优化Guidelines指令遵循

prompt/guideline_prompt.md

@@ -18,7 +18,7 @@
 - **优先级排序**: 确定任务执行顺序
 
 ### 3. 计划阶段 (Planning)
-- **步骤制定**: 详细的执行步骤
+- **步骤制定**: 详细的执行步骤，步骤数量根据实际需求调整，每个步骤应具体可执行
 - **应急预案**: 准备应对可能的障碍
 
 ---
@@ -46,11 +46,7 @@
 - **可用资源**: [可以利用的工具和资源]
 
 ### ⚡ 执行步骤
-1. [第一步的具体行动]
-2. [第二步的具体行动]
-3. [第三步的具体行动]
-4. [第四步的具体行动]
-5. [第五步的具体行动]
+[按1,2,3...n顺序列出执行步骤]
 
 ## 输出语言 (Language)
 {language}

routes/chat.py

@@ -245,7 +245,7 @@ async def enhanced_generate_stream_response(
         if guideline_reasoning:
             # 用###分割guideline_reasoning，取最后一段作为Guidelines
             guidelines_text = guideline_reasoning.split('###')[-1].strip() if guideline_reasoning else ""
-            final_messages = append_user_last_message(final_messages, f"\n\nGuidelines:\n{guidelines_text}\nPlease follow these guidelines in your response.")
+            final_messages = append_user_last_message(final_messages, f"\n\nGuidelines:\n{guidelines_text}\nPlease follow these guidelines step by step.")
 
         # 第三阶段：agent响应流式传输
         logger.info(f"Starting agent stream response")
@@ -361,7 +361,7 @@ async def create_agent_and_generate_response(
     if guideline_reasoning:
         # 用###分割guideline_reasoning，取最后一段作为Guidelines
         guidelines_text = guideline_reasoning.split('###')[-1].strip() if guideline_reasoning else ""
-        final_messages = append_user_last_message(final_messages, f"\n\nGuidelines:\n{guidelines_text}\nPlease follow these guidelines in your response.")
+        final_messages = append_user_last_message(final_messages, f"\n\nGuidelines:\n{guidelines_text}\nPlease follow these guidelines step by step.")
         pre_message_list.append({"role": "assistant","reasoning_content": guideline_reasoning+ "\n"})
 
     # 非流式响应