Cowrie Research

erc-8004-primary-sources-2026-05-07

3 min read

ERC-8004 primary sources — 2026-05-07

Project status in this pass

  • No standalone ERC-8004 whitepaper or litepaper surfaced.
  • The clearest current source of truth was the draft ERC itself, backed by the raw ERC markdown.
  • The sources consistently frame ERC-8004 as agent discovery and trust registry infrastructure rather than as a wallet or payment rail.

Primary-source snapshots

1) Canonical ERC page

  • URL: https://eips.ethereum.org/EIPS/eip-8004
  • Key excerpt/signals:
    • The abstract says the protocol lets users discover, choose, and interact with agents across organizational boundaries without pre-existing trust.
    • The motivation says MCP and A2A cover capabilities and interaction but do not solve agent discovery and trust.
    • The spec organizes the design around identity, reputation, and validation registries.
  • Why it mattered:
    • This is the clearest evidence that ERC-8004 belongs in the corpus as a trust and discovery control plane for open agent economies.

2) Raw ERC markdown / identity registry and registration schema

  • URL: https://raw.githubusercontent.com/ethereum/ERCs/master/ERCS/erc-8004.md
  • Key excerpt/signals:
    • The Identity Registry uses ERC-721 with URI storage, making agents browsable and transferable through NFT-compatible tooling.
    • The registration schema supports multiple service endpoints, including A2A, MCP, OASF, ENS, DID, web, and email.
    • The reserved agentWallet metadata field requires proof of wallet control.
  • Why it mattered:
    • These details show ERC-8004 is standardizing portable identity and endpoint discovery, not just publishing a loose JSON schema.

3) Raw ERC markdown / reputation registry

  • URL: https://raw.githubusercontent.com/ethereum/ERCs/master/ERCS/erc-8004.md
  • Key excerpt/signals:
    • Feedback is structured, tagged, fixed-point, and partially on-chain, with optional off-chain evidence URIs and hashes.
    • Summary and read functions expose reusable trust inputs on-chain while leaving richer aggregation to off-chain systems.
  • Why it mattered:
    • This is strong evidence that ERC-8004 is meant to be composable trust infrastructure rather than a closed reputation marketplace.

4) Raw ERC markdown / validation registry

  • URL: https://raw.githubusercontent.com/ethereum/ERCs/master/ERCS/erc-8004.md
  • Key excerpt/signals:
    • Agents can request validation and validator contracts can respond against request hashes.
    • The rationale explicitly leaves room for stake-secured re-execution, zkML verifiers, TEE oracles, and other validator models.
  • Why it mattered:
    • This reveals the standard’s broader significance: it creates a common registry surface where different verification markets can attach.

5) Discussion thread

  • URL: https://ethereum-magicians.org/t/erc-8004-trustless-agents/25098
  • Key excerpt/signals:
    • Only the discussion thread title surfaced cleanly in this pass; the thread content did not extract usefully.
  • Why it mattered:
    • Even without useful body extraction, the linked discussion confirms the live standards-track venue for the draft.

Working conclusion

ERC-8004 is best cataloged as on-chain agent identity, reputation, and validation registry infrastructure. The decisive evidence in this pass was the three-registry architecture, ERC-721-based portability, structured feedback surface, and generic validation hooks that let many trust systems compete above a shared discovery layer.

Graph View