#development-security

AI CLI Agents Vulnerable to Git Config Code Injection

Security researchers uncover critical flaws across AI coding assistants where weaponized repositories execute arbitrary shell commands bypassing sandbox controls.