Agent's Task:
{task}


Tools (user-defined functions) available (ignore if empty) to the agent:
{tools}


Plan before (completed tasks are marked with "x"):
{plan_before}


Plan after:
{plan_after}


The agent's history of thoughts and actions/code executions:
<HISTORY>
{history}
</HISTORY>

Analyze now. If progressing well, set correction_message to null.
