基础模型驱动的工业智能体：技术成熟度、能力变迁与未竟之挑战（英文版）

Foundation models, particularly large language models, are increasingly inte-grated into agent architectures for industrial tasks such as decision support,process monitoring, and engineering automation. Yet evidence on their pur-poses, capabilities, and limitations remains fragmented across domains. Thiswork examines how mature foundation-model-based agent systems are in indus-trial contexts, how their functional profile differs from conventional agentsystems, and which limitations persist. A