CVE-2025-54795: 
JavaScript vulnerability analysis and mitigation

Claude Code, an agentic coding tool, was found to contain a critical command injection vulnerability (CVE-2025-54795) in versions below 1.0.20. The vulnerability was discovered by Elad Beber from Cymulate during Anthropic's Research Preview phase and was disclosed on August 4, 2025. The flaw affects the command parsing functionality of Claude Code, potentially allowing attackers to bypass the confirmation prompt system (GitHub Advisory, NVD).

The vulnerability stems from an error in command parsing that enables attackers to bypass Claude Code's confirmation prompt system and execute untrusted commands. The issue has been assigned a CVSS v4.0 base score of 8.7 (High), with the vector string CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N. The vulnerability is classified as CWE-78 (Improper Neutralization of Special Elements used in an OS Command Injection) (GitHub Advisory).

The vulnerability allows attackers to execute arbitrary commands without proper authorization, potentially leading to high impacts on system confidentiality, integrity, and availability. Successful exploitation requires the ability to inject untrusted content into a Claude Code context window (GitHub Advisory).

The vulnerability has been patched in Claude Code version 1.0.20. Users on standard Claude Code auto-update received this fix automatically after release. All versions prior to 1.0.24 have been deprecated and users have been forced to update, making current users unaffected by this vulnerability (GitHub Advisory).