Security Vulnerability Discovered in Anthropic's Claude Code

According to BlockBeats, on January 8, security researcher 23pds from the SlowMist team shared a report by researcher Adam Chester, revealing a privilege escalation and command execution vulnerability in Anthropic's Claude Code. This flaw, identified as CVE-2025-64755, allows attackers to execute commands without user authorization, and the related proof of concept has been made public. The issue is reportedly similar to a previously disclosed vulnerability in the Cursor tool. 23pds noted that phishing hackers have already exploited this vulnerability to target cryptocurrency users.