Old Version
News Brief

China Flags Risks in OpenClaw, Releases Safety Rules

The National Vulnerability Database of China (NVDB) under the Ministry of Industry and Information Technology (MIIT) recently released "six dos and six don'ts" to prevent potential security risks associated with OpenClaw.

By NewsChina Updated May.1

The National Vulnerability Database of China (NVDB) under the Ministry of Industry and Information Technology (MIIT) recently released "six dos and six don'ts" to prevent potential security risks associated with OpenClaw. 

An open-source autonomous AI agent developed by Austrian developer Peter Steinberger who now works with OpenAI, OpenClaw, nicknamed "Lobster" in China, connects large language models with local systems and is able to perform real-world tasks through natural languages. After targeted training, it supports intelligent office work, software development, content management, document editing and financial transactions. 

However, as the training and use require high-level system permissions, there could be significant risks such as supply chain attacks, sensitive information leaks and device hijacking, which may be exacerbated by third-party paid training services. 

The NVDB's guidelines say Open- Claw users must download the latest official version, restrict internet exposure and control permissions, while prohibiting dangerous operations such as using third-party mirror versions and granting administrator-level authorization.

Print