JSON Schema (v0.1)

Structural schema for llmo.json v0.1 documents.

A formal JSON Schema (draft 2020-12) defines the structural shape of a conforming llmo.json document. The raw schema is published at:

https://llmo.org/spec/v0.1/schema.json

The schema’s $id is https://llmo.org/spec/v0.1/schema.json. Tools that fetch schemas by $id will resolve to that URL.

What the schema validates

• Presence and types of required top-level fields (llmo_version, entity, claims, valid_from, valid_until, document_id).
• Constant value of llmo_version ("0.1").
• Shape of the entity object, including regex constraints on primary_domain, aliases, and well-known external_ids (Wikidata QID, DUNS, LEI, DID).
• Shape of each claim object.
• Shape of the statement payload for each of the eight core claim types (identity, canonical_urls, official_channels, product_facts, personnel, disavowal, supersedes, pointer), applied via conditional allOf / if / then constructs.
• Shape of the optional signature object on both documents and claims.
• Pattern constraints on identifier strings and domain literals.

What the schema does not validate

The JSON Schema is a structural check only. It does not validate the following conformance-level concerns, which are the validator’s responsibility (see spec §5):

• That valid_until is within 365 days of valid_from.
• That entity.primary_domain matches the domain serving the file.
• That URLs inside claims resolve to the entity’s primary domain or declared aliases.
• That a signed document’s JWS verifies against the JWKS at /.well-known/llmo-keys.json.
• That RFC 8785 (JCS) canonicalization of the signed payload reproduces the signed bytes.
• That the document is actually served at /.well-known/llmo.json over HTTPS.
• That supersedes references prior document_id values that existed.

A document can pass the JSON Schema and still fail minimal, standard, or strict conformance.

Using the schema

Common usage patterns:

• JavaScript / TypeScript. Use ajv with the 2020-12 meta-schema enabled (new Ajv2020()). Load the schema with ajv.addSchema(schema, '...') or pass directly to ajv.compile(schema).
• Python. Use jsonschema with Draft202012Validator. Wrap the document with validator.validate(document).
• Go. Use santhosh-tekuri/jsonschema/v5 or similar; both support draft 2020-12.
• Editor integration. Many JSON editors (VS Code, JetBrains) will autocomplete against the schema if the document declares "$schema": "https://llmo.org/spec/v0.1/schema.json" at the top, or if the editor is configured to associate the schema with files matching llmo.json.

Note: llmo.json files do not include a $schema field in the protocol itself, because the binding is by file name and discovery path rather than by in-document declaration. Tools that want schema-driven autocomplete while authoring may add one locally, but it is not part of the published document.

Download

• Raw JSON: /spec/v0.1/schema.json
• Canonical $id: https://llmo.org/spec/v0.1/schema.json