Looking back: OpenSSF data expands GitHub's malware alerts

At the end of July, GitHub announced that it would add data from the OpenSSF (Open Source Security Foundation) malicious-packages repository to its Advisory Database. Since then, the Microsoft subsidiary has issued security alerts for eight package ecosystems: crates.io, Go, Maven, NuGet, npm, PHP Composer, PyPI and RubyGems. Before that, it only supported npm packages. Now the OpenSSF is taking stock: More than 238,000 reports of malicious packages have made their way into the OpenSSF repository.

Launched in 2023 with more than 15,000 entries, the OpenSSF repository is a public, community-maintained database that uses the Open Source Vulnerability (OSV) format. GitHub was one of the three original data sources at launch. More than 28,000 malware reports now come from GitHub, the OpenSSF notes in its case study blog post . It also names the other contributors to malicious-packages: the OpenSSF’s own Package Analysis project, as well as Amazon Inspector, Checkmarx, Google and ReversingLabs.

Expanding its own detection methods to cover eight ecosystems would have taken years, Ankit Kumar Honey, senior engineering manager at GitHub and head of the Dependabot team in Supply Chain Security, estimated, according to GitHub’s August 2026 announcement . So GitHub used a source that was already available. “Meanwhile, OpenSSF has already solved the aggregation problem for everybody”, Honey said.

GitHub developed its own import tool to bring in the data. It checks reports, standardizes records and processes withdrawn advisories. At the same time, it prevents GitHub’s own contributions from being imported twice. Safeguards include limits on import batches, traceability back to upstream commits and the ability to roll back entire batches.

The opt-in malware alerts are available through Dependabot, GitHub’s service for monitoring a project’s dependencies. They’re available in every repository where Dependabot alerts are enabled, but must also be turned on there separately. Organizations and enterprises can roll them out in bulk through Security Configurations. Anyone who has already enabled malware alerts gets the expanded coverage automatically.

You can also malware reports in the GitHub Advisory Database using the type:malware filter. False positives can occur when an internal package has the same name as a malicious public package. You can reduce them with Dependabot auto-triage rules . However, the built-in rule “Dismiss package malware alerts”, which is disabled by default, also dismisses alerts for dependencies that are actually malicious in every version.

This article was originally published in German . It was translated with technical assistance and editorially reviewed before publication.