Skip to main content

What are you looking for?

Explore our services and discover how we can help you achieve your goals

Client Workspace

Software supply chain security: SBOMs, signed builds and dependency policy in adoption order

Software supply chain security works best adopted in order: first know every dependency and pin it, then produce an SBOM with every build, then write a dependency policy, then sign artifacts with build provenance, and last refuse to deploy anything unverified. Netbase sequences it this way because each step relies on the one before.

Book a solution review See the related solution

By Netbase Delivery and Quality Desk · Reviewed by David (CEO) · Updated · 11 min read

star

This guide is for CTOs, engineering managers and product owners of business applications whose customers, auditors or regulators have started asking for SBOMs and signed releases. If you want the scanning and review done with your team, see application security assurance.

In this guide

What does software supply chain security cover in a business application?

Software supply chain security covers everything that enters a release without being written by your team: open-source packages and their transitive dependencies, base container images, build tools and plugins, the CI system itself, and the registry the final artifact is published to. A flaw or a malicious change in any of these ships with your code.

The OWASP Top 10:2025 made this a category of its own, A03 Software Supply Chain Failures, and describes it as breakdowns or compromises in building, distributing or updating software. That is wider than "vulnerable libraries": it includes a compromised maintainer account, a tampered build step and an update pushed to every system at once.

For a mid-market team, three controls answer most of the questions buyers now ask. An SBOM says what is inside a release. A dependency policy says what is allowed in. Signing with build provenance says who built the artifact, from which source, and that nobody changed it afterwards.

Why does the order of adoption matter?

The order matters because each control consumes the output of the one before it. A signature over an artifact whose contents nobody can list proves only that an unknown thing was not altered. A dependency policy without an inventory cannot be enforced, and an SBOM generated from unpinned dependencies describes a build nobody can reproduce.

Teams that start with the most visible control, usually signing, often end up with signed images that no deploy step checks. Starting from the inventory makes every later step cheaper, because the data it needs already exists.

Which control proves what?

Read the "does not prove" column before promising a customer anything.

Control What it proves What it does not prove Effort on the team
SBOM per build Which components and versions are in this exact release That the components are safe, or that the list is complete if it was not generated from the build Low once automated in CI; someone must store and query the files
Dependency admission policy Which new or upgraded packages are allowed in, and who approved an exception That an allowed package has no undisclosed flaw or malicious release Medium: rules, a review step and an exception register
Signed artifacts with provenance Who built the artifact, from which source and build, and that it was not changed afterwards That the source or the build platform itself was trustworthy Medium to high: signing in CI plus verification before deploy

The CISA 2026 Minimum Elements for a Software Bill of Materials, published with the NSA, the FBI and international partners, calls an SBOM an ingredients list and a building block of supply chain risk management; it replaces the NTIA minimum elements of 2021. Its value comes from querying it when a new vulnerability is disclosed.

How do you adopt the controls in order?

Each stage has an entry condition and a test for "done". Do not start the next stage until the current one passes its test.

  1. Inventory and lock files

    Entry: the application builds in CI from version control. Commit lock files for every package manager, pin base images by digest, and list the build tools and CI plugins in use. Done when two builds of the same commit resolve identical dependency versions.

  2. An SBOM for every build

    Entry: stage 1 passes. Generate an SBOM in a standard format such as CycloneDX or SPDX as a build step, store it next to the artifact, and load it into a tracker such as OWASP Dependency-Track that matches components against new disclosures. Done when you can answer "which releases contain package X version Y" in minutes, not days.

  3. A dependency admission policy

    Entry: SBOMs are queried, not only stored. Write the rules below, add a review step for any new direct dependency, and block the merge for a new package with a known critical issue. Done when every exception has a named owner and an expiry date.

  4. Signed artifacts with build provenance

    Entry: builds run on a hosted CI platform, not a developer laptop. Sign the artifact and its SBOM in CI with a managed signing key and record provenance. The SLSA Build track starts at Build L1, where provenance exists, and reaches Build L2 when a hosted platform signs it, which prevents tampering after the build. Keyless signing with Sigstore removes long-lived signing keys. Done when every production artifact has a signature and provenance record.

  5. Verify before deploy

    Entry: stage 4 covers every production artifact. Make the deploy step refuse an artifact whose signature, builder identity or source repository does not match, and document a break-glass path with named approval. Done when an unsigned test artifact is rejected in staging.

The dependency policy template

Keep it to one page of testable clauses that reviewers can apply in a pull request.

  • Source

    Rule to write
    Packages only from the official registry or an internal mirror over a secure connection
    How it is checked
    Registry configuration in CI; review of any new source
  • Admission

    Rule to write
    A new direct dependency needs a stated purpose, an active maintainer and an acceptable licence
    How it is checked
    Pull request template plus reviewer sign-off
  • Pinning

    Rule to write
    Exact versions through lock files; base images by digest
    How it is checked
    Build fails on an unpinned reference
  • Known issues

    Rule to write
    No new dependency with a known critical vulnerability
    How it is checked
    Dependency check on every merge
  • Updates

    Rule to write
    Security updates follow the patching policy; other updates on a regular cadence
    How it is checked
    Update bot pull requests and the patch register
  • Exceptions

    Rule to write
    Named owner, reason, compensating control and expiry date
    How it is checked
    Exception register reviewed at each release
  • Removal

    Rule to write
    Unused dependencies are removed when found
    How it is checked
    SBOM diff between releases

The OWASP A03 guidance points the same way: official sources over secure links, signed packages, tracked transitive dependencies, and separation of duties so no single person can write code and promote it to production without oversight.

When should you not adopt a stage yet?

A stage is the wrong next step when the one beneath it is missing.

  • No reliable CI. If releases are built by hand, fix the pipeline first; signing a manual build only signs an unknown process. A DevOps and platform engineering review is the better spend.
  • A short-lived prototype. For a proof of concept that will not hold customer data, lock files and a dependency check are enough. Add SBOMs and signing when the product is going to production.
  • Signing with nobody verifying. If no deploy step or customer will check signatures, stop at stage 3 and plan stage 5 with stage 4, not years after it.

Where AI fits

AI helps with the reading around these controls, not with the decision. A model can summarise a vulnerability advisory against your SBOM, draft the purpose and risk note for a new dependency, propose an upgrade path with a regression test. Engineers still decide whether a package is admitted, whether a finding applies, and whether an exception is acceptable.

AI coding assistants also create a new supply chain risk: they sometimes suggest packages that do not exist, which attackers can register. The security review gate for AI-generated code adds a check that every suggested dependency exists and is the intended one. Netbase works with the major commercial and open-source AI tools and models, chosen per project, and runs AI-assisted engineering under human review.

What delivery record exists, and what does not?

Netbase's published security practices are secure code review and version control, TLS in transit and AES at rest, role-based access control, MFA for admin dashboards, vulnerability scanning and penetration testing, and disaster recovery; NDAs, data processing agreements and SLAs are available on request. Security by design is part of how Netbase approaches delivery.

Netbase holds ISO 27001 certification and a SOC 2 Type II attestation for its own operations. These describe how Netbase works; they do not certify a client's build pipeline, SBOMs or signed artifacts.

Delivery follows a six-step lifecycle from discovery and architecture planning to ongoing support, with weekly reviews and a named project manager, which is where the dependency policy, open exceptions and the next adoption stage are reviewed. Netbase works as offshore development partner on a multi-tenant cloud ERP for a US client, a SaaS product where one compromised dependency would reach every tenant.

Most Netbase projects are delivered on fixed-price contracts. No Netbase record publishes SBOM counts, signing coverage, supply chain incidents or remediation times, and none is offered here as evidence that a particular control works.

Alternatives: who runs the programme, and how to choose

Ownership matters more than tools. There are three common set-ups.

  • The product team owns it. Best for one product with a stable team. Risk: the exception register and the SBOM tracker stop being reviewed when deadlines press.
  • A platform team provides it. Best when several products share CI; every team inherits the same SBOM, signing and verification steps. Risk: one policy tuned for one stack blocks another.
  • A development or assurance partner sets it up with the team. Best when there is no in-house security skill or a customer review is close. Ask who maintains the policy afterwards and how exceptions are approved. The secure development lifecycle guide shows where these steps sit next to static and dynamic testing.

Choose by three criteria: who can tell a real dependency risk from noise fastest, who can change the pipeline without stopping releases, and who will own the exception register a year from now.

Limits of this guide

  • It covers application dependencies, builds and artifacts, not hardware, firmware or vendor-hosted services.
  • It is not compliance advice; a customer contract, an auditor or a regulator decides what evidence is required.

Plan the next step with a Netbase consultant

Frequently asked questions

Yes, if you sell or run software others depend on. A scan tells you about known issues today; an SBOM stored with each release lets you answer, when a new vulnerability is disclosed, which shipped versions contain the affected component.

Usually not. A signature proves an artifact was not changed after the build, but without an inventory you cannot say what the signed artifact contains. Produce SBOMs first, then sign the artifact and its SBOM together.

Most mid-market teams should aim for SLSA Build L2 first: provenance generated and signed by a hosted build platform. Build L3, which hardens the platform against tampering during the build, makes sense when customers or regulators ask for it.

A named technical owner, usually the engineering lead of the product, should approve each exception with a reason and an expiry date. Exceptions without an owner or an end date turn the policy into a list of suggestions.

Next step

If you want SBOMs, a dependency policy and signed releases set up in your pipeline and owned by the team that ships the code, book a solution review.

Application security assurance for web, SaaS and AI features Application security assurance for web, SaaS and AI features

Netbase provides application security testing for product and engineering leaders to find and fix vulnerabilities before release, including in AI features. We review code with AI-assisted triage, scan dependencies and test running applications, then rank findings and retest fixes. Netbase's own alignment with ISO 27001 and SOC 2 Type II covers Netbase's operations, never your application.

Learn More
line
AI-assisted DevOps and platform engineering for routine releases AI-assisted DevOps and platform engineering for routine releases

Netbase provides DevOps and platform engineering for SaaS and commerce teams to make releases routine instead of risky. We set up CI/CD pipelines, repeatable environments and observability on AWS, Google Cloud, DigitalOcean or Cloudflare, with AI assistance for tests, reviews and incident triage, so every change ships the same way and problems surface before customers notice.

Learn More
line
SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules SaaS product accelerator: launch an AI-ready SaaS on proven Netbase modules

A SaaS product accelerator is a set of reusable Netbase modules for accounts, billing, roles and integrations that helps founders and product teams launch subscription software faster, with room for in-product AI from the first release. Reusing these modules can shorten development time, and the approach is proven on Printcart, Netbase's own web-to-print SaaS.

Learn More
line

Work with Netbase remotely

Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.

Contact Netbase

Discuss a project

Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.
Project enquiries

project@netbasejsc.com

WhatsApp

+84 937 869 689

Office address

91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam

Get in touch

Tell us what you want to build, modernize, or operate.

Tell us what you want to build, modernize, or operate.

Contact Netbase