feat: Add support for internal LiteLLM models (Qwen3, Gemma-4)

- Add dynamic routing in ChatController for Qwen3-Coder and Gemma-4-31B
- Update scaffold.html and chat.html model selection dropdowns
- Maintain existing OpenRouter fallback via application.yml

dat-gateway/src/main/resources/application.yml

@@ -9,6 +9,7 @@ spring:
   lifecycle:
     timeout-per-shutdown-phase: 20s
   ai:
+    # 기본(외부) 모델로 사용할 OpenRouter 설정입니다.
     openai:
       api-key: ${OPENROUTER_API_KEY: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}
       base-url: https://openrouter.ai/api/v1
@@ -18,6 +19,13 @@ spring:
           model: cohere/north-mini-code:free
           temperature: 0.3
 
+# 신한라이프 다중 모델 동적 라우팅을 위한 커스텀 설정
+shinhan:
+  ai:
+    base-url: https://dev-iam-litellm.shinhanlife.co.kr:18020
+    qwen-key: sk-UJ2IRenvaMPbY3ozQdj6zw
+    gemma-key: sk-EH107wYBZuU6RUTqthR17A
+
 server:
   port: 8081
   http2: