Malicious & suspicious packages

This is a collection of campaigns of malicious, suspicious and other unwanted or risky packages, that target Python Package Index ecosystem (more may come in the future).

Data are sourced (with rare exceptions) from what see in my analysing lab, and are exported here periodically. The classification is mostly arbitrary and may not follow any strict criteria.

All published packages have been manually checked. See some statistics for PyPI here.

Content

Packages are grouped in campaigns, which describe the activity. Usually, there is no additional description on the package level. As sometimes the behaviour changes over time, the campaign description may not describe exactly the package behaviour.

The current focus is on packages that are created for malicious purposes, and not on hacked versions of legitimate packages. You should generally assume all versions to be affected. Except of older packages, I now try to list affected versions. They are also automatically updates.

Malicioussnes classification

Each campaing has assigned a category, which describe how "malicious" I assessed the given campaign and how convinced I was about it. For instance, packages that do just a basic installation check or collect basic information (and have no other purpose, commonly used for dependency confusion checks) will fall into probably pentest category, indicating low harmfullness. Note that such an activity is anyway forbidden by PyPI's Acceptable Use Policy.

For a definition of a malicious package, see explanation in OpenSSF project.

Using data to secure your environment

You can secure your environment by using one of machine-readable formats available in the GitHub repository: the custom JSON or OSV Schema to build custom integration and leverage all provided data, including those with lower confidence, like highly suspisicous category.

Three categories (malicious, probably pentest and pentest) are also automatically ingested into OpenSSF Malicious Packages project, and from there into the main OSV.dev database. You can use one of their official scanners or third party tools to easy scan your dependencies. Note there is a delay between publishing data here and injesting them into OSV.dev.

Disclaimer

Data are presented as-is without any guarantee. Detection, classification, analyse & co. are done as a hobby activity, you use the information at your own risk. There are possible mistakes or highly opinionated classifications.