Inconsistency with summary_mode for MultiStepAgent._generate_updated_plan

[Jus-Codin]

In MultiStepAgent._generate_updated_plan it mentions that summary_mode is False so as to not extract the previous planning steps from memory.

smolagents/src/smolagents/agents.py

Lines 432 to 435 in 2f300dc

	def _generate_updated_plan(self, task: str, step: int) -> Tuple[ChatMessage, ChatMessage]:
	# Do not take the system prompt message from the memory
	# summary_mode=False: Do not take previous plan steps to avoid influencing the new plan
	memory_messages = self.write_memory_to_messages()[1:]

However, in PlanningStep.to_messages, the previous plan is added to the message history if summary_mode is False.

smolagents/src/smolagents/memory.py

Lines 160 to 166 in 2f300dc

	if not summary_mode: # This step is not shown to a model writing a plan to avoid influencing the new plan
	messages.append(
	Message(
	role=MessageRole.ASSISTANT, content=[{"type": "text", "text": f"[PLAN]:\n{self.plan.strip()}"}]
	)
	)
	return messages

The comment in the above snippet also references hiding the step from a model writing a plan, which means that when writing a plan, summary_mode should be set to True. Am I missing something out here?

[Jus-Codin]

Just figured out that summary_mode is intended to be used to provided agents with a managed agent's summary of work. However, the above mentioned comments should probably be either removed or changed to avoid confusion