Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
SLSA — Supply-chain Levels for Software Artifacts — is a graded set of requirements for build integrity, centred on provenance: a signed record of how an artefact was produced and from what.

What Is the SLSA Framework?
SLSA, pronounced “salsa”, stands for Supply-chain Levels for Software Artifacts. It is a specification that grades how trustworthy a build process is, and provides a common vocabulary for saying so. Google published the first version in 2021, drawing on its internal Binary Authorization for Borg, and it is now developed under the Open Source Security Foundation. Version 1.0 arrived in April 2023 and reorganised the specification into tracks.
Everything in the Build track rests on one artefact: provenance. A signed, machine-readable statement describing what was built, from which source at which commit, by which build platform, with which parameters.
| Level | Requirement | What it stops |
|---|---|---|
| Build L1 | Provenance exists and is published | Nothing by itself; it makes the build process visible and inspectable |
| Build L2 | Provenance is signed and generated by a hosted build service | Tampering with the artefact or its provenance after the build |
| Build L3 | The build platform is hardened; provenance is unforgeable and builds are isolated | A compromised build job forging provenance or influencing another build |
The jump that matters is from L2 to L3. At L2 the build service signs the provenance, so you know which platform produced the artefact. At L3 the signing key is unavailable to the build job itself, and builds cannot affect one another — meaning even a malicious build definition cannot produce provenance claiming it came from somewhere else.
Version 1.0 deliberately narrowed the scope. The earlier draft folded source review, two-person approval, and reproducible builds into its top level; those concerns moved to separate tracks, with a Source track and a Build Environment track in development. The result is a Build track that organisations can actually achieve.
Provenance is published as an in-toto attestation: a statement naming the artefact by digest, a predicate type identifying it as SLSA provenance, and a predicate describing the build. The build definition records the source repository and commit, the entry point, and the parameters; the run details record the builder identity and timestamps.
A verifier can then enforce meaningful policy: this image must have been built by our organisation’s reusable workflow, from the main branch of this specific repository, from a commit that exists in that repository. That is a far stronger statement than a signature alone, which only says somebody signed it.
For most teams this is less work than it sounds, because the build platforms have done it. GitHub Actions can generate signed provenance through the slsa-github-generator reusable workflows or the native artefact attestation feature, reaching Build L3 for many project types. npm provenance, generally available since April 2023, publishes SLSA provenance for packages built in a public workflow. Google Cloud Build, GitLab, and several commercial platforms produce conformant provenance too.
A sensible sequence:
That last point is where most adoptions stall. Generating provenance nobody checks is paperwork. Verification belongs in an admission controller, a deployment gate, or the package installer, and it must fail closed.
It makes no claim about code quality. An artefact at Build L3 can be riddled with vulnerabilities, depend on an abandoned library, or do exactly what a malicious contributor intended — provenance describes how it was built, not what it does. Dependencies are a separate problem: SLSA grades your build, not the builds of the packages you consume, which is why it sits alongside dependency scanning and SBOMs rather than replacing them.
The threat model it addresses is specific and real: the SolarWinds Orion build compromise, the Codecov bash uploader incident, and the recurring pattern of attackers targeting build infrastructure because it is trusted and rarely watched.
Supply-chain Levels for Software Artifacts, pronounced salsa. It is a specification maintained under the Open Source Security Foundation that defines graded requirements for build integrity, with each level demanding stronger guarantees about how an artefact was produced and how trustworthy the record of that production is. The name is usually written in capitals and the levels are cited as Build L1 to L3.
An SBOM lists what is inside an artefact: the components, versions, and dependencies. SLSA provenance describes how the artefact came to exist: which source, which builder, which parameters. One is an inventory, the other a production record. They answer different questions during an incident and are typically published together as signed attestations attached to the same artefact.
In most cases by using a build platform that already provides it rather than by building infrastructure yourself. GitHub Actions with the SLSA generator workflows or native artefact attestations reaches Build L3 for many project types, as do several hosted build services. The requirement is that provenance is signed by the platform with a key the build job cannot access, and that builds are isolated from one another.
No. Provenance only protects you when something verifies it and refuses artefacts that fail. That means a policy at deployment or admission checking that the artefact was built by the expected workflow, from the expected repository and branch, and rejecting anything else. Teams frequently generate attestations for months before anyone enforces them, which provides a record but no protection.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.