Breaking News
China’s national cybersecurity response agency has issued a warning about security risks associated with the agentic artificial intelligence tool OpenClaw, citing weak default protections and potential vulnerabilities that could expose users to cyberattacks.
In a statement posted on its WeChat account, the National Computer Network Emergency Response Technical Team of China said the tool has “extremely weak default security configurations” and should be deployed with caution.
The agency warned that attackers could exploit the system by embedding malicious instructions in webpages or distributing compromised plugins designed for OpenClaw. It also noted that several previously disclosed vulnerabilities could enable credential theft and other serious attacks if the software is not properly secured.
Authorities also highlighted the risk of user error, warning that OpenClaw users could accidentally delete important data while interacting with the agent-based system.
To mitigate these risks, the CERT recommended deploying the tool inside isolated containers and ensuring that its management ports are not exposed to the public internet. It also advised organizations to enforce strict authentication and access controls, disable automatic updates and limit access to third-party plugins.
Security concerns around OpenClaw have also been raised by analysts at Gartner, which earlier this year described the software as an “unacceptable cybersecurity risk” for business environments unless run in isolated test systems using temporary credentials.
The warning comes as OpenClaw adoption grows rapidly in China. The CERT said downloads and usage have surged in recent months, partly due to major domestic cloud platforms offering one-click deployment options.
Chinese internet company Tencent recently launched an OpenClaw-based tool called Work Buddy, which allows users to deploy the system and connect it with multiple chat platforms within minutes.
Following the security advisory, local reports indicated that some government agencies and state-owned banks in China have banned the use of OpenClaw, reflecting growing concerns about the security implications of rapidly deployed AI tools in sensitive environments.
In a statement posted on its WeChat account, the National Computer Network Emergency Response Technical Team of China said the tool has “extremely weak default security configurations” and should be deployed with caution.
The agency warned that attackers could exploit the system by embedding malicious instructions in webpages or distributing compromised plugins designed for OpenClaw. It also noted that several previously disclosed vulnerabilities could enable credential theft and other serious attacks if the software is not properly secured.
Authorities also highlighted the risk of user error, warning that OpenClaw users could accidentally delete important data while interacting with the agent-based system.
To mitigate these risks, the CERT recommended deploying the tool inside isolated containers and ensuring that its management ports are not exposed to the public internet. It also advised organizations to enforce strict authentication and access controls, disable automatic updates and limit access to third-party plugins.
Security concerns around OpenClaw have also been raised by analysts at Gartner, which earlier this year described the software as an “unacceptable cybersecurity risk” for business environments unless run in isolated test systems using temporary credentials.
The warning comes as OpenClaw adoption grows rapidly in China. The CERT said downloads and usage have surged in recent months, partly due to major domestic cloud platforms offering one-click deployment options.
Chinese internet company Tencent recently launched an OpenClaw-based tool called Work Buddy, which allows users to deploy the system and connect it with multiple chat platforms within minutes.
Following the security advisory, local reports indicated that some government agencies and state-owned banks in China have banned the use of OpenClaw, reflecting growing concerns about the security implications of rapidly deployed AI tools in sensitive environments.
See What’s Next in Tech With the Fast Forward Newsletter
Tweets From @varindiamag
Nothing to see here - yet
When they Tweet, their Tweets will show up here.
