fix: Use node.name in node_label calculation (minimal change)

- Add name field to WorkflowNode model
- Extend node_label priority: node.name > prompt_slug > node_type-id
- No new fields in NodeExecutionState (uses existing debug_prompt_slug)
- Simpler approach than previous attempt to avoid 504 timeout

backend/workflow_executor.py

@@ -167,7 +167,10 @@ async def execute_workflow(
             )
 
             # Add human-readable label to node_state for debug UI
-            node_label = node.prompt_slug if hasattr(node, 'prompt_slug') and node.prompt_slug else f"{node.type.value}-{node_id[:8]}"
+            # Priority: node.name (user-defined) > prompt_slug > node_type-id
+            node_label = node.name if hasattr(node, 'name') and node.name else (
+                node.prompt_slug if hasattr(node, 'prompt_slug') and node.prompt_slug else f"{node.type.value}-{node_id[:8]}"
+            )
             if not node_state.debug_prompt_slug:
                 node_state.debug_prompt_slug = node_label
 

backend/workflow_models.py

@@ -196,6 +196,7 @@ class WorkflowNode(BaseModel):
     """
     id: str = Field(..., description="Eindeutige Knoten-ID")
     type: NodeType = Field(..., description="Knotentyp")
+    name: Optional[str] = Field(None, description="Node-Name (vom Editor)")
     position: Optional[Position] = Field(None, description="Position im visuellen Editor")
 
     # ANALYSIS-Knoten