New to Rust? Grab our free Rust for Beginners eBook Get it free →
Key Components in a Software Bill of Materials

To review the software in a product, you need a record of its components and how they connect. A software bill of materials (SBOM) records those components and relationships for a product.
This article explains which fields make an SBOM useful and how teams build and maintain one.
TL;DR
An SBOM identifies the components in a software product and records how they connect.
- Identify the subject product or build, document author, and creation time.
- Record each component’s supplier, name, version, and available identifier.
- Include direct and indirect dependency relationships, with unknowns and scan boundaries stated.
- Validate a machine-readable export and refresh it when the delivered software changes.
What is a software bill of materials?
An SBOM records a product’s software components and how they relate. CISA’s July 2026 minimum-elements guidance updates and replaces the NTIA’s 2021 baseline. Treat older checklists as historical rather than current guidance.
SPDX and CycloneDX are machine-readable formats. A VEX statement can report whether an advisory applies to a product. CycloneDX models components and services, while SPDX provides package and relationship information.
What information makes an SBOM useful?
The table separates document-level details from fields attached to individual components.
| Information | What it records | How it helps |
|---|---|---|
| Document author | Person or tool that produced the record | Identifies the source for follow-up questions |
| Creation timestamp | When the inventory snapshot was created | Helps assess freshness before a release comparison |
| Document identifier | Identifier for the exported document | Distinguishes one inventory file from another |
| Subject | Product, release, build, image, or other artifact under review | Keeps the inventory tied to the correct deliverable |
| Supplier | Organization that supplies the named package | Helps distinguish packages with similar names and identify the upstream source |
| Component name and version | Published package name and resolved version | Supports precise lookup and advisory matching for the installed release |
| Package identifier | PURL, CPE, or another stable ecosystem identifier when available | Provides a consistent reference across tools and registries |
| Dependency relationship | Which component depends on or includes another | Shows direct and nested paths into the product, not just a flat name list |
| Artifact hash | Fingerprint calculated from artifact bytes | Helps compare the same artifact across build stages |
| Provenance | Available source or build context | Shows where the component data came from |
| License and copyright | Declared license information and copyright notices | Flags data that may need human review |
Retain original exports with normalized copies so reviewers can trace changes back to their source.
How do teams build and maintain an SBOM?
Build the inventory from the resolved or delivered software, then keep its scope and release identity attached to the export. This reduces the chance that a declared dependency list is mistaken for the full shipped composition.
Step 1: Discover direct and nested components
Collect first-party modules and third-party packages, including dependencies brought in by other packages. In a Node.js project, package manifests and lockfiles are useful inputs, but they describe different things: a manifest can state acceptable version ranges, while package-lock.json records npm’s resolved package tree. A generator may also inspect built artifacts or container images when the delivered composition differs from those files.
Step 2: Assign component identities
CodeForGeek’s guide to tilde and caret ranges explains why a package.json range may differ from the installed version. Record the package name and supplier from source data, and leave unresolved identity fields unknown for follow-up.
Step 3: Record dependency relationships
Retain the direct-versus-transitive distinction in the selected schema. That lets a reviewer follow a nested library through its parent packages to the product.
Step 4: Validate, label, and refresh the document
Export in a format the recipient supports, then validate it against the format specification.
Record the subject and creation time for each build. Name the generator and the input types it examined.
Generate a new record when the build changes, and retain the prior one for comparison.
A schema check cannot prove that every package was discovered. Compare the export with build inputs and state which inputs or runtime areas the scan excluded.
What should you check when SBOM data is incomplete?
Confirm the subject and release before treating a missing or unexpected entry as a product property.
| Case | What to check | Practical response |
|---|---|---|
| Supplier or version unknown | Whether source files or artifact data identify the package | Keep the value unknown and check the originating manifest or supplier |
| Similar package names | Supplier, ecosystem, package identifier, and exact version | Keep entries separate until the identifiers show they refer to the same package |
| Nested dependency missing | Whether the scan included transitive packages and the resolved build | Regenerate from the resolved build or delivered image, then compare coverage |
| Runtime service omitted | Whether remote APIs or dynamically loaded software sit outside the build scan | Document the boundary and add service context when the format supports it |
| Inventory does not match a release | Subject, build identifier, timestamp, and artifact association | Use the record created for the release under review |
| Old inventory timestamp | Whether the record was generated for the artifact that shipped | Generate a fresh record for that release and retain the prior copy |
| License declaration missing | Whether source metadata is absent or a scan did not find it | Mark unknown and route the package for human review |
| Vulnerable component listed | Whether the advisory applies to the product’s use and configuration | Check vendor guidance or VEX status before assessing applicability |
CISA’s SaaS transparency guidance explains that a customer may not be able to inspect a provider’s running environment. Ask which service instance the inventory covers and how provider-managed components are represented.
Agree on an update trigger and delivery route, then keep each notice with the relevant service release.
CISA’s 2024 transparency paper describes three practice levels for SBOM attributes: minimum expected, recommended, and aspirational. The SBOM-a-Rama agenda from September 2024 covered generation references, sharing roles, interoperability, VEX, and collection management, all parts of the process beyond producing a file.
When a producer cannot supply a recommended attribute, note whether the reason is scan coverage, unavailable source data, or a disclosure boundary.
AI product inventories can include models and related assets beyond ordinary package dependencies. CISA’s 2026 guidance notes that these software types may need additional elements, so document the inventory’s scope instead of assuming a package list covers them.
How does an SBOM support business and release planning?
An SBOM gives teams a shared set of component records for release review, supplier questions, and security triage.
- The producer maintains the export and answers questions about fields or provenance.
- The reviewer compares unexpected components with the change request and test evidence.
Agree on a delivery method and an update trigger before exchanging inventories. When a changed record arrives, compare it with the prior copy and send questions about identity or scope to the producer.
| Planning question | Evidence to inspect | Useful next action |
|---|---|---|
| Which deliverable is covered? | Product, release, build, and timestamp metadata | Match the record to the artifact under review |
| What changed? | Added, removed, or version-changed component records | Route an unexpected change to its maintainer |
| Can an alert be assessed? | Package identifiers, versions, and dependency paths | Check product-specific applicability and test evidence |
| Who receives the record? | Recipient and sharing scope for the release | Send the product-specific view and retain internal response data |
Compare candidate and shipped inventories to identify component changes, and retain the prior record to trace what changed. Route unexpected entries for review alongside test results, risk acceptance, and the reason for the change.
This resembles master planning because teams coordinate inputs and timing, but its planning object is software composition. Use the release identifier as the join point between the inventory, build record, vulnerability workflow, and recipient copy. That choice helps prevent component data from being attached to the wrong build.
How do costing and inventory concepts apply to software?
Manufacturing systems plan physical materials from assembly quantities and stock balances. For software, link purchasing and production records to products or releases instead of translating them into package counts.
The phrase “bill of materials” can refer to either record, so state whether a workflow concerns physical assembly or software composition.
| Business task | Software record to use | How to connect it |
|---|---|---|
| Cost review | Supplier and package identities, plus prices and contract terms held in procurement or finance | Join records by a stable product or supplier reference |
| License review | Declared license and copyright data, with source and timestamp | Route absent or ambiguous declarations for human review |
| Inventory search | Component identity and version within each release | Search across releases, then confirm the exact artifact before acting |
| Release planning | Candidate and shipped-build inventories | Compare records after a dependency change and retain both versions |
When a separate license or vulnerability service enriches the inventory, keep its findings separate and record the service’s provenance and timestamp.
Business management software can connect product records to purchasing, production schedules, and inventory balances. Software governance can connect SBOM data to a release record, build pipeline, vulnerability workflow, or procurement system.
For a workflow check, confirm that a sample record names its component version, dependency path, generation time, and unknown fields. Then verify that it points to the intended artifact and can be compared with another release.
Conclusion
Review component identity, dependency relationships, release scope, and scan boundaries together. Read CISA’s 2026 minimum elements for the updated baseline, then use the CycloneDX overview or SPDX package specification for format details. For generation and sharing resources, see NTIA’s SBOM resource hub.
FAQ
These answers cover common questions about SBOM fields and formats.
What are the main components of an SBOM?
An SBOM identifies the subject product and its components, including supplier, name, version, identifiers where available, and dependency relationships. Document author and timestamp help establish source and freshness.
Which formats can an SBOM use?
SPDX and CycloneDX are machine-readable SBOM formats. Choose one supported by the tools and recipients in your workflow, then validate it against its specification.
Does an SBOM show whether a vulnerability is exploitable?
An SBOM lists components and relationships. Check product-specific vulnerability applicability separately, including VEX information where available.
When is an SBOM hash useful?
A hash can help compare artifact bytes from different build stages. Use package names and identifiers to locate the component itself.




