ERC-8004 Mainnet Launch Gives Agents a Trust Layer

ERC-8004 mainnet launch marks a shift from agent demos to shared infrastructure. Week in Ethereum News highlighted the launch on March 20, but the real builder takeaway is narrower: Ethereum now has live agent identity and reputation registries, even as the standard stays in Draft and its validation layer remains unfinished.

What ERC-8004 mainnet launch actually shipped

ERC-8004, titled "Trustless Agents," is an ERC proposal for discovering agents and establishing trust through three registries: Identity, Reputation, and Validation. The EIP says those registries are meant to let agents interact across organizational boundaries without pre-existing trust, and it explicitly frames the problem as a gap left by existing agent communication standards. On March 20, Week in Ethereum News called the mainnet launch the highlight of the week, and the ERC-8004 contracts repository now lists deployed IdentityRegistry and ReputationRegistry contracts on Ethereum mainnet, alongside deployments on Base and Abstract. That is why the launch matters. This is no longer only a design discussion on Ethereum Magicians. Parts of the standard are now live as shared infrastructure that developers can integrate against.

The deployment pattern also shows the project is thinking beyond one chain. The EIP says the registries are expected to be deployed as per-chain singletons, and notes that an agent registered and receiving feedback on one chain can still operate and transact on others. The contracts repository lines up with that model by listing mainnet deployments across several EVM networks plus Sepolia test deployments. For builders, that is the more useful signal inside the launch. ERC-8004 is trying to make identity and trust portable across agent markets that may execute, pay, and settle in different places.

ERC-8004 EIP

Why agent builders needed more than MCP and A2A

The core motivation in the EIP is straightforward. MCP can expose capabilities, and A2A can handle messaging and task orchestration, but neither standard solves discovery and trust on its own. ERC-8004 tries to fill that missing layer. The proposal says developers need a way to discover agents and evaluate them in untrusted settings if open, cross-organizational agent markets are going to work at all. Without that, most agent systems drift back toward closed directories, curated marketplaces, or reputation controlled by one platform.

That is the strongest reason to care about the launch. ERC-8004 does not try to replace every agent protocol. It tries to become the trust and identity layer that sits underneath them. The standard's registration file can advertise A2A endpoints, MCP endpoints, ENS names, DIDs, email addresses, and other service pointers, which makes it more like a portable agent passport than a single-purpose API spec. For Ethereum builders, that matters because it keeps discovery composable. A team can build its own marketplace, coordinator, or evaluation layer without having to invent a new identity primitive for each product.

Web3 Builder coverage

How the registries work in practice

The Identity Registry is an ERC-721 with URI storage. In plain terms, an agent gets an NFT-style onchain handle whose tokenURI points to a registration file containing metadata and service endpoints. The EIP says this gives each agent a portable, censorship-resistant identifier, and the spec lets that registration file advertise several endpoint types at once, from A2A agent cards to MCP endpoints and onchain wallet references. The standard also reserves an agentWallet field that must be re-verified if ownership changes, using EIP-712 signatures for EOAs or ERC-1271 for smart contract wallets. That detail is easy to miss, but it matters because it ties advertised payment or control addresses back to the onchain identity rather than leaving them as unverified strings.

The Reputation Registry is the second live piece. The EIP describes it as a standard interface for posting and fetching feedback signals, with simple filtering and composability onchain and more advanced aggregation expected offchain. That split is sensible. Onchain storage gives public, queryable raw signals. Offchain scoring lets builders create more sophisticated trust products without forcing one canonical ranking algorithm into the base standard.

agent identity archive

The Validation Registry is where the design becomes more ambitious. ERC-8004 envisions validators posting responses tied to request hashes, potentially using stake-secured re-execution, zkML proofs, or TEE oracles. But the contracts repository warns that this part of the spec is still under active update and discussion with the TEE community, with a follow-up revision expected later in 2026. That means the current mainnet launch is real, but partial.

Why the launch matters even though the spec is still Draft

The most honest way to read this launch is as a live start, not a finished standard. The EIP page still labels ERC-8004 as "Draft," and the proposal's citation section itself refers to the document as "ERC-8004: Trustless Agents [DRAFT]." The repository also makes clear that the validation registry is not yet stable. So builders should avoid reading "mainnet launch" as "finalized Ethereum standard." It means deployed contracts and active experimentation, not full closure on the specification.

That distinction is exactly what makes the story useful. Ethereum often ships this way: interfaces begin attracting integrations before the governance and implementation edges are fully settled. In ERC-8004's case, that means developers can already build against live identity and reputation registries while treating validation as an evolving extension rather than a settled base layer. The advantage is obvious. Teams do not need to wait for the entire stack to be complete before testing discovery, registration, and feedback flows in production. The risk is also obvious. If the unfinished parts change materially, early integrations may have to adapt.

There is already evidence the standard is being pulled into adjacent work. ERC-8183, "Agentic Commerce," recommends integrating with ERC-8004 for onchain reputation and attestation, and describes patterns for mapping job outcomes into trust signals or using hooks to enforce reputation-aware policies. That is the kind of composability signal builders should watch more closely than generic hype around AI agents on Ethereum.

agentic commerce archive

What builders should watch next

The next test is not whether more demos appear. It is whether ERC-8004 becomes a default integration target for agent products that need portable identity, public reputation, or machine-readable service discovery. Builders should watch three things. First, whether major agent frameworks and wallets start resolving ERC-8004 registration files directly. Second, whether marketplaces and commerce protocols actually write feedback into the reputation layer instead of keeping ratings in app databases. Third, whether the validation registry matures into something credible enough for higher-stakes use cases.

That last point will decide how far the standard can go. Identity and feedback are enough for discovery and low-stakes marketplaces. They are not enough for sensitive workflows where developers want verifiable execution or stronger guarantees about what an agent actually ran. The EIP explicitly leaves room for multiple trust models, from reviewer feedback to zkML and TEE-backed validation. If those extensions land cleanly, ERC-8004 could become useful infrastructure for open agent economies. If they do not, the launch will still matter, but mostly as a discovery standard with a reputation rail attached.

ERC-8183 Agentic Commerce

Ethereum has now put a live onchain trust layer for agents in front of builders. The harder question is whether enough products will use it before the unfinished parts of the spec settle into place.