You reviewed the code. You scanned the dependencies. You signed the release. But can you prove that the software running in production is the exact software your organization intended to ship?
Modern applications are created through far more than source code. Repositories, package registries, open-source projects, CI runners, reusable workflows, build images, compilers, caches, cloud identities, signing systems, artifact stores, containers, and deployment controllers all influence what eventually reaches production.
Every one of those systems creates a trust decision.
Software Build Integrity and Dependency Trust is a practical guide to designing software delivery systems in which those decisions can be verified rather than merely assumed.
Instead of treating software supply-chain security as a collection of scanners and compliance checkboxes, this book shows how to build a measurable chain of evidence from authorized change to production deployment.
Inside, you'll learn how to:
- Protect repositories, branches, tags, maintainers, workflow files, and release paths
- Control third-party components before they enter sensitive environments
- Defend against namespace confusion, malicious updates, compromised maintainers, and unsafe transitive relationships
- Treat CI/CD platforms as privileged security infrastructure
- Separate untrusted validation, trusted compilation, publication, signing, promotion, and deployment
- Build reproducible, hermetic, isolated, and evidence-producing environments
- Use immutable digests to identify exactly what was tested, approved, distributed, and deployed
- Design signing systems around identities and policy rather than shared long-lived secrets
- Generate and use SBOMs, VEX information, attestations, and machine-readable evidence
- Apply SLSA concepts without turning maturity levels into meaningless badges
- Enforce promotion and deployment decisions through policy as code
- Replace permanent automation credentials with short-lived workload identities
- Prepare for compromised runners, malicious components, stolen signers, poisoned caches, and altered release workflows
- Build forensic evidence that allows responders to quickly determine where affected components were built and deployed
- Scale strong controls across hundreds or thousands of repositories without creating release bureaucracy
Real-world incidents involving SolarWinds, Codecov, PyTorch, and xz Utils demonstrate how different parts of the delivery chain can fail—and why no single security control is enough.
The book also includes a practical maturity model, a structured learning path, a 90-day implementation playbook, a seven-layer reference architecture, a ten-question architecture review, a decision matrix, and detailed operational checklists that teams can adapt to real environments.
Whether you're a software engineer, DevOps or DevSecOps professional, platform engineer, cloud engineer, security engineer, architect, SRE, engineering leader, or technology risk professional, this book will help you answer the question that increasingly matters:
Why should this exact software be trusted to run?
Build faster when appropriate. Verify before trust. Make evidence part of the delivery system.
