MAL-2026-10760 Malicious code in abseil-py (PyPI)

Distribution name abseil-py name-squats Google's widely-used absl-py (Abseil Python common libraries); the import name abseil_py also differs from the legitimate absl module. During bdist_wheel, setup.py walks the parent-process chain via psutil and matches process names against a hardcoded list of AI coding agent markers (claude-code, codex, cursor, gemini, grok, copilot, aider, opencode, qwen, amazon-q). On a match, _report() POSTs a JSON payload containing the package name, event, matched agent label, and timestamp to the hardcoded endpoint https://vero-research.com/ using urllib.request.urlopen with a spoofed browser User-Agent. The init prints a 'research study' disclaimer only after installation has already occurred and the beacon has fired. The behavior transmits information about the installer's environment (which AI coding tool triggered the install) to an author-controlled third-party endpoint without prior disclosure or opt-out.

Installing the package or importing the module exfiltrates basic information about the host, and the package has no other purpose.

Category: PROBABLY_PENTEST - Packages looking like typical pentest packages, but also anything that looks like testing, exploring pre-prepared kits, research & co, with clearly low-harm possibilities.

Campaign: GENERIC-standard-pypi-install-pentest

Reasons (based on the campaign):

• The package contains code to exfiltrate basic data from the system, like IP or username. It has a limited risk.

• The package overrides the install command in setup.py to execute malicious code during installation.