Command Execution Vulnerability in RTK AI by RTK
CVE-2026-54555
7.8HIGH
What is CVE-2026-54555?
The RTK AI product, prior to version 0.42.2, has a vulnerability within its command processing mechanism that allows potentially harmful commands to be executed without proper user validation. The issue arises because the permission splitter fails to correctly handle certain shell constructs, permitting a hidden command to be executed when an allowed prefix is present. This flaw can lead to unauthorized commands running silently in the background, thereby bypassing the designed security protocols intended for user authorization.
Affected Version(s)
rtk < 0.42.2
