fix: SSE streaming - WorkflowNode label attribute and ai_insights column name

- workflow_executor.py: Generate node_label from prompt_slug or node.type (WorkflowNode has no label attribute)
- prompts.py: Fix INSERT statement - use 'created' column instead of 'created_at'

SSE endpoint now works correctly for workflow execution streaming.

backend/routers/prompts.py

@@ -1555,8 +1555,8 @@ async def execute_unified_prompt_stream(
                     with get_db() as conn:
                         cur = get_cursor(conn)
                         cur.execute(
-                            """INSERT INTO ai_insights (profile_id, scope, content, metadata, created_at)
-                               VALUES (%s, %s, %s, %s, NOW())""",
+                            """INSERT INTO ai_insights (profile_id, scope, content, metadata, created)
+                               VALUES (%s, %s, %s, %s, CURRENT_TIMESTAMP)""",
                             (profile_id, prompt_slug, content, json.dumps({"prompt_type": result['type']}))
                         )
                         conn.commit()

backend/workflow_executor.py

@@ -171,10 +171,12 @@ async def execute_workflow(
 
             # NEW: Progress-Callback aufrufen (für SSE Streaming)
             if progress_callback:
+                # Create a meaningful label for the node
+                node_label = node.prompt_slug if hasattr(node, 'prompt_slug') and node.prompt_slug else f"{node.type.value}-{node_id[:8]}"
                 await progress_callback("node_complete", {
                     "node_id": node_id,
                     "node_type": node.type,
-                    "node_label": node.label,
+                    "node_label": node_label,
                     "status": node_state.status.value,
                     "total_nodes": len(graph.nodes),
                     "completed_nodes": len([ns for ns in node_states if ns.status in [NodeStatus.COMPLETED, NodeStatus.SKIPPED]]),