Your supply chain isn't compromised. It's working.

At least 70 of Microsoft’s own open source projects on GitHub now return the same line when you try to load them: “Access to this repository has been disabled by GitHub Staff due to a violation of GitHub’s terms of service.” Microsoft pulled them after hackers breached the projects and injected password-stealing malware into the code.

Many of the affected repos relate to Azure and to tools developers use alongside AI coding apps like Claude Code, Gemini’s command line interface, and VS Code. The malware ran when a user opened one of the compromised tools inside their AI coding app, lifting passwords and other credentials off the machine. Security firm Cloudsmith and the community malware-analysis site OpenSourceMalware were among the first to flag it. How many people pulled the affected code is not known.

Microsoft confirmed it removed the repositories, a move first reported by 404 Media. Spokesperson Ben Hope said the company “temporarily removed some repositories as we investigated potential malicious content,” that “some of these repos have been restored after review, while others may remain offline while work continues,” and that Microsoft “notified a small number of customers who may have pulled down content from the affected repositories.” The company did not say how many customers were affected.

Nothing in the pipeline broke. The code sat in a first-party Microsoft repo on a Microsoft-owned host, and it flowed to developers through the same pull they run every day. That is what makes it a supply chain attack rather than a smash-and-grab: the attacker’s code inherited all the trust Microsoft’s name carries, and the distribution path delivered it exactly as designed. The source was legitimate, the download succeeded, the install ran. The supply chain did its job on behalf of whoever got into the repo.

Getting into the repo is the part that stands out. Solo maintainers of open source projects get targeted often, sometimes through months of trust-building before a malicious commit lands. A vendor the size of Microsoft, with the resources to defend its own GitHub org, is a rarer catch.

This is the second known compromise of Microsoft’s open source projects in a few weeks, per Ars Technica. In mid-May, researchers reported that Durable Task, a Microsoft project that helps developers build apps, was hacked. OpenSourceMalware describes the current incident as a “re-compromise” of Durable Task, which points to one of two things: either the first cleanup missed something and the attacker still had a foothold, or this is a separate breach into the same project.

For anyone who pulls from these repos, the practical read is that a first-party source is not a clean bill of health. If you or your build systems pulled Microsoft open source content recently, especially Azure tooling or anything that runs inside an AI coding app, treat the credentials on those machines as exposed and rotate them. Rotate cloud keys first; the reason a project like this is worth breaching is that the developers using it tend to hold access to cloud systems and customer data. Credential theft that triggers when a developer opens the compromised tool means the live workstation, with whatever it can reach, is the prize.

Microsoft says it will contact affected customers directly through its established support channels if further action is needed. Watching for that notice is reasonable; waiting for it before rotating is not, given the company has not published a list of affected projects or a count of impacted users.

The re-compromise detail is the one to hold onto. “Restored after review” is Microsoft’s language for repos it has put back, and the Durable Task sequence shows that a review which clears a project once does not guarantee the attacker is gone. A repo coming back online is a weaker signal than a repo being clean.

See also: NordVPN for tunneled traffic when operating outside controlled networks.

---

#ad Contains an affiliate link.