Solidus Agents
How this works, explained once
Agent identity on-chain: passports, ERC-8004, delegated authority, and how a machine proves who it is acting for.
47 articles
- How to verify our ERC-8004 agent passport yourselfAn agent's own word about itself isn't worth much, anything can claim an identity.
- What is ERC-8004? The Ethereum agent identity standard, explainedERC-8004, titled "Trustless Agents," is an open Ethereum standard proposed August 13, 2025 by Marco De Rossi (MetaMask), Davide Crapis (Ethereum Foundation), Jordan Ellis (Google), and Erik Reppel…
- What is an agent passport?A passport is not a claim you make about yourself. It's issued by someone other than its holder, and its value comes entirely from that fact: a stranger at a border can check it independently,…
- Why an AI agent needs more than a wallet addressA wallet address can prove it controls funds and can sign a transaction.
- Selective disclosure for AI agents: what BBS+ proves without revealingThe unlinkability implementation described here is unaudited.
- How operator KYC works for AI agents (and why it isn't KYC-as-a-service)Solidus Agents does not perform KYC. It links to a KYC credential issued somewhere else, by verify.solidus.network, and that link, not a KYC check of its own, is what this page explains.
- The ERC-8004 Validation Registry: what evidence actually goes in itThe unlinkability implementation described here is unaudited.
- Giving a Coinbase AgentKit wallet a verifiable identityAgentKit gives an agent a wallet, and via x402, a way for that wallet to pay for things.
- What is a trust registry for AI agents?A trust registry answers "should I trust whoever issued this": a different question from ERC-8004's own Reputation Registry, which answers "has this specific agent performed well." They're related.
- Agent Card and Agent Passport: One Is On Chain, One Is a File We ServeThe passport is a token in a public registry contract. It has an identifier, an owner address and a URI string, and it is readable by anybody with a node and no account with us.
- How Agent Onboarding Works, and Why It Is Not Know Your Agent YetAn operator record is created, and it gets its own tenant.
- Do Our Credentials Also Land as Attestations? The Code Says Yes and the Running Process Says NoWrite the same credential twice. Once in our own format, and once as an attestation in a public attestation registry, with the same payload underneath.
- Machine Identity, Explained: Ours Revokes and It Does Not ExpireMachine identity is older than the agent conversation. Service accounts, certificates issued to servers, and workload identity systems that hand a running process a credential proving what it is have…
- What Is a Machine Customer, and Why Nobody Has Built One End to EndA machine customer is software that buys on its own initiative, within limits somebody set, rather than a person clicking a button and software carrying out the click.
- What Is x402, and What Our Header Does Beside Itx402 is a protocol for paying for an HTTP request. A server answers a request with the payment-required status, states what it wants, and the client pays and retries.
- How to Register an AI Agent, and the One Field to Decide Before You StartThe creation route defaults custody to managed. Omit the field and you get managed custody.
- How to Issue a Capability Credential, and What the Bundle in the Response Actually IsTwo credential types through this path, an ownership binding and a capability scope, and a claims object that is untyped at the boundary and validated in the service against a schema for that type.
- How a Relying Party Verifies an Agent Credential, and Why `valid` Is Not the Field You WantOne header. Inside it, a selective-disclosure proof, the issuer's public key, the messages the agent chose to disclose, a timestamp, and optionally a status reference and a credential identifier.
- How Revocation Works, and the One Case Where the Authoritative Answer Is the Stale OneMarks the credential revoked in the database, idempotently, and safely if two requests race.
- Agent Identity SDK Quickstart, With the Three Things a Quickstart Usually HidesBoth are at 0.1.4 on the public registry. The repository declares 0.1.0, so anybody reading our source instead of the package is reading a different thing from the one they installed.
- Agent Credential Webhook Events, and Why the Timestamp Is Not Where You Will Look For ItSubscribe an endpoint and a list of events, and the subscribe response returns a signing secret.
- Credential-Gated API Access, and the Line Between What the Proof Covers and What It Does NotReplace the anonymous API key with a verified identity. Instead of a shared secret that proves somebody once received a string, the caller presents a proof that a named agent holds a credential saying what it may do.
- Registries and the Marketplace Word: We Write to One of Three, and the Third Is Only an AddressThe standard defines three. One for identity, one for validation evidence, one for reputation.
- Who Needs Agent Identity Today, and the Honest Answer About How We Would Know
- Composing With Olas, and a Claim We Found in Our Own Package That We Will Not RepeatOur product contract puts their adapter in the compose-with column and requires additive framing.
- Composing With OriginTrail: A Thesis With No Code Behind It, Said First Rather Than LastWe ship four framework adapters. This is not one of them.
- The Agent Card Schema, Field by Field, and the Two Fields You Cannot Branch OnA JSON document served without authentication at the address the passport token points to.
- Base Sepolia and Ethereum Mainnet for ERC-8004: The Address Does Not Tell You the Network
- What an ERC-8004 Anchor Costs: Gas Units We Measured, and Why We Will Not Give You a PriceRegistering an agent in the identity registry consumed between 183,744 and 200,844 gas, with a mean near 184,750.
- A Threat Model for Agent Credentials, Written From What We Measured Rather Than What We IntendedThere is no value behind these credentials yet, which is the only reason the gaps below are cheap today.
- Agent Credentials and the European Wallet: A Roadmap Target, and a Question Nobody Upstream Has SettledThis page describes engineering state and a roadmap. It is a reading of our own code and our own plans, not legal advice, and nothing on it should be relied on as a statement about what any regulation requires of you.
- What Agent Identity Actually Costs: There Is No Price, and the Free Tier Promises Two Agents It Will Not Give YouOur own pricing page publishes a free tier at zero and a paid tier whose price reads "coming soon", with billing described as not yet available and the tier's details as still being finalised.
- How Agent Identity Companies Make Money, and What Our Own Meter Says We IntendAnd other companies' revenue is not ours to report. Our internal notes carry funding figures and scoring tables for several of them; none of it was verified and none of it appears here, on the same…
- Does Anyone Need Agent Identity Yet?The audience page asks WHO would need this and answers with hypotheses, each paired with what would kill it.
- Could AI Agents Have a Trust Score?This surface computes no score, stores no score and serves no score.
- One Identity Primitive for Humans, Agents and Machines: True of the Identifier, Not of the CredentialOne identity primitive that works for a person, for an agent and for a machine is the sentence this company organises itself around.
- Before Agent Identity Had a Name: What Is Actually Datable, and Why We Claim No PrescienceSoftware has been proving what it is to other software for decades.
- Is This Just Another Blockchain Identity Project? Partly Yes, and Here Is the Part That Is NotYes, we run our own network, and the identifier resolves there.
- What Is Know Your Agent? A Borrowed Word Doing Work Nobody Has Earned"Know your customer" is a regulatory obligation. It has a defined scope, entities supervised against it, records that must be kept, and consequences when it is done badly.
- Why Nobody Prices Agent Identity Yet: The Natural Unit Is the One You Cannot MeterThe event that carries the value in agent identity is verification, and verification is designed not to be observable.
- Proof of Personhood and Agent Identity: Opposite Questions, and Ours Does Not Answer Either One AlonePersonhood asks whether there is exactly one human here. Agent identity presupposes there is none, and asks who is accountable instead.
- What Is AP2? A Payment Protocol Whose Own Glossary Puts a Human Back in the LoopA payment protocol for agents has to answer a question a payment rail never had to: on whose authority is this agent spending, and how would anyone check.
- Verifying Agent Identity at an MCP Server: What Changes When You Replace the API Key, and What Does NotBoth an API key and a credential bundle are bearer credentials, so this is not a migration that makes theft harder.
- What Is ERC-8211? Nothing. We Checked, and Here Is Exactly How
- KYC, KYB, KYE, KYA, KYM: A Test for Telling an Obligation From a Coinage, and One Answer We DeclineThis page describes our own engineering state and offers a way of thinking.
- The elizaOS Plugin: The One Adapter That Speaks the Host's Language, and the Same Custody DefaultTheir runtime runs agents. We issue credentials. Different layers, and our product contract puts this adapter in the compose-with column, which is also the only description the code supports.
- MiCA and Autonomous Agent Payments"We do not move money" is a fact about a system. "Therefore we are outside a regulatory regime" is a legal conclusion about a company.