Task decomposition can be done through simple prompting by LLM, task-specific instructions, or human inputs. These components work together to enable autonomous behavior by allowing the agent to plan, ...