OpenClaw Research: A Systematic Survey of Large Language Model Agents in Open Deployment

Shuo Lu^♦,†, Kecheng Yu^♦,†, Siru Jiang^♦,†, Yinuo Xu^♦, Bing Zhan^♦, Yanbo Wang^♦, Changxin Ke^♣, Yuan Xu^♦, Xin Xiong^♠, Xinyun Zhou^¶, Yihua Shao^♦, Zhengbo Wang^△,♦, Lijun Sheng^△,♦, Aijing Yu^♠, Haosen Yang^#, Yunpu Ma^♮, Hao Tang^‖, Nicu Sebe^°, Tat-Seng Chua^§, Philip Torr^★, Ran He^♦, Jian Liang^♦,‡

^♦Institute of Automation, Chinese Academy of Sciences   ^♣Institute of Computing Technology, Chinese Academy of Sciences   ^♠Institute of Information Engineering, Chinese Academy of Sciences   ^¶Zhejiang University   ^△University of Science and Technology of China   ^#University of Surrey   ^♮Ludwig Maximilian University of Munich   ^‖Peking University   ^°University of Trento   ^§National University of Singapore   ^★University of Oxford

Abstract

Autonomous agents powered by large language models are moving from curated demos to persistent, open-world deployment. We formalize this setting through an agent-system tuple A = ⟨ π, env, pop, substrate ⟩ and derive four principles of openness: Open Policy, Open Environment, Open Population, and Open Substrate. These principles structure the taxonomy around five research areas: Learning & Evolving, Safety & Security, Claw Society, Infrastructure & Systems, and Applications. This survey provides a roadmap for governing LLM agents as they move beyond laboratory settings into large-scale open deployment.

📖 Citation

@article{lu2026openclaw,
  title   = {OpenClaw Research: A Systematic Survey of Large Language Model Agents in Open Deployment},
  author  = {Shuo Lu and Kecheng Yu and Siru Jiang and Yinuo Xu and Bing Zhan and Yanbo Wang and Changxin Ke and Yuan Xu and Xin Xiong and Xinyun Zhou and Yihua Shao and Zhengbo Wang and Lijun Sheng and Aijing Yu and Haosen Yang and Yunpu Ma and Hao Tang and Nicu Sebe and Tat-Seng Chua and Philip Torr and Ran He and Jian Liang},
  journal = {arXiv preprint},
  year    = {2026}
}