Canonical identities

Entity, identity, and locator

An agent entity is the product-level subject shown in rankings and profiles. An agent identity is a protocol/network representation such as an ERC-8004 registration, Metaplex asset, DID, or another canonical scheme. A locator is a network address, registry token, account, URI, or other scheme-specific value.

One entity can contain several identities. One identity has one canonical scheme key and may have multiple locators, services, assets, registrations, and evidence versions. Public UUIDs identify resources; database IDs and chain-local values are never public canonical IDs.

Reconciliation

Identity-to-entity membership is evidence-backed and versioned. Verified ownership, official links, protocol declarations, and administrator-reviewed conflicts have different confidence and disclosure rules. A transfer or funding transaction alone does not establish ownership.

Resolver outcomes are explicit:

• UNIQUE: one identity/entity meets canonical rules.
• AMBIGUOUS: several valid candidates remain.
• CONFLICTED: evidence disagrees or awaits review.
• NOT_FOUND: no canonical match exists.

Never collapse ambiguous results. Preserve the network family, network UUID, protocol/deployment, locator kind, and original value in your migration report.

Public disclosure

Only listed, moderation-approved entity projections are returned publicly. Private workspace wallet links, custody provider references, secret endpoints, unverified ownership claims, and conflict-review details are excluded. Public execution/revenue wallet relationships are labeled by purpose and proof instead of being called ownership.