Prompt Injection in Python virtualenv

Executive Summary

CVE-2026-113723 is a command injection vulnerability in the virtualenv tool affecting versions 21.14.8 and earlier. It occurs when malicious prompt text is passed to virtualenv and later executed when the environment is activated and the prompt is drawn. The issue stems from improper handling of shell prompts in bash and other shells, where attacker-controlled input like $(touch PWNED) could execute commands with the user's privileges.

Useful 0 Not Useful 0

Detection Guidance

Check virtualenv version with pip show virtualenv. If version is 21.14.8 or earlier, the system is vulnerable. Inspect activation scripts in virtual environments for prompt handling issues, particularly in bash and zsh shells.

Useful 0 Not Useful 0

Impact Analysis

An attacker could trick you into activating a virtual environment with a malicious prompt. When you view the prompt in certain shells (like ksh93 or bash in compatibility mode), embedded commands execute automatically. This could allow the attacker to create files, modify data, or perform other actions with your user privileges. The attack requires both malicious prompt input and your activation of the environment.

Useful 0 Not Useful 0

Mitigation Strategies

Upgrade virtualenv to version 21.14.9 or later using pip install --upgrade virtualenv. Alternatively, set VIRTUAL_ENV_DISABLE_PROMPT=1 to disable prompts entirely. Regenerate all virtual environment activation scripts after upgrading.

Useful 0 Not Useful 0