Cybersecurity researchers have identified a malicious Python package that purports to be an offshoot of the popular requests library and has been found concealing a Golang-version of the Sliver command-and-control (C2) framework within a PNG image of the project’s logo.
The package employing this steganographic trickery is requests-darwin-lite, which has been downloaded 417 times prior to it being taken down from the Python Package Index (PyPI) registry.
Requests-darwin-lite “appeared to be a fork of the ever-popular requests package with a few key differences, most notably the inclusion of a malicious Go binary packed into a large version of the actual requests side-bar PNG logo,” software supply chain security firm Phylum said.
The changes have been introduced in the package’s setup.py file, which has been configured to decode and execute a Base64-encoded command to gather the system’s Universally Unique Identifier (UUID).
In what’s an interesting twist, the infection chain proceeds only if the identifier matches a particular value, implying that the author(s) behind the package is looking to breach a specific machine to which they are already in possession of the identifier obtained through some other means.
This raises two possibilities: Either it’s a highly targeted attack or it’s some sort of a testing process ahead of a broader campaign.
Should the UUID match, the requests-darwin-lite proceeds to read data from a PNG file named “requests-sidebar-large.png,” which bears similarities with the legitimate requests package that ships with a similar file called “requests-sidebar.png.”
What’s different here is that while the real logo embedded within requests has a file size of 300 kB, the one contained inside requests-darwin-lite is around 17 MB.
The binary data concealed in the PNG image is the Golang-based Sliver, an open-source C2 framework that’s designed to be used by security professionals in their red team operations.
The exact end goal of the package is currently unclear, but the development is once again a sign that open-source ecosystems continue to be an attractive vector to distribute malware.
With a vast majority of codebases relying on open-source code, the steady influx of malware into npm, PyPI, and other package registries, not to mention the recent XZ Utils episode, has highlighted the need for addressing issues in a systematic manner that otherwise can “derail large swaths of the web.”