Fix handling multiple actions requested by grounding agent in an iteration

src/khoj/processor/operator/operator_agent_binary.py

@@ -221,6 +221,9 @@ Focus on the visual action and provide all necessary context.
             else:
                 message = f"**Action Result**: {json.dumps(result_content)}"
                 images = []
+            if not images:
+                action_result["content"] = [{"type": "text", "text": message}]
+            else:
                 action_result["content"] = construct_structured_message(
                     message=message,
                     images=images,