How we took malware advisories beyond npm

GitHub’s Dependabot previously flagged malware only in npm packages, leaving other ecosystems vulnerable to credential-stealing attacks at installation. The company now incorporates OpenSSF’s malicious-packages repository into its Advisory Database, which aggregates over 15,000 reports in OSV format from community and automated sources. This expansion covers eight ecosystems: npm, PyPI, Maven, RubyGems, NuGet, Go, crates.io, and PHP Composer, allowing Dependabot to alert developers to malicious dependencies across a broader range of package managers.

The new pipeline reuses existing importer patterns to process OSV records, validating fields and schema compliance before normalization. Valid records are converted into feed entries with preserved upstream data, while invalid or retracted reports are rejected. The system handles mismatches in ecosystem naming conventions, version formats, and deduplication to avoid redundant advisories. For example, GitHub’s own npm malware reports are filtered out to prevent circular imports, ensuring only new, externally sourced threats are flagged.

Security reviews addressed concerns about upstream data reliability, opting for automated publishing of malware advisories due to the urgency of credential theft risks. Unlike vulnerability advisories, malware reports require rapid dissemination, as delays could allow attackers to exploit compromised packages. The pipeline includes three layers of protection to mitigate false positives or compromised data feeds, prioritizing speed over human review for these binary-flagged threats.

Dependabot malware alerts are opt-in and can be enabled at the repository, organization, or enterprise level. Once activated, Dependabot matches dependencies against the Advisory Database, including historical malware advisories. Additionally, GitHub now alerts users to vulnerable GitHub Actions workflows, further expanding Dependabot’s coverage to include CI/CD security vulnerabilities.