Malicious code in agentaix (PyPI)

-= Per source details. Do not edit below this line.=-

Source: amazon-inspector (c403a9ecab78c519698c7faebcb0f9de4f8af059a4c3f5803acf8c534923138a)

On import agentaix, src/agentaix/__init__.py executes drop_readme_on_desktop() at module top level, copying the bundled file readme.txt to ~/Desktop/readme.txt. The file is not documentation: it is a Python remote-access client with a #!/usr/bin/env python3 shebang that registers username@hostname with a hardcoded command-and-control server at https://googleforum.pythonanywhere.com and polls /api/client/<id>/... for operator commands. Handlers implement arbitrary shell execution via subprocess.run(command, shell=True) (handle_exec), arbitrary file read with base64-encoded upload back to the C2 (handle_download), plus upload, delete, rename, and open operations. The dropper stages the payload under a benign filename in a highly visible location to induce the user to run it; once run, the operator obtains interactive shell and arbitrary file read on the host, enabling theft of SSH keys, cloud credentials, and other secrets. pyproject.toml ships placeholder author metadata (Your Name / [email protected], github.com/yourname/agentaix), consistent with an unattributed malicious upload.

Source: kam193 (f42a135102abc5c66418237d4c6d9f8ee608570f8d4d5146463924666e0338e1)

The package installs a persistent job that allows the attacker for exfiltrating files and executing remote code on the victim's machine.

Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.

Campaign: 2026-10-agentaix

Reasons (based on the campaign):

• files-exfiltration

• peristence-autorun

• The package contains code to execute remote commands (probably limited to a specific set) on the victim's machine.

• rat

• persistence

• Downloads and executes a remote malicious script.