Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
in-toto is a framework for securing a software supply chain end to end: a signed layout defines the steps a release must pass through, and signed link metadata proves each one actually happened.

What Is in-toto?
in-toto answers a question that signing a finished artefact cannot: was this software produced by the process we said it would be produced by? A signature on a release binary proves who published it. in-toto proves that the source was reviewed, the tests ran, the build used the expected inputs, and that the thing published is the thing the build produced, with each step attested by whoever performed it.
It came out of the Secure Systems Lab at New York University, the same group behind The Update Framework, and is now developed under the Cloud Native Computing Foundation.
Three pieces make up the model.
The layout is a signed document written by the project owner. It names each step in the supply chain, who is authorised to perform it, which commands are expected, and how the artefacts flow between steps. It also defines inspections, which are checks run at verification time against the final product.
The link metadata is produced by each functionary as they do their work. A link records the materials consumed, the products created, the command executed, and its return code, signed with the functionary’s key. Wrapping an existing build command is deliberately trivial:
in-toto-run --step-name build --key builder.key \
--materials . --products dist/ -- make release
Verification combines the two. The verifier checks that every step in the layout has link metadata signed by an authorised key, and then enforces the artefact rules that connect steps: MATCH requires that the products of one step are exactly the materials of the next, while CREATE, DELETE, MODIFY, ALLOW, DISALLOW, and REQUIRE describe what each step may legitimately do to the files it touches.
The MATCH rule carries most of the weight. It is what catches a file that appeared between testing and packaging, which is exactly the shape of the SolarWinds Orion compromise, where malicious code was injected during the build rather than committed to source.
The part of in-toto most people now encounter is not the full layout model but the attestation format it defines. An in-toto attestation is a signed JSON statement with three elements: a subject identifying the artefact by cryptographic digest, a predicate type naming what kind of claim is being made, and a predicate holding the claim itself.
That envelope is now the common carrier for supply chain metadata. SLSA provenance is an in-toto predicate. So are SBOM attestations, vulnerability scan results, and test outcomes attached to a container image with Cosign. Teams routinely produce and consume in-toto attestations without ever writing a layout.
The strength is coverage. Most supply chain controls secure one link: a signature on the artefact, a scan of the dependencies, a protected branch. in-toto describes the whole chain and makes a gap in it visible, because a missing or unsigned link fails verification just as loudly as a tampered one.
The cost is that someone must write and maintain the layout, distribute functionary keys, and keep the document accurate as the pipeline changes — and pipelines change constantly. A layout that no longer matches reality gets bypassed, and a bypassed control is worse than none because it still appears in the compliance evidence. This is why adoption has concentrated on the attestation format, which delivers a large share of the benefit for a fraction of the operational burden. Datadog was an early production adopter, applying it to the pipeline that builds its agent integrations.
For most organisations the sensible sequence is: emit SLSA provenance from your build platform as an in-toto attestation, attach an SBOM attestation alongside it, verify both at deployment with an admission policy, and only then consider a full layout for the handful of releases where end-to-end process assurance genuinely justifies the effort — signed firmware, regulated products, or anything where a build-time compromise would be unrecoverable.
At Privia Security, we review build and release pipelines against this kind of threat model as part of our cybersecurity consultancy work.
A release signature proves that a named identity published a particular file. in-toto proves that the file came out of the process you defined: the expected steps ran, in order, performed by authorised parties, with the output of each step feeding the next unchanged. It detects tampering between steps, which a signature applied at the end cannot, because that signature is applied to whatever the build produced.
You are already using part of it. SLSA provenance is published as an in-toto attestation, so the format and signing envelope come from this framework. The full layout model goes further by describing every step in the chain and the artefact flow between them. SLSA is the easier starting point, and the layout model is worth the effort only where end-to-end process assurance is a real requirement.
A statement in the layout about what a step may do to files. MATCH requires that artefacts produced by one step appear unchanged as inputs to another, which is the rule that detects injection between steps. CREATE, DELETE, and MODIFY declare legitimate changes, while ALLOW and DISALLOW handle everything else. Together they turn an informal pipeline description into something a verifier can enforce.
Datadog was the first large production adopter, applying it to the pipeline that builds its agent integrations. Beyond full deployments, the attestation format is used very widely: SLSA provenance from GitHub Actions, SBOM and scan attestations attached with Cosign, and policy engines that verify them at admission all speak in-toto. Most teams meet the format long before they meet the layout model.
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.