Anthropic admitted that Claude Code had embedded steganography code targeting Chinese users, calling it an "abuse prevention experiment," and promised to roll back the code tomorrow.
Related
Anthropic responded to the Claude Code accusation of containing hidden code to detect Chinese users: It is true, and a complete rollback will be implemented tomorrow.
According to Foresight News , Thariq Shihipar, a member of the Claude Code team, responded to the issue of Anthropic embedding hidden code in Claude Code targeting Chinese users, stating that the mechanism was an "experimental" measure launched by the team in March 2026 to prevent unauthorized account resale and prevent model distillation attacks. The team has since deployed stronger mitigation measures and originally planned to take the experiment offline; the relevant PRs have been merged and will be fully rolled back in the new version released on July 2, 2026.
The Ministry of Industry and Information Technology issued a risk warning regarding the potential security backdoors in the AI programming tool Claude Code.
According to Odaily Odaily, the Cybersecurity Threat and Vulnerability Information Sharing Platform (NVDB) of the Ministry of Industry and Information Technology recently discovered that the AI programming tool Claude Code has a security backdoor vulnerability, which poses a serious threat. Claude Code is an AI programming tool developed by Anthropic, an American company, which can autonomously complete code writing and repair tasks based on text requirements. Due to its built-in monitoring mechanism, it can send sensitive information such as user location and identity to remote servers without the user's consent. The affected Claude Code versions are 2.1.91 to 2.1.196. It is recommended that relevant units and users immediately conduct a comprehensive investigation. For development terminals that have installed the above-mentioned affected versions, they should immediately uninstall or upgrade to the latest secure version that has removed the relevant backdoor code. Strengthen the control of external access permissions and traffic monitoring of development tools within the core business network segment to prevent the unauthorized transmission of sensitive data.
Anthropic: Claude Cowork will be available on mobile and web.
Odaily Odaily reports that Anthropic has announced that Claude Cowork will be available on mobile and web platforms, and will also extend its Claude Code double collaboration (Cowork) credits until August 5th. (Jinshi)
Claude Code can also be compromised? Accused of using strangers' passwords to modify databases across servers.
According to Beating's monitoring, a user submitted a report on GitHub stating that when using Anthropic's command-line AI assistant Claude Code, server IP addresses, usernames, and root passwords that did not belong to the user appeared out of thin air in the AI's dialogue context. Subsequently, the local AI assistant directly read these passwords, automatically connected to someone else's server via SSH, and performed write modifications to the database. In other words, the user's AI, using someone else's account and password, mistakenly connected to and modified someone else's production database. Community technical staff analyzed that the root cause of the problem may be the failure of the "prompt prefix caching" isolation mechanism of the large model. To reduce computational costs and speed up the process, the cloud-based large model caches the preceding text of user dialogues; if the cache keys between different users collide or become confused, someone else's confidential cache may be incorrectly concatenated into your dialogue. If this conjecture is true, any developer using Claude Code faces the risk of their server usernames, passwords, and core source code being obfuscated and leaked to other users. However, it cannot be ruled out that this is merely a case of the model's illusion coincidentally guessing the real IP address and weak password, or that local project history contaminated the context. This issue has been tagged with the security category (area:security) by GitHub's automated system, and all parties are awaiting official verification and conclusions.
Alibaba has internally announced a complete ban on Claude, effective July 10.
According to ChainCatcher, Alibaba announced today a complete ban on Anthropic's Claude series products. All employees are required to uninstall multiple models including Sonnet, Opus, and Fable, as well as Agent products including Claude Code. The ban will officially take effect on July 10th. Previously, Alibaba had encouraged employees to freely use external models, including Claude, through a large reimbursement policy. Sources indicate that this "reverse ban" is closely related to a series of recent frictions between the two companies. In mid-June, Anthropic accused Alibaba in the US Senate of using a large number of fake accounts to conduct "industrial-grade model distillation attacks," and subsequently implemented large-scale account bans and other risk control measures targeting Chinese users. Furthermore, developers recently revealed that Claude Code had a hidden detection mechanism built into previous versions, secretly collecting and marking data on Chinese time zones and domain names of several Chinese technology companies, including Alibaba. Although Anthropic subsequently acknowledged this "experimental" measure and announced its rollback and removal in the latest version, it still sparked widespread controversy and a crisis of trust within the industry.
Claude Code's real-time collaboration feature is available to individual users, while Pro and Max subscriptions can generate real-time web dashboards.
According to Beating, Anthropic announced that it will roll out the Live Artifacts feature of its terminal agent programming tool, Claude Code, to Pro and Max individual subscription accounts. Previously, this ability to convert command-line interactions into web dashboards in real time was only available to team and enterprise users. Now, individual developers can directly issue commands in the terminal when writing code or analyzing data using Claude Code, publishing session-generated data, prototypes, or architecture diagrams as independent web pages. The generated web pages will refresh in real time as the terminal task is executed. By sharing a private link, team members or customers can view the latest progress at any time without developers needing to repeatedly send screenshots or share terminal screens. Currently, these web pages are statically hosted, unable to send network requests to external networks from within the page, nor can they save form inputs or perform multi-route redirects. Users can use this feature after logging in to the terminal.