Search
The China Academy of Information and Communications Technology, working hand-in-hand with research teams from Shanghai Jiao Tong University and Nanjing University, carried out a comprehensive security review of the open-source autonomous agent framework, OpenClaw. Their efforts led to the identification and verification of a serious command injection vulnerability driven by Large Language Models (LLMs), along with successful validation of the attack chain across multiple mainstream model environments. The collaborative team promptly reported the vulnerability to the GitHub community and provided crucial assistance in mitigating this high-risk security threat. Moving forward, the China Academy of Information and Communications Technology remains committed to ongoing security risk monitoring, ensuring robust and secure applications within the industry.
-
C114 Communication Network -
Communication Home
