AIBOM workbench

Forty catalogued models with owner and jurisdiction already recorded, plus live resolution against the Hugging Face Hub for any model at all — and an advisory check on the loading stack for both.

Step 1

What you need to provide

Only the model ID is required. Everything else either comes from the Hub, or is something the Hub structurally cannot know — jurisdiction and legal entity have no field in CycloneDX or SPDX and no registry publishes them. Arriving here from a catalogue model pre-fills those two.

The org/name path from the model's Hub URL.

Blank resolves the head and pins whatever SHA comes back.

The Hub gives an org handle, not a legal person.

Decides whether hosted use is an export. Goes in as a property.

The BOM subject.

Checked against the advisories locally. Nothing is sent anywhere.

Coverage

Which bill of materials covers what

Every advisory in this tool sits in the SBOM and AIBOM overlap. That is the whole argument for merging them rather than filing an AIBOM and calling the job done.

SBOM AIBOM packages, versions weights, lineage +-------------------+---------+-------------------+ | | loader | | | syft, trivy | stack | transformers | | lockfiles | <- CVEs | reads this | +---------+---------+---------+---------+---------+ | | crypto | all | signing | | | | libs | three | sigstore| | +---------+---------+---------+---------+---------+ | CBOM | | algorithms, keys | +-----------------------------+ SBOM only packages, versions, transitive dependencies AIBOM only weights, base-model lineage, training datasets CBOM only algorithms, key lengths, certificates SBOM n AIBOM the loader stack. EVERY published CVE lives here. SBOM n CBOM crypto libraries (openssl, cryptography) AIBOM n CBOM model signing and attestation all three a signed, pinned, verified artifact Not shown, because they do not overlap: HBOM accelerator, firmware, country of origin SaaSBOM endpoints, trust boundaries, tools and MCP servers

Scoring

Methodology

Twelve fields are scored. A field counts as present only if the upstream owner published it. A value supplied by the catalogue populates the BOM but does not score, and is tagged atlas:provenance:* to say so — otherwise the score would measure this catalogue's completeness rather than the vendor's.

FieldSourcePathAvailability
commit_shaHub APImodel_info().shaAlways, if the repo resolves
weight_file_hashesHub APIsiblings[].lfs.sha256Almost always
safe_serializationHub APIa .safetensors sibling existsUsually
model_architectureRepo fileconfig.json → architecturesUsually
parameter_countHub APIsafetensors index totalOften
declared_licenceModel cardYAML licenseUsually
pipeline_taskModel cardYAML pipeline_tagUsually
base_modelModel cardYAML base_modelSometimes
training_datasetsModel cardYAML datasetsRarely published
evaluation_resultsModel cardmodel-index blockRarely published
owner_entityCataloguethis atlasAlways
owner_jurisdictionCataloguethis atlasAlways

Why the catalogue reads 17%

Ten of the twelve fields require a live Hub call. Generated offline, only the two catalogue-sourced fields are present: 2 ÷ 12 = 16.67%, rounded to 17%. The uniform score is the tell — it measures the absence of a network call, not the models. Resolving live produces a spread.

Why hosted models are scored out of 6

Six fields describe a downloadable artefact. A hosted API model has no repository, no weight files and no config.json, so those cannot exist. Counting them as missing would imply a gap the vendor could close by publishing a file it will never have. They are reported as not applicable and excluded from the denominator.

Why 100% is the wrong target

Even fully resolved, almost nothing reaches 100%. training_datasets and evaluation_results are absent from most model cards because owners choose not to publish them. Expect roughly 75% for a well-documented model and 50% for a typical one. The score measures upstream disclosure, not the quality of your inventory.

Advisories

The full advisory set

No CVE targets model weights. Weights are data, not executable software, so nothing is published against them. Every advisory below sits in the loader — the SBOM and AIBOM overlap.