News for hackers. Stuff that matters.
Re:Post Mortem: axios NPM supply chain compromise by redoh on Friday April 03, 2026 @11:22PM (#47635352)
The part that keeps bugging me about incidents like this is how the clearest signal was already there and nobody acted on it. Every legitimate v1 release had OIDC provenance attestations, and the malicious one just... didn't. That's a binary check you could automate in any CI pipeline in about ten minutes, but almost nobody does.

I've started adding a provenance verification step to our deploy pipelines after the event-stream incident years ago, and it's caught weird stuff twice now, both times just maintainers accidentally publishing from their local machine instead of CI, not actual attacks, but the point is the same mechanism would catch this. The real problem isn't that we lack signing infrastructure or better NPM policies. It's that the ecosystem has trained everyone to treat dependency updates as a solved problem the moment you have a lockfile. Lockfiles protect you from silent republishing of existing versions, but they do nothing when the attack comes as a "legitimate" new patch release. Diffing your lockfile for unexpected new transitive deps on every deploy is table stakes, and most teams I've worked with still don't do it.

Re:Post Mortem: axios NPM supply chain compromise by cyberax on Friday April 03, 2026 @11:48PM (#47635506)
Another point: do NOT use the "~" or "^" versions for automatic updates. Just lock everything tight in your package files. Then have an alert on the lockfile changes.
Re:Post Mortem: axios NPM supply chain compromise by stingraycharles on Saturday April 04, 2026 @12:31AM (#47635772)
Please no AI posts on HN.
Re:Post Mortem: axios NPM supply chain compromise by panstromek on Saturday April 04, 2026 @03:05AM (#47636628)
Well, the hack didn't survive more than 2-3 hours if I'm not mistaken. I don't think that counts as "nobody acted on it."
Re:Post Mortem: axios NPM supply chain compromise by panstromek on Saturday April 04, 2026 @03:10AM (#47636667)
Actually, from the OP, the timeline is:

> March 31, 00:21 UTC: [email protected] published with [email protected] injected

> March 31, around 01:00 UTC: [email protected] published with the same payload

> March 31, around 01:00 UTC: first external detections

> March 31, around 01:00 UTC: community members file issues reporting the compromise. The attacker deletes them using the compromised account.

So it was found out almost immediately.