OpenAgentID documentation
Start here

DID resolution

Resolve documents through explicit backends, then verify the trust required by your action.

DID parsing, document retrieval, cryptographic verification and authority confirmation are separate contracts. A returned document is not automatically permission to execute an action.

Current resolver library

oas-resolve defines an async Resolver trait and utility implementations:

  • InMemoryResolver for explicit local/test documents.
  • CachingResolver for time-bounded cached resolution.
  • FallbackResolver for priority-ordered backends.
  • AnchoredResolver for the configured authority-anchor checks.
  • ResolutionMetadata and typed ResolveError results.

Concrete HTTP, DHT or blockchain backends must implement the resolver contract; the utility crate does not itself prove a globally deployed resolution network.

Resolve safely

Parse with the actual OasDid type, select trusted resolver backends, retrieve the document, validate its structure and proofs, and apply the lineage/revocation/authority policy needed by the operation. For agent:instance, the kind has two components; a fixed-count string split is not a DID parser. Identifiers need not all be a multibase public key.

Cache invalidation and fallback order are application choices. An unavailable required authority backend should produce a denied/unverifiable operation, not an invented positive result. Inspect typed errors to distinguish a malformed DID, unresolved document, signature failure and unavailable authority evidence.

See resolver modules, AEGIS verification, and anchor policy.

On this page