The glossary holds the canonical definition of every term this site uses, one per term. This page is the quickstart's short version: a summary line per term, each linking to the glossary section that owns the detail. It defines nothing the glossary does not.
Roles and records
- Human owner: the verified human principal behind every agent. Glossary
- Agent (sub-identity): a broker record with an id (
agt_...), a dedicated inbox address and delegated scopes; the identity the model acts under. Glossary - Harness: the client that connects to the broker's MCP endpoint and runs the agent. Glossary
- Broker: the service that verifies the human, holds the sub-identity and inbox, mints assertions, records acceptances and receipts, and exposes the MCP tools. Glossary
- Relying party (provider): a service that accepts brokered agent signups by publishing discovery and verifying an assertion. It may publish an optional broker-issued verification credential. Glossary
- Provider verification credential: the broker-issued RP attestation binding origin, audience, identity, level and keys. Glossary
- Delegated scopes: the permissions the agent holds, embedded in
delegation.scopes. Glossary
Artifacts
- Discovery document: the identity-only JSON at
/.well-known/agent-signup.jsonon the provider's origin. Glossary - Assertion: a short-lived Ed25519-signed JWT that proves the agent, the verified human owner and the delegated ToS acceptance. Glossary
- Receipt (
signup_token): the provider's RP-signed proof of signup, recording the identity facts, the accepted ToS and an optional context pointer. Noexpin v1.2. Glossary - Context guide: a hash-pinned URL to the provider's post-signup guide, usually a
SKILL.md. Glossary - Ledger: the acceptance ledger and the service signups ledger, two records with one name. Glossary
- Kill switch: the per-agent revocation control in the portal. Glossary
Term to where it lives
| Term | What it is | Where it lives |
|---|---|---|
| Human owner | Verified human principal | Broker account, human.id in the assertion |
| Agent | Sub-identity with an inbox | Broker record, agent.id in the assertion |
| Harness | MCP client holding the session | Your machine |
| Discovery | Identity-only JSON | Provider origin, /.well-known/agent-signup.json |
| Terms | Exact legal text plus hash | terms in discovery |
| Assertion | Short-lived signed JWT | Returned by identity_signup or identity_prove, sent to the provider |
| Receipt | RP-signed proof of signup | signup_token in the signup response |
| Provider verification credential | Optional broker-signed identity | rp_attestation in discovery |
| Context guide | Hash-pinned vendor guide | Provider origin, linked by context.url |
| Acceptance | Durable consent event | Broker ledger, one per agent and terms hash |
| Signup proof | Recorded account attribution | Broker ledger, written by Contract F or identity_record_signup |
| Revocation | Human kill switch | Broker, sets revoked |
| Suspension | Operator protective action | Broker, sets suspended |
Next
- How it works puts these terms in order.
- Trust and limits explains what revocation and the ledger do not cover.
- The specification defines the field-level contract.