Securing Python's binary dependencies

PROJECT

Improving security by bringing hidden binary dependency relationships to light.

About


A large amount of infrastructure relies on Python for its ability to easily interface with compiled libraries, particularly AI and scientific software. However, package management tooling in the Python ecosystem does not record these binary dependency relationships, which most frequently occur when Python code calls compiled libraries written in languages such as C or Rust.

This means that developers cannot track or remediate CVEs in these dependencies; cannot know when these libraries become unmaintained “Bernies”; and cannot offer support to the maintainers of their binary dependencies. Given that around 20.4% of popular Python packages include binary dependencies, this status quo puts critical Python-based technology, from AI to built infrastructure, at risk of cyberattacks and degradation.

We propose a solution then can make a large positive impact on the security of the Python ecosystem with a comparatively small investment; can be integrated as a default into existing tools; and is structured so that it can be, in principle, ported to other language ecosystems in the future.

Resources:

Our team

Today's Balance
£0.00
Project funded
£53,531

Contribute


Become a financial contributor.

Financial Contributions

Custom contribution
Donation
Make a custom one-time or recurring contribution.