In its design for automatic terminal command execution, AI Code offers two options: Execute safe commands and execute all commands. The description for the form…

CVE-2026-30304 affects AI Code's automatic terminal command execution feature. The vulnerability allows attackers to use prompt injection attacks to bypass safety mechanisms by wrapping malicious commands in templates that mislead the AI model into classifying them as 'safe' commands. This results in arbitrary command execution without user approval, circumventing the intended security controls in the 'Execute safe commands' mode.