diff --git a/backend/agent_factory.py b/backend/agent_factory.py
index 2734cf2..2d36b14 100644
--- a/backend/agent_factory.py
+++ b/backend/agent_factory.py
@@ -87,7 +87,7 @@ def create_agent(**kwargs) -> Agent:
     model_from_config = (config or {}).get("model_select")
     model_from_kwargs = kwargs.pop("model", None)
 
-    model_name = model_from_kwargs or model_from_config or "gemini-2.5-flash"
+    model_name = model_from_config or model_from_kwargs or "gemini-2.5-flash"
     model = OpenAIChatCompletionsModel(model_name, openai_client=client)
 
     # Safety: ensure no stray 'model' remains in kwargs to avoid duplicate kwarg errors
diff --git a/backend/service/mcp_client.py b/backend/service/mcp_client.py
index e6a35c7..9f55d64 100644
--- a/backend/service/mcp_client.py
+++ b/backend/service/mcp_client.py
@@ -2,7 +2,7 @@
 Author: ai-business-hql qingli.hql@alibaba-inc.com
 Date: 2025-06-16 16:50:17
 LastEditors: ai-business-hql ai.bussiness.hql@gmail.com
-LastEditTime: 2025-11-25 15:37:50
+LastEditTime: 2025-12-24 19:03:58
 FilePath: /comfyui_copilot/backend/service/mcp-client.py
 Description: 这是默认设置,请设置`customMade`, 打开koroFileHeader查看配置 进行设置: https://github.com/OBKoro1/koro1FileHeader/wiki/%E9%85%8D%E7%BD%AE
 '''
@@ -255,7 +255,7 @@ IF the user wants to:
 - Ask questions about the current workflow (e.g., "how does this work?", "explain the workflow").
 - Keywords: "analyze", "explain", "understand", "how it works", "workflow structure", "分析", "解释", "怎么工作的", "解读".
 
-**ACTION:**
+**ACTION:
 - You MUST call `get_current_workflow` to retrieve the workflow details.
 - Then, based on the returned workflow data, provide a detailed analysis or explanation to the user.
 
@@ -307,11 +307,11 @@ You must adhere to the following constraints to complete the task:
             agent_input = messages
             log.info(f"-- Processing {len(messages)} messages")
 
-            # from agents import Agent, Runner, set_trace_processors, set_tracing_disabled, set_default_openai_api
-            # from langsmith.wrappers import OpenAIAgentsTracingProcessor
-            # set_tracing_disabled(False)
-            # set_default_openai_api("chat_completions")
-            # set_trace_processors([OpenAIAgentsTracingProcessor()])
+            from agents import Agent, Runner, set_trace_processors, set_tracing_disabled, set_default_openai_api
+            from langsmith.wrappers import OpenAIAgentsTracingProcessor
+            set_tracing_disabled(False)
+            set_default_openai_api("chat_completions")
+            set_trace_processors([OpenAIAgentsTracingProcessor()])
 
             result = Runner.run_streamed(
                 agent,