AppSecNews
SCA Free Established

GitHub Dependabot

by GitHub

GitHub's built-in dependency scanner and update bot, matching your dependency graph against the GitHub Advisory Database and opening pull requests that bump vulnerable packages.

Visit docs.github.com (leaves AppSecNews, opens in a new tab) Leaves AppSecNews for the vendor's own site.

No endorsements yet

Run GitHub Dependabot in production? A named recommendation helps the next team shortlisting it.

Recommend this tool

Endorsers verify their identity through LinkedIn. Titles and companies are self declared, shown as they were when each person signed, and reviewed by an editor before anything is published. Endorsements are never paid for.

What we still need to verify : 2 points in this profile are not yet confirmed against vendor documentation.
  • Current ecosystem support list, which expands over time: verify against GitHub documentation
  • Feature availability differences between public repositories, private repositories and Advanced Security: confirm

Treat these points as unconfirmed. They are open items in the catalog's verification queue, and this note stays until each is checked against the vendor's documentation.

What it does

Dependabot has two parts that are easy to confuse. The first is alerting: GitHub builds a dependency graph for a repository by parsing its manifest and lock files, then matches that graph against the GitHub Advisory Database, a curated feed drawn from CVE data, ecosystem security advisories and researcher reports, with GitHub's own review applied to affected version ranges. When a match appears, an alert is raised against the repository, including for transitive dependencies the graph resolved.

The second part is updating. Configured through a file in the repository, Dependabot opens pull requests that bump packages, either on a schedule for version updates or in response to an advisory match for security updates. Each pull request includes the changelog, release notes and commit range for the bump, plus a compatibility signal derived from whether other public repositories' test suites passed on the same upgrade. Grouping rules let related packages move together rather than arriving separately.

Where it fits

It lives entirely inside GitHub, configured by whoever owns the repository, usually developers rather than a security team. Alerts surface in the repository security tab and updates arrive as ordinary pull requests going through your normal review and CI. That is its real strength: remediation shows up as a code change where work already happens. You need dependencies declared in supported manifest formats and a test suite good enough that developers trust a green build on a version bump.

Strengths

  • Zero setup cost for alerting on GitHub-hosted code, which removes the usual excuse for having no composition scanning.
  • Curated affected-version ranges in the GitHub Advisory Database beat raw NVD matching for ecosystem packages.
  • Remediation arrives as a reviewable pull request rather than a ticket, closing the loop between finding and fix.

Limitations

  • No reachability analysis whatsoever. If a vulnerable package is present, you get an alert, regardless of whether the affected function is ever called.
  • Pull request volume on a large or fast-moving dependency tree is a genuine burden, and teams routinely respond by ignoring the bot, which is worse than not running it.
  • Transitive vulnerabilities often cannot be fixed by Dependabot alone, because the direct dependency's maintainer has to update first, and it is GitHub-only, so code hosted elsewhere gets nothing.

Who it suits

The correct default for any team on GitHub. Turn it on, tune the grouping, and you have a baseline. It is not sufficient alone for an organization needing prioritization by exploitability or reachability, centralized policy across many repositories, license compliance, or audit-grade evidence. Those teams should treat it as a complement to a dedicated composition analysis platform, not a replacement.

Used GitHub Dependabot? Recommend it under your own name and title.

Recommend this tool