feat(conversations): 统一对话 ID 与标题，敏感词结构化返回，新增删除接口与 CORS 支持

- schemas：ConversationResponse/ConversationDocOut 增加 title；ConversationDocOut 增加 id（_id 可选）
- schemas：新增 SensitiveWordInfoResponse、DeleteResult；MessageResponse/MessageSendResult 的 sensitive_words_found 改为结构化列表

关联问题：#7 #8 #9 #10 #11 #12

app/schemas/conversation.py

@@ -2,6 +2,19 @@ from typing import List, Optional
 from pydantic import BaseModel
 from datetime import datetime
 
+class SensitiveWordInfoResponse(BaseModel):
+    """敏感词信息结构化模型
+    字段说明：
+    - word: 敏感词文本
+    - category: 一级分类
+    - subcategory: 二级分类
+    - severity: 严重程度（数值越大越严重）
+    """
+    word: str
+    category: Optional[str] = None
+    subcategory: Optional[str] = None
+    severity: Optional[int] = None
+
 class MessageCreate(BaseModel):
     content: str
 
@@ -10,17 +23,20 @@ class MessageResponse(BaseModel):
     content: str
     timestamp: datetime
     contains_sensitive_words: bool
-    sensitive_words_found: List[str]
+    sensitive_words_found: List[SensitiveWordInfoResponse]
 
 class ConversationResponse(BaseModel):
     id: str
+    title: Optional[str] = None
     messages: List[MessageResponse]
     created_at: datetime
     updated_at: datetime
 
 class ConversationDocOut(BaseModel):
-    _id: str
+    id: str
+    _id: Optional[str] = None
     user_id: Optional[str] = None
+    title: Optional[str] = None
     messages: List[MessageResponse]
     created_at: datetime
     updated_at: datetime
@@ -28,7 +44,11 @@ class ConversationDocOut(BaseModel):
 class CreatedId(BaseModel):
     id: str
 
+class DeleteResult(BaseModel):
+    deleted: bool
+    message: Optional[str] = None
+
 class MessageSendResult(BaseModel):
     contains_sensitive_words: bool
-    sensitive_words_found: List[str]
+    sensitive_words_found: List[SensitiveWordInfoResponse]
     assistant_response: str
\ No newline at end of file