CVE-2026-24887: Claude Code has a Command Injection in find Command Bypasses User Approval Prompt
Claude Code is an agentic coding tool. Prior to version 2.0.72, due to an error in command parsing, it was possible to bypass the Claude Code confirmation prompt to trigger execution of untrusted commands through the find command. Reliably exploiting this required the ability to add untrusted content into a Claude Code context window. This issue has been patched in version 2.0.72.
Other sources
Due to an error in command parsing, it was possible to bypass the Claude Code confirmation prompt to trigger execution of untrusted commands through the find command. Reliably exploiting this required the ability to add untrusted content into a Claude Code context window.
Users on standard Claude Code auto-update have received this fix already. Users performing manual updates are advised to update to the latest version.
Claude Code thanks https://hackerone.com/alexbernier for reporting this issue!
— GitHub
Affected Software
Event History
Frequently Asked Questions
What is the severity of CVE-2026-24887?
CVE-2026-24887 is considered a high-severity vulnerability due to its potential for command injection.
How do I fix CVE-2026-24887?
To resolve CVE-2026-24887, upgrade to Claude Code version 2.0.72 or later.
What software is affected by CVE-2026-24887?
CVE-2026-24887 affects the @anthropic-ai/claude-code package versions prior to 2.0.72.
What type of vulnerability is CVE-2026-24887?
CVE-2026-24887 is a command injection vulnerability that allows execution of untrusted commands.
Is user approval prompt for commands bypassed in CVE-2026-24887?
Yes, CVE-2026-24887 allows the bypassing of the user approval prompt due to a flaw in command parsing.