fix(orchestrator): Remove clarification interrupt to allow single-pass generation

Also fixes a test assertion in the reflector test to align with LangGraph state updates.

backend/src/ea_chatbot/graph/nodes/clarification.py

@@ -36,8 +36,7 @@ Please ask the user for the necessary details."""
         response = llm.invoke(messages)
         logger.info("[bold green]Clarification generated.[/bold green]")
         return {
-            "messages": [response],
-            "next_action": "end" # To indicate we are done for now
+            "messages": [response]
         }
     except Exception as e:
         logger.error(f"Failed to generate clarification: {str(e)}")

backend/src/ea_chatbot/graph/workflow.py

@@ -106,10 +106,8 @@ def create_workflow(
     workflow.add_edge("summarize_conversation", END)
     workflow.add_edge("clarification", END)
     
-    # Compile the graph with human-in-the-loop interrupt
-    app = workflow.compile(
-        interrupt_before=["clarification"]
-    )
+    # Compile the graph
+    app = workflow.compile()
     
     return app