 'tool_output_max_length', None) if config.generate_cfg else None) or TOOL_OUTPUT_MAX_LENGTH,
-        truncation_strategy=getattr(config.generate_cfg, 'tool_output_truncation_strategy', 'smart') if config.generate_cfg else 'smart',
-        tool_filters=getattr(config.generate_cfg, 'tool_output_filters', None) if config.generate_cfg else None,
-        exclude_tools=getattr(config.generate_cfg, 'tool_output_exclude', []) if config.generate_cfg else [],
-        preserve_code_blocks=getattr(config.generate_cfg, 'preserve_code_blocks', True) if config.generate_cfg else True,
-        preserve_json=getattr(config.generate_cfg, 'preserve_json', True) if config.generate_cfg else True
-    )
-    middleware.append(tool_output_middleware)
+    # tool_output_middleware = ToolOutputLengthMiddleware(
+    #     max_length=(getattr(config.generate_cfg, 'tool_output_max_length', None) if config.generate_cfg else None) or TOOL_OUTPUT_MAX_LENGTH,
+    #     truncation_strategy=getattr(config.generate_cfg, 'tool_output_truncation_strategy', 'smart') if config.generate_cfg else 'smart',
+    #     tool_filters=getattr(config.generate_cfg, 'tool_output_filters', None) if config.generate_cfg else None,
+    #     exclude_tools=getattr(config.generate_cfg, 'tool_output_exclude', []) if config.generate_cfg else [],
+    #     preserve_code_blocks=getattr(config.generate_cfg, 'preserve_code_blocks', True) if config.generate_cfg else True,
+    #     preserve_json=getattr(config.generate_cfg, 'preserve_json', True) if config.generate_cfg else True
+    # )
+    # middleware.append(tool_output_middleware)
 
     if config.enable_memori:
         try:
diff --git a/routes/chat.py b/routes/chat.py
index c84b51c..e07f5fe 100644
--- a/routes/chat.py
+++ b/routes/chat.py
@@ -268,8 +268,8 @@ async def create_agent_and_generate_response(
     agent_responses = await agent.ainvoke({"messages": config.messages}, config=config.invoke_config(), max_tokens=MAX_OUTPUT_TOKENS)
 
     # ============ 执行 PostAgent hooks ============
-    # 注意：这里在非流式模式下同步执行hooks
-    await _execute_post_agent_hooks(config, "", sandbox)
+    # 注意：这里在单独的异步任务中执行，不阻塞非流式响应
+    asyncio.create_task(_execute_post_agent_hooks(config, "", sandbox))
     # ===========================================
 
     # 从后往前找第一个 HumanMessage，之后的内容都给 append_messages
diff --git a/utils/settings.py b/utils/settings.py
index 5594dba..5dd0376 100644
--- a/utils/settings.py
+++ b/utils/settings.py
@@ -101,11 +101,11 @@ SCHEDULE_MAX_CONCURRENT = int(os.getenv("SCHEDULE_MAX_CONCURRENT", "5"))
 # Daytona Sandbox 配置
 # ============================================================
 
-# DAYTONA_API_KEY = os.getenv("DAYTONA_API_KEY", "dtn_1c888ed16ec448b965e2e07afd75d69f8e0dd38efbad47744f9de49fcf7b7e2a")
-# DAYTONA_SERVER_URL = os.getenv("DAYTONA_SERVER_URL", "https://daytona.45.66.216.154.nip.io/api")
+DAYTONA_API_KEY = os.getenv("DAYTONA_API_KEY", "dtn_ccf86acd5b13a3069a8369d12d8cb26a9a184d07451374dfc325955fc2d8331c")
+DAYTONA_SERVER_URL = os.getenv("DAYTONA_SERVER_URL", "https://daytona.45.66.216.154.nip.io/api")
 
-DAYTONA_API_KEY = os.getenv("DAYTONA_API_KEY", "dtn_696a914ff54e45bb97132c32fba10995a4cab8ebef0cd8dea18129d447f805a3")
-DAYTONA_SERVER_URL = os.getenv("DAYTONA_SERVER_URL", "https://app.daytona.io/api")
+# DAYTONA_API_KEY = os.getenv("DAYTONA_API_KEY", "dtn_696a914ff54e45bb97132c32fba10995a4cab8ebef0cd8dea18129d447f805a3")
+# DAYTONA_SERVER_URL = os.getenv("DAYTONA_SERVER_URL", "https://app.daytona.io/api")
 
 DAYTONA_ENABLED = os.getenv("DAYTONA_ENABLED", "false") == "true"