Tool profilesBased on official documentation, not tested by us

OSV-Scanner: check the libraries your project uses for known vulnerabilities, free and open source

OSV-Scanner is a free Google tool that checks your project's dependencies against the open OSV.dev vulnerability database. Based on the official README and docs, not tested by us: what it does, what it sends out, its limits and what we could not confirm.

Type

Dependency vulnerability scanner (command line)

License and cost

Apache-2.0 licence, free

Reviewed

3 Nov 2026

Version

Not shown on the pages we read (the README calls the current line a V2 beta)

Best for

Teams that build with npm, pip, Go, Rust and similar packages and want a free check of known flaws in their libraries.

Watch out for

A clean result only means no known flaw matched. The fix command is experimental and can run scripts.

← All tool profilesNo. 6 / 11

In one sentence

OSV-Scanner reads the list of libraries your project depends on and checks each one against OSV.dev, an open database of known vulnerabilitiesVulnerability A weakness in software, a setting or a process that an attacker could use to get in or do harm. Tap again to open the glossary. If a library you use has a known flaw, it tells you. It is a supply-chain check: the risk is not in your code but in the code you pulled in.

Who it is for

Founders, indie makers and small teams that build with packages from npm, pip, Go, Rust and similar, which is most projects made with AI help, and want a free check before shipping. It is a command-line tool.

What the documentation says

Everything below comes from the README and the two documentation pages we read on 2 November 2026.

  • Maintainer and licence: it lives in Google's GitHub organisation, and the README calls it an officially supported frontend to OSV.dev. The licence is Apache-2.0.
  • What it scans: source directories (recursively), container images, vendored C/C++ code, and it can also check licences against deps.dev data.
  • Languages and package managers: the README lists C/C++, Dart, Elixir, Go, Java, JavaScript, PHP, Python, R, Ruby and Rust, with managers such as npm, pip, yarn, Maven, Go modules, Cargo, gem, Composer and NuGet. It says to see the docs for the full list.
  • Commands: osv-scanner scan -r <folder> scans a folder recursively, scan image <image>:<tag> scans a container image, and an experimental fix command suggests upgrades.
  • Output: the --format option sets the format (the examples use JSON and a "vertical" text view), --output-file saves it, and --serve shows an HTML report on port 8000.
  • Offline use: matching against a local copy of the database is possible; the usage page names the flag --offline-vulnerabilities.
  • Configuration: --config takes a TOML file that applies to everything scanned, and package overrides with an ignore action are mentioned.
  • What it sends out: according to the README, queries go to the OSV.dev and deps.dev APIs, and optionally to Maven Central, npm and PyPI. Only package names, versions, ecosystems and file hashes are sent, not your source code.
  • Call analysis: the README says it can cut false alarms by checking whether the vulnerable function is actually used.

The good

  • Free, Apache-2.0, backed by an open and community-fed database.
  • Covers many languages with one tool, and also container images.
  • Per the README, it does not upload your source code.
  • It can work offline.

The limits it admits

  • The README describes the current line as a beta.
  • The fix command is experimental, and the docs warn that it can run your package manager's scripts and contact external registries. Use it only on projects you trust.
  • It can only find vulnerabilities that are in the database. A flaw nobody has reported is invisible to it.

When not to use it

  • If you believe a clean result means your project is safe. It means no known flaw matched, not that none exists.
  • As a replacement for updating. Finding a vulnerable library does not fix it; you still have to upgrade or replace it.
  • If your project uses a language or file type outside the lists in the docs, which we could not check in full.

What it costs you

Time more than money. We have no measured setup time, because we did not run it. The ongoing work is reading the findings, deciding which ones matter for how you use the library, upgrading, and re-scanning. Running it in your pipeline is possible, but we did not read the pipeline documentation.

Typical setup mistakes

  • Scanning once and forgetting. New vulnerabilities are published every day, so a project that was clean last month may not be today. Re-scan on a schedule.
  • Scanning without the lockfile. The scan reads the package files it finds. Without the lockfile, the exact versions you really run may not be seen. This is our reading of how lockfile scanning works, not a quote from the docs.
  • Ignoring everything that looks minor. The docs offer ways to ignore a package, and a wide ignore can hide a real problem.
  • Running fix on code you do not trust. The docs warn it can run scripts.
  • Not checking what a finding means. A vulnerable function you never call is not the same as one you use. That is what call analysis is for.

How to check that it works

Official documentation and our approach, not something we ran: in a throwaway test project, scan a folder with an old dependency you know has a published vulnerability, and expect the tool to list it. Then upgrade the dependency and scan again. Use a test project, not a real one.

Where to start

We did not test the tool, so we do not offer a tutorial. Follow the official documentation: httpsHTTPS The secure version of the web address protocol: it encrypts what travels between a visitor and your site and checks that the site is the real one. Tap again to open the glossary://google.github.io/osv-scanner/

What we could not confirm

We did not download, install or run OSV-Scanner, so everything above is the maintainers' own description. The pages we read showed no release number or date, and the docs home page did not cover installation, supported files or exit codes. We did not read the pages on exit codes, output formats, the configuration file syntax, or pipeline use, so we cannot say what its exit codes mean or how it behaves in a pipeline. We do not know how accurate it is, how many false alarms it gives, or how long it takes on a large project. We have not compared it with other dependency scanners.

Glossary: vulnerability, supply chainSupply chain attack An attack that reaches you through something you already trust and install, such as a code package, a tool or a provider. Tap again to open the glossary, CVECVE A public ID number, like CVE-2026-12345, given to one specific security flaw so everyone can talk about the same bug. Tap again to open the glossary, patchPatch (security update) A fix released by the maker of a piece of software to close a flaw or a bug. Tap again to open the glossary. See also the news section for real supply-chain cases.

More free tools

Other profiles in this section.

Sources

Follow us

New profiles and security news: YouTube (@safet2ship) and Instagram (@safet2ship).

YouTube channel · Instagram

Get the free runbooks

One-page security checklists for founders who build with AI. You get a confirmation email first, then the runbooks.

Send me the runbooks