China’s National Vulnerability Database (NVDB) has issued a security alert about Claude Code. The agency claims the AI coding assistant contains an alleged backdoor that can transmit user information without consent.
The warning covers several versions of Anthropic’s tool. Chinese authorities urge organizations to remove the affected software or install the latest version. Anthropic has not publicly responded to the claims.
China Raises Data Collection Concerns
According to the NVDB, the alleged backdoor can send users’ geographic location and identity-related information to remote servers.
The agency claims the monitoring mechanism works without user permission. If true, it could create privacy and security risks for organizations that rely on the AI coding assistant.
The NVDB operates under China’s Ministry of Industry and Information Technology. It serves as the country’s national vulnerability database.
Several Claude Code Versions Are Affected
The advisory applies to Claude Code versions 2.1.91 through 2.1.196.
Chinese authorities recommend reviewing all affected systems immediately. They also advise users to uninstall the impacted versions or upgrade to the latest release. According to the advisory, the newer version removes the alleged backdoor.
The agency also recommends stronger network monitoring. It urges organizations to tighten controls over development tools that communicate with external services.
Alibaba Has Already Restricted Claude Code
The warning follows Alibaba’s recent decision to ban employees from using Claude Code for work.
Instead, the company directed staff to use its own AI coding platform, Qoder.
Reports suggest Alibaba became concerned about features that could identify users located in China. Those concerns prompted the company to restrict the tool before the government issued its latest advisory.
Anthropic Has Yet to Comment
Anthropic has not publicly responded to the allegations.
As a result, the claims remain unverified. It also remains unclear whether independent security researchers have confirmed the alleged monitoring mechanism.
AI Coding Tools Face Growing Scrutiny
The Claude Code warning reflects growing concerns about AI-powered development tools.
Governments and businesses continue to examine how these platforms collect, process, and transmit user information. The latest advisory also shows how geopolitical tensions are increasingly shaping discussions around AI security and software supply chains.
0 responses to “China Warns of Alleged Backdoor in Anthropic’s Claude Code”