CVE-2025-58764: Claude Code rg command had Command Injection that allowed bypass of user approval prompt for command execution
Claude Code is an agentic coding tool. Due to an error in command parsing, versions prior to 1.0.105 were vulnerable to a bypass of the Claude Code confirmation prompt to trigger execution of an untrusted command. Reliably exploiting this requires the ability to add untrusted content into a Claude Code context window. Users on standard Claude Code auto-update will have received this fix automatically. Users performing manual updates are advised to update to version 1.0.105 or the latest version.
Other sources
Due to an error in command parsing, it was possible to bypass the Claude Code confirmation prompt to trigger execution of an untrusted command. Reliably exploiting this requires the ability to add untrusted content into a Claude Code context window.
Users on standard Claude Code auto-update will have received this fix automatically. Users performing manual updates are advised to update to the latest version.
Thank you to the NVIDIA AI Red Team for reporting this issue!
— GitHub
Affected Software
Event History
Frequently Asked Questions
What is the severity of CVE-2025-58764?
CVE-2025-58764 has been rated as a high severity vulnerability due to the potential for untrusted command execution.
How do I fix CVE-2025-58764?
To fix CVE-2025-58764, upgrade to Claude Code version 1.0.105 or later to ensure the command parsing error is resolved.
What versions are affected by CVE-2025-58764?
CVE-2025-58764 affects all versions of Claude Code prior to 1.0.105.
Can CVE-2025-58764 be exploited remotely?
Yes, CVE-2025-58764 can be exploited remotely if the attacker can introduce untrusted content within the application.
What types of attacks does CVE-2025-58764 enable?
CVE-2025-58764 enables attacks that can trigger execution of malicious commands without proper user consent.