diff --git a/agenthub/langchains_agent/utils/prompts.py b/agenthub/langchains_agent/utils/prompts.py
index 1c7f297d2081..a5dca94858ed 100644
--- a/agenthub/langchains_agent/utils/prompts.py
+++ b/agenthub/langchains_agent/utils/prompts.py
@@ -165,11 +165,15 @@ def get_request_action_prompt(
 def parse_action_response(response: str) -> Action:
     parser = JsonOutputParser(pydantic_object=_ActionDict)
     action_dict = parser.parse(response)
+    if 'content' in action_dict:
+        # The LLM gets confused here. Might as well be robust
+        action_dict['contents'] = action_dict.pop('content')
+
     action = ACTION_TYPE_TO_CLASS[action_dict["action"]](**action_dict["args"])
     return action
 
 def parse_summary_response(response: str) -> List[Action]:
     parser = JsonOutputParser(pydantic_object=NewMonologue)
     parsed = parser.parse(response)
-    thoughts = [ACTION_TYPE_TO_CLASS[t['action']](**t['args']) for t in parsed['new_monologue']]
-    return thoughts
+    #thoughts = [ACTION_TYPE_TO_CLASS[t['action']](**t['args']) for t in parsed['new_monologue']]
+    return parsed['new_monologue']