AI and data stack advisories

Severe, 6 weeks1496Projects326

1496 severe, 6 weeks · 326 projects

AnthropicGHSA-qxfv-fcpc-w36x

Command Injection in Claude Code rg command allowed bypass of user approval prompt for command execution

Anthropic

CVE-2025-58764 · Published Sep 9, 2025

High8.7
Fix: upgrade to 1.0.105 or later
GitHub advisory

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!

Affected versions

PackageAffectedFixed in
@anthropic-ai/claude-code
npm
< 1.0.1051.0.105
Details and references
CVSS 4.0
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
Severity from
GitHub (reviewed advisory)
Weakness
CWE-94

More Anthropic advisories

All Anthropic
High 8.7Oct 3, 2025Command execution prior to Claude Code startup trust dialogCVE-2025-59536fixed in v1.0.111Low 2.3Oct 3, 2025Permission deny bypass through symlinkCVE-2025-59829fixed in v1.0.120High 7.7Sep 24, 2025Claude Code Vulnerable to Arbitrary Code Execution via Plugin Autoloading with Specific Yarn VersionsCVE-2025-59828fixed in v1.0.39High 8.7Sep 9, 2025Claude Code vulnerable to arbitrary code execution caused by maliciously configured git emailCVE-2025-59041fixed in 1.0.105High 8.7Sep 2, 2025Claude Code Vulnerable to Arbitrary Code Execution Due to Insufficient Startup WarningGHSA-ph6w-f82w-28w6fixed in v1.0.87High 7.1Aug 15, 2025Permissive Default Allowlist Enables Unauthorized File Read and Network Exfiltration in Claude CodeCVE-2025-55284fixed in v1.0.4