Researchers Uncover PyPI Package Hiding Malicious Code Behind Image File

A malicious package discovered on the Python Package Index
(PyPI) has been found employing a steganographic trick to conceal
malicious code within image files.

The package in question, named “apicolor,” was
uploaded to the Python third-party repository on October 31, 2022,
and described as a “Core lib for REST API,” according to Israeli
cybersecurity firm Check Point^[1]. It has since been
taken down^[2].

Apicolor, like other rogue packages^[3]
detected recently, harbors its malicious behavior in the setup
script used to specify metadata associated with the package, such
as its dependencies.

This takes the form of a second package called “judyb” as well
as a seemingly harmless PNG file (“8F4D2uF.png”) hosted on Imgur,
an image-sharing service.

“The judyb code turned out to be a steganography module,
responsible [for] hiding and revealing hidden messages inside
pictures,” Check Point explained.

The attack chain entails using the judyb package to extract
obfuscated Python code embedded within the downloaded image, which,
upon decoding, is designed to retrieve and execute a malicious
binary from a remote server.

The development is part of an ongoing trend where threat actors
are increasingly setting their sights on the open source ecosystem
to exploit the trust associated with third-party software to mount
supply chain attacks.

Even more troublingly, such malicious libraries can be
incorporated into other open source projects and published on
GitHub, effectively broadening the scope and scale of the
attacks.

“These findings reflect careful planning and thought by a threat
actor, who proves that obfuscation techniques on PyPI have
evolved,” the company said.