The paper introduces a new task called Proactive Agent Planning, which requires language agents to predict clarification needs based on user-agent conversations and agent-environment interactions, invoke external tools to collect valid information, and generate plans to fulfill user demands. To study this practical problem, the authors establish a new benchmark dataset named Ask-before-Plan. They propose a novel multi-agent framework, Clarification-Execution-Planning (CEP), consisting of three agents specialized in clarification, execution, and planning. The CEP framework includes trajectory tuning for the clarification and execution agents and a memory recollection mechanism for the dynamic execution agent. Extensive evaluations and comprehensive analyses validate the effectiveness of the proposed CEP framework. The paper also discusses the limitations and ethical considerations of the work.The paper introduces a new task called Proactive Agent Planning, which requires language agents to predict clarification needs based on user-agent conversations and agent-environment interactions, invoke external tools to collect valid information, and generate plans to fulfill user demands. To study this practical problem, the authors establish a new benchmark dataset named Ask-before-Plan. They propose a novel multi-agent framework, Clarification-Execution-Planning (CEP), consisting of three agents specialized in clarification, execution, and planning. The CEP framework includes trajectory tuning for the clarification and execution agents and a memory recollection mechanism for the dynamic execution agent. Extensive evaluations and comprehensive analyses validate the effectiveness of the proposed CEP framework. The paper also discusses the limitations and ethical considerations of the work.