Researcher: counterfeit packages on the official Python repository, downloaded ~5,000 times, contained code that installed cryptomining software
Context & Ripple Effects
This is an early entry in a pattern the coverage keeps confirming: public package registries as a malware delivery channel. Two years earlier, researchers found 725 Ruby libraries on RubyGems carrying clipboard-stealing malware aimed at bitcoin payments, and months before this report a researcher documented GitHub Actions being abused to run cryptominers on GitHub's own infrastructure — same monetization, different entry point.
What changed here is the vector: not a compromised legitimate package but outright counterfeits uploaded to PyPI itself and pulled down roughly 5,000 times. The later arc shows the problem compounding — ctx and phpass were later compromised to steal AWS credentials, and GitGuardian found nearly 3,000 of 450,000 PyPI projects exposing credentials in code — meaning a malicious package on the registry sits on top of an existing secrets-hygiene problem that multiplies the blast radius.
First-order effects
- Developers who installed the counterfeit packages got cryptomining software running on their machines, burning compute they paid for — and PyPI's name-trust assumption is what let ~5,000 downloads happen before detection.
Second-order effects
- Registry operators face pressure to add package-verification and scanning layers, while the same attackers pivot across ecosystems — the RubyGems, GitHub Actions, and later ctx/phpass incidents show one technique retooled per platform rather than one-off attacks.
Third-order effects
- If counterfeits and credential leaks keep co-occurring on PyPI, dependency resolution shifts from trusting a registry name to requiring provenance and integrity checks — and the 2026 Mistral PyPI compromise suggests the pattern extends to AI-model packages, not just libraries.
The trend: Open-source package registries are consolidating as attackers' preferred supply-chain entry point, with each incident pushing ecosystems toward provenance verification over name-based trust.