feat(llm): pass extra_body via model_kwargs in LangChain

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

backend/app/services/llm_client.py

@@ -110,10 +110,16 @@ class LLMClient:
             os.environ.setdefault("OPENAI_API_KEY", self.settings.llm_api_key)
             os.environ.setdefault("OPENAI_BASE_URL", self.settings.llm_base_url)
 
+            # Pass thinking/reasoning disable params via model_kwargs.
+            # LangChain's ChatOpenAI forwards model_kwargs as top-level
+            # request parameters, which is equivalent to OpenAI SDK's extra_body.
+            model_kwargs = self._build_extra_body() or None
+
             self._langchain_model = init_chat_model(
                 model=self.model,
                 model_provider="openai",
                 temperature=0.0,
+                model_kwargs=model_kwargs,
             )
         return self._langchain_model