fix: disable end_prompt, filter more thinking patterns from antaf

end_prompt was being sent to antaf as user message, causing LLM
to output its thinking process. Disabled end_prompt and added
more thinking patterns to filter.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

config/.config.yaml

@@ -13,9 +13,7 @@ prompt: ""
 system_error_response: "抱歉，小虎现在有点忙，咱们稍后再聊。"
 
 end_prompt:
-  enable: true
+  enable: false
-  prompt: |
-    请你以关心对方健康的语气结束对话，提醒注意休息、按时吃药、保持好心情，像一位贴心的家庭医生告别。
 
 wakeup_words:
   - "你好小智"

modules/antaf/antaf_llm.py

@@ -26,12 +26,17 @@ class LLMProvider(LLMProviderBase):
     def _is_thinking(text):
         """检测蚂蚁阿福的内心思考/推理过程，这些不应该发给用户"""
         thinking_patterns = [
-            "用户问", "用户说", "用户的", "用户可能", "用户真正",
+            "用户问", "用户说", "用户的", "用户可能", "用户真正", "用户让我",
-            "我得", "我会", "我在想", "我决定", "我要",
+            "我得", "我会", "我在想", "我决定", "我要", "我需要确保", "我需要",
             "语气比较", "感觉他", "让他知道", "让他觉得",
             "先安抚", "得先", "不想表现",
             "整体语气", "这样能", "这样他",
             "所以我", "还带了个",
+            "首先记录", "是否有其他", "在这种情况下",
+            "必须遵循", "角色设定", "确保回复",
+            "避免任何生硬", "技术性的表达",
+            "提供的日期", "作为测试", "模拟使用",
+            "需要确认或澄清", "普遍适用性",
         ]
         for p in thinking_patterns:
             if p in text: