Open Source Security
PyPI Trusted Publishing Common Pitfalls
PyPI trusted publishing removed a whole class of token leaks, but teams keep tripping over the same half-dozen configuration mistakes. Here is what to watch for.
Jan 28, 20267 min read
Deep dives, practical guides, and incident analyses from engineers who build Safeguard. No fluff, no vendor FUD — just what you need to ship secure software.
PyPI trusted publishing removed a whole class of token leaks, but teams keep tripping over the same half-dozen configuration mistakes. Here is what to watch for.
The Safeguard Research team analyzed first-quarter 2026 malicious package telemetry across npm, PyPI, RubyGems, and crates.io. Here is what the data shows.
How a GitHub Actions cache poisoning attack pushed a crypto miner into Ultralytics 8.3.41 on PyPI, and what engineering teams should actually change.
Three PyPI packages impersonating Alibaba's AI Labs SDK exfiltrated .gitconfig data from developer machines in a regionally targeted 2025 espionage campaign.
Typosquatting on PyPI reached industrial scale in 2024, with attackers using automated tooling to register thousands of malicious package names targeting common misspellings of popular libraries.
Q1 2024 brought typosquats, stealer campaigns, and a week-long new-user freeze on PyPI. Here is what the attacks looked like and how to defend.
Build a pre-install guard that catches typosquatted npm, PyPI, and RubyGems dependencies using Levenshtein distance, download-count heuristics, and registry APIs.
A practical pre-install verification workflow for PyPI packages covering sigstore attestations, maintainer checks, and sdist auditing.
Every pip install is a trust decision. The Python Packaging Authority has spent years hardening the ecosystem, but the attack surface remains vast and the threat actors are persistent.
Weekly insights on software supply chain security, delivered to your inbox.