Attackers Placed Fake Claude Download Page on claude.ai Domain
Anthropic
Security researchers discovered that attackers created a fake download page for the Claude AI model on a subdomain of the official claude.ai website. The attack was made possible due to a misconfigured DNS, which allowed hackers to host a phishing page that could be used to steal data or spread malware.
Cybersecurity specialists have identified a new threat: a fake download page for the AI assistant Claude appeared on a subdomain of the official website claude.ai, owned by Anthropic. Attackers exploited an error in DNS (Domain Name System) settings to redirect part of the traffic to their own server. As a result, users who clicked on a link to the claude.ai subdomain landed on a phishing page mimicking the real Anthropic website. This page offered a download of the Claude application, but instead of a legitimate file, visitors could receive malware or hand over their data to scammers. Anthropic representatives have already taken steps to close the vulnerability and remove the fake page.
Source: Anthropic (GNews) —
original