China Academy of Information and Communications Technology Jointly Discovers OpenClaw Critical Vulnerability with Universities and Assists in Rapid Patching

MetaMuskRat · 2026-03-18T08:43:47+00:00

Recently, a research team conducted a security audit of the open-source agent framework OpenClaw and discovered LLM-driven command injection vulnerabilities, completing validation of the attack chain. The findings have been reported to GitHub to assist in fixing this critical security risk. China Academy of Information and Communications Technology will continue to monitor the security risks of this framework.

MetaMuskRat

2026-03-18 08:43:47

Abstract generation in progress

People’s Financial News, March 16 — Recently, the China Academy of Information and Communications Technology (CAICT), together with Shanghai Jiao Tong University and Nanjing University, formed a joint research team to conduct an in-depth security audit of the open-source autonomous intelligent agent framework OpenClaw. Through static analysis and dynamic practical testing, they discovered and verified a serious vulnerability: LLM-Driven Command Injection. The research team has completed attack chain verification across various mainstream model environments and reported the issue to the GitHub community, assisting in the prompt repair of this high-risk security flaw. CAICT will continue to monitor the security risks of OpenClaw and support the industry in safe application development.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

1 Likes