"We have checked our own environments thoroughly and found no traces of compromise. We suspect this may be part of the broader GitHub infrastructure breach carried out by the TeamPCP hacking group in May 2026: https://techcrunch.com/2026/05/20/github-says-hackers-stole-..."

Greater HN collective, please help me metaphorically double-click on this. I've poked around a bit but didn't find out much more than the given link. What are we concerned about the hack possibly having accomplished?

Because stealing repos is bad enough... but are we saying it's possible that commits can now magically appear in repos from hackers? I don't want to raise any alarms if I'm misreading this or if we're early in the news cycle, but if that's possible, I and a lot of other people reading this need to have some immediate conversations with a lot of people. So... is that what this is saying? Or am I misreading it? I sure hope so.

I was impacted. found weird spam repos that later were deployed on cloudflare redirecting my domains.

meanwhile the gitea running on my metalbox for nearly a decade has seen no compromise and 100% uptime when cloudflare has gone down repeatedly

im rethinking the whole "go where crowd is" , while great from evolutionary point of view, its the complete opposite. Where the crowd gathers online is the most dangerous place.

it's the same with linux viruses. they were always a possibility, but because linux is not popular, they were never an issue.

Linux is absolutely popular for servers. If you put a WordPress installation on the IPv4 address space, or any other kind of PHP you usually find a webshell has appeared after just a few minutes.

true, i get these attempts on my server daily. but here too you got less popular alternatives, so the same principle applies.

Don't use github actions. Don't use toolchains that auto execute stuff.

Simple as that, because that's the attack surface.

https://cookie.engineer/weblog/articles/malware-insights-git...

I wrote that article December 2024. Still ongoing, Microsoft. Best enterprise security practices, I suppose shrugs ...

So in summary:

* GitHub's backwards priorities end up causing a hack on their systems.

* Hackers use their newly gained powers to compromise other people's repos.

* GitHub dectects compromised repo, and suspends the account of its maintainer, so they cannot warn nor act against it to protect or at least warn their community of users.

"I cause a fire, and later ban you for getting burned."

No wonder people are leaving.

Where are they going? If its not self hosted I don't see it not ending up like github.

Why do people not like gitlab? I’ve always found it a better experience than github

Gitlab's UI changes every now and then, for seemingly no reason. The UI is very full of stuff (hard to find your way around), and very slow. Notably in the past months, they've changed the issues/tickets board into a "work items" board which feels infinitely slower to load, has such a vague meaning that nobody can find it (especially when translated), and brings exactly 0 use to anyone i know. They just seem to be doing that with every feature and every part of the interface.

On the server side, gitlab was always very hard to selfhost with many moving parts, many requirements, and using much resources. gitlab-runner is not very explicit about things when you're not in the happy path (why is it not picking up jobs?).

I'm not even a minimalist. I've been running gitea/forgejo for the past 8 years or so and it's been a miracle in comparison: lightweight server, easy setup/upgrades, and super simpler UI/UX that everybody understands on the first try. Forgejo (gitea community fork) learns from everything that Github historically made good (UX) without any enshitiffication in sight (developed by a non-profit). I highly recommend it.

If you're leaving based on security failures, Gitlab is not the place to go.

same. so much more intuitive

codeberg

I had a repo with more than a dozen forks banned on GitHub for some unclear TOS violations. Ticket has been sitting for a week plus now, asking for clarification and guidance.

So, it lives in codeberg now. https://codeberg.org/nelsonjchen/op-replay-clipper

this just looks like a reskinned gitea

It's a running a fork (codeberg specific) of a fork of gitea called forgejo (https://codeberg.org/forgejo/forgejo) so it's not surprising. The people behind it were a bit miffed at Gitea doing some questionable commercial endeavors in their view and also not dog-fooding Gitea for Gitea.

huh i did not know that . thanks for forgejo guess im moving

There exist competent operations people and competent developers.

Seems like it's similar to the attack reported in this other HN post: https://news.ycombinator.com/item?id=48409869

We're working on an antiworm. One of our customers got affected.

Our tool already discovers infected repositories and mitigates/removes the implants from the filesystem.

Please revoke/rotate all your tokens and passwords that were used in the infected repositories, the worm is pretty sophisticated.

https://github.com/Team-Rockstars-Security/antimiasma

Looking at the setup.js it seems to be an infostealer which posts the found details to a newly created github repo (on the victims account) or a command and control server. As far as I can tell it looks for github secrets and kubernetes cluster secrets.