ml2000PyPI
Advisory published Updated
ml2000 is a confirmed malicious PyPI package (MAL-2026-4756) that runs destructive / sabotage code (malicious version 0.1.4). Do not install it — remove it immediately and rotate any exposed credentials.
Malicious code in ml2000 (PyPI)
What this malware does
On invoking the ml2000 CLI with no arguments, interactive_menu() in src/ml_labs/generator.py writes a batch file and launches it via subprocess.Popen(["cmd.exe", "/c", bat_path], creationflags=DETACHED_PROCESS | CREATE_NO_WINDOW). The batch script runs taskkill /IM WindowsTerminal.exe /F, taskkill /IM cmd.exe /F, taskkill /IM powershell.exe /F, then pipx uninstall ml2000, then deletes itself. The use of detached/no-window flags hides this from the user, and the README advertises only ML notebook code generation — the destructive behavior is undisclosed. This is install/use-time destruction of installer-side resources: open terminal sessions are force-killed (causing loss of unsaved work in any other shell the user has open) and the package removes itself behind the user's back. Project metadata is also placeholder (Your Name <[email protected]>), corroborating that this is not a legitimate maintained release.
Malicious versions
Indicators of compromise (SHA-256)
Detection & response playbook
Destructive / sabotageFind it
Search your lockfiles and build artifacts for ml2000 (version 0.1.4).
If you installed it — respond
ml2000 carries a destructive/sabotage payload. Remove it immediately, restore any affected data from clean backups, and verify integrity of build outputs that may have been tampered with.
Did it already run?
If ml2000 was installed, its post-install payload may already have run. Removing the package does not undo that — check outbound connections and credential use from the install window onward.
Frequently asked questions
Campaign
References
Credits
- Amazon Inspector · finder
Detect & block this
O3 blocks ml2000-class packages before install and in CI — and if it already ran, its runtime egress monitoring catches the malicious outbound activity and severs the channel.