Skip to content
↑↓ to move · Enter to openEsc to close
Glossary · Identity

Agent-to-agent authentication

Updated

When a purchasing agent talks to a supplier's sales agent, neither side has a person present to vouch for it. Each needs some way to learn which agent it is talking to and who is accountable for it.

The term covers several layers: authenticating the channel, authenticating individual requests, and identifying the accountable operator. Different standards address different layers.

Definition

Agent-to-agent authentication is how one AI agent establishes which agent it is dealing with, and who operates it, before exchanging data or completing a task with it.

Also known as: A2A authentication, machine-to-machine agent authentication

How it relates to AgenID

AgenID covers the operator-identity layer. An agent presents its identifier; the other agent resolves it, re-verifies the operator's Ed25519 signature and reads the declared purpose. The learn guide walks through the flow. The check needs no account or API key, and the verifying agent does not have to trust AgenID, because the registry cannot forge an operator signature.

AgenID does not authenticate live requests today: nothing binds a specific HTTP request to a registered agent. Request signing is the layer approaches like Web Bot Auth work on.

Why it matters

Agents increasingly act on both sides of a transaction. A stable identity lets history and accountability attach to the agent across platforms instead of restarting at every service. Without a shared identity layer, each pair of agents needs a bilateral arrangement such as exchanged API keys, which does not work for agents meeting for the first time.

The reference deployment at agenid.com issues nothing above L1_REGISTERED today. L2 and higher require a VerificationAssertion signed by a root authority key, and that key has not been created yet.

What it does not mean

Authentication answers who; it does not answer what the agent is allowed to do. Delegated authority from a principal is a separate object that v1.1.1 does not define.

Frequently asked questions

Can two agents verify each other with AgenID today?

Each can resolve the other's identifier and re-check the operator signature. Binding that identity to a particular live request is not implemented yet.

Is this the same as OAuth between services?

No. OAuth grants access to resources. AgenID identifies the accountable operator of an agent.

Where does the A2A Agent Card fit?

It handles discovery and capabilities. See Agent Card.

What agenid.com issues today

The reference deployment at agenid.com issues DECLARED and L1_REGISTERED only. L2_DOMAIN_VERIFIED and higher require a VerificationAssertion signed by a root authority key, and that key has not been created yet. L5 is reserved by the specification and is not issuable at all.

Identity is not permission

Which agent is acting and who declared it is one question. What that agent is allowed to do is another, and it stays with the party on the other side and with the operator. Protocol v1.1.1 defines no signed authorization object.

Sources

  1. A2A Protocol specification
  2. RFC 9421: HTTP Message Signatures

Give your agent an identity

Register an agent in the browser — the signing key is generated on your device and never sent to AgenID — or resolve an existing AgenID and re-check its signature yourself.