feat(workflow): add per-agent vision configuration

backend/services/workflow_engine.py

@@ -18,6 +18,8 @@ class AgentStageConfig:
     llm_resource_id: str
     asr_resource_id: str
     tts_resource_id: str
+    vision_enabled: bool
+    vision_model_resource_id: str | None
     tool_ids: tuple[str, ...]
     knowledge_base_id: str | None
     knowledge_mode: str
@@ -119,6 +121,10 @@ class WorkflowEngine:
             llm_resource_id=str(source.get(llm_key) or ""),
             asr_resource_id=str(source.get(asr_key) or ""),
             tts_resource_id=str(source.get(tts_key) or ""),
+            vision_enabled=bool(source.get("visionEnabled", False)),
+            vision_model_resource_id=(
+                str(source.get("visionModelResourceId") or "") or None
+            ),
             tool_ids=tuple(str(tool_id) for tool_id in source.get("toolIds") or []),
             knowledge_base_id=knowledge_base_id or None,
             knowledge_mode=(
@@ -139,6 +145,14 @@ class WorkflowEngine:
             turn_config=dict(turn_config),
         )
 
+    def uses_vision(self) -> bool:
+        """Return whether any effective Agent stage needs a camera stream."""
+        return any(
+            self.agent_stage_config(node_id).vision_enabled
+            for node_id, node in self.nodes.items()
+            if node.get("type") == "agent"
+        )
+
     def prompt_for(self, node_id: str, store: DynamicVariableStore) -> str:
         """Build the Agent system prompt according to its inheritance setting."""
         prompt = store.render(str(self.data(node_id).get("prompt") or "").strip())