feat：意图识别优化

app/schemas/requirement.py

@@ -23,6 +23,10 @@ class RequirementChange(ExtensibleModel):
     target_description: NonBlankString | None = Field(
         default=None, description="目标的自然语言描述，例如右侧区域、用户所说的那个图"
     )
+    candidate_ids: list[NonBlankString] = Field(
+        default_factory=list,
+        description="仅 unresolved 时填写：当前大屏中可能是该目标的真实组件 ID；没有可能的组件时留空，不虚构",
+    )
     operation: NonBlankString = Field(description=f"需求操作。{OPERATION_GUIDE}")
     parameters: JsonObject = Field(
         default_factory=dict,
@@ -46,6 +50,9 @@ class RequirementChange(ExtensibleModel):
             raise ValueError("只有 nodes 范围可以提供 target_ids")
         if self.target_scope == "unresolved" and self.target_description is None:
             raise ValueError("目标待确认时必须提供 target_description")
+        # 目标已确定时候选没有意义，直接丢弃，避免模型多填字段导致整份需求失效
+        if self.target_scope != "unresolved" and self.candidate_ids:
+            self.candidate_ids = []
         return self