Executive brief
A vulnerability in iTerm2, a popular terminal application for macOS, allows attackers to execute malicious code on a user's computer simply by having the user view a text file. By including specific hidden terminal commands in a file (like a readme.txt), an attacker can trick the software into running a hidden executable file located in the same directory. This could lead to a full system compromise if a user downloads and inspects a malicious folder or repository.
Technical details
iTerm2 fails to properly validate the origin of SSH conductor protocol messages, allowing untrusted terminal output (e.g., from 'cat'ing a file) to impersonate a legitimate conductor session. By emitting specific DCS 2000p and OSC 135 sequences, an attacker can manipulate iTerm2's state machine into constructing a 'run' command. Because iTerm2 writes these commands back to the local PTY when no real SSH session exists, the local shell interprets the base64-encoded payload as a command. If the attacker places an executable in the working directory with a name matching the predictable base64 output (e.g., 'ace/c+aliFIo'), the shell will execute it. This is addressed in version 3.6.10 by adding random UUID prefixes to pre-framer commands.
Affected products
- iTerm2 iTerm2 Up to and including 3.6.9
Timeline
- 2026-03-30: disclosed: Vulnerability reported to iTerm2 maintainers.
- 2026-03-31: patched: Fix committed to iTerm2 repository.
- 2026-04-18: advisory: Public disclosure by Calif.io.
- 2026-04-20: patched: Stable version 3.6.10 released.