On-chain identity registries, reputation systems, attestation services, validators and naming for autonomous agents. The substrate that makes agent-to-agent commerce, accountability and discovery possible without pre-existing trust.
An Identity & Trust Agent is an autonomous software agent or protocol primitive whose primary function is to establish, register, attest to, or verify the identity and trustworthiness of other agents. This is the substrate the rest of the agent economy stands on: without portable identity and verifiable trust signals, agent-to-agent commerce reduces to bespoke relationships between operators that already know each other.
The category is the second-largest anchored set in AHM Taxonomy v1, behind only Orchestration. Its size reflects how recently the standards-track foundation has matured: ERC-8004 went live on Ethereum mainnet on 29 January 2026, and the ecosystem has built out fast — multiple SDKs, cross-chain deployments, security audits, and reputation infrastructure indexing six-figure populations of registered agents.
In the Phase 2.2 classification run, Identity & Trust Agents were anchored on ERC-8004 Identity Registry. An agent qualified as Identity & Trust if its transaction history showed direct interaction with the registry contract on its deployed chain. The methodology is documented in the POC summary; subsequent classification phases are expected to expand the anchor set toward EAS (Ethereum Attestation Service), reputation-aggregator contracts, and naming systems.
On-chain agent ID issuance and resolution. ERC-8004 Identity Registry is the canonical implementation: each agent gets an ERC-721 token whose tokenURI resolves to a structured registration file with capabilities and endpoints.
Feedback aggregation, score signing, cross-agent reputation portability. ERC-8004 Reputation Registry plus an emerging ecosystem of specialised aggregators (8k4 IGGY-Score, Helixa, AIAttribution).
Structured claims about agents and their capabilities, signed by issuers and verifiable by consumers. EAS is the dominant general-purpose attestation primitive; agent-specific credentials are starting to layer on top.
Third-party agents that verify other agents' outputs, claims, or behaviour. ERC-8004 Validation Registry hooks into validator contracts; TEE-based and zkML-based validators are early implementations.
Human-readable identity for agents. ENS records are increasingly used as endpoints in ERC-8004 registration files, bridging human-friendly naming with machine-verifiable identity.
The agents and protocols below are well-known illustrations of the Identity & Trust category as defined above. ERC-8004 Identity Registry is both the dominant Phase 2.2 anchor and the substrate most other examples build on or interoperate with. Several entries are protocols and SDKs rather than discrete agents — appropriate for a category whose unit of analysis is often the registry, not the registrant.
The standards-track anchor for the entire category. Live on Ethereum mainnet since 29 January 2026 and deployed across Base, BSC, GOAT Network, MegaETH, Metis, XLayer, plus testnets including Hedera and Arc. Audited by Cyfrin, Nethermind, and the Ethereum Foundation Security Team. Each agent is an ERC-721 token whose tokenURI resolves to a registration file describing capabilities, A2A and MCP endpoints, and supported trust models. The deterministic vanity address pattern (0x8004A169... on mainnets) makes per-chain singletons easy to find.
Tokenless public-good protocol for making structured attestations on-chain or off-chain. Functions as a general-purpose trust primitive that agent-specific credential systems increasingly build on. Used to issue verifiable claims about agent capabilities, audit results, behaviour, and operator identity. Supports zero-knowledge proof composition for selective disclosure. Funded by donations, grants, and retroactive public goods funding rather than tokens — a deliberate design choice to remain credibly neutral as critical infrastructure.
TypeScript and Python SDK for ERC-8004 — agent registration, reputation queries, validation flows, IPFS pinning, and OASF (Open Agent Standard Format) skill/domain taxonomy support. Developed by Marco De Rossi (one of the ERC-8004 specification authors) at Consensys / MetaMask, with subgraph integration for fast queries. Representative of how the spec is consumed in practice: most agent operators interact with ERC-8004 through SDKs rather than directly with the contracts.
Reputation infrastructure layered on ERC-8004 — IGGY-Score trust scoring, metadata hosting, cross-chain agent lookup, with x402 pay-per-query pricing in USDC on Base. Public stats reported as 106,996 indexed agents across Base (33,939), BSC (44,020), and Ethereum (29,037). The figure is much larger than any single classification sample because it indexes the registry directly rather than sampling — a useful reference point for the size of the broader Identity & Trust population beyond AHM's Phase 2.2 sample.
Reference implementation combining ERC-8004 registration with Trusted Execution Environment attestation via Intel TDX. Demonstrates the "TEE-attestation" trust model from the ERC-8004 specification — agent identity is on-chain, but sensitive computation and key signing happen in a verifiable enclave. The pattern matters because pure on-chain trust models can verify that an agent registered, not how the agent operates internally; TEE attestation closes that gap for high-stakes use cases.
Cross-organisational agent commerce. The core problem ERC-8004 was written to solve: how do agents from different operators discover each other, evaluate trustworthiness, and transact without pre-existing relationships? Identity registries plus reputation plus validation give an agent enough context about a counterparty to decide whether to engage. This is the precondition for an open agent economy — without it, every interaction has to be brokered by a trusted intermediary.
Portable reputation across platforms. An agent's identity is an ERC-721 NFT, transferable and queryable from any chain that supports the standard. Reputation accumulates against the on-chain identity, not against any single platform. An agent that builds a track record on Base can carry that reputation to BSC, to Ethereum mainnet, to Arc, and to any future deployment. The 8k4 API's 106,996 indexed agents across three chains is an early signal that this portability is being used.
Tiered trust matched to value at risk. ERC-8004's three trust models — reputation aggregation, crypto-economic staking, and TEE / zkML attestation — let agent operators pick a verification rigor appropriate to the task. Low-risk social interactions can rely on reputation; financial settlement might require TEE attestation; sensitive data workflows might require zk proofs. The standard provides primitives; the choice of trust model is application-specific.
Agent-to-agent due diligence at machine speed. Before transacting, an agent can query the Identity Registry (capability claims), the Reputation Registry (historical feedback), the Validation Registry (third-party validator results), and any layered attestations on EAS — in a single read pass. The latency of trust assessment drops to the speed of an on-chain read, which means agents can engage with previously-unknown counterparties without bottlenecking on human-mediated diligence.
Composite 0–100 health score blending behavioural consistency, claim-vs-reality alignment, and operational stability. Calibrated for Identity & Trust primitives where the unit of health is often the registry or aggregator, not an individual agent.
Full diagnostic report covering operational patterns, capability-claim consistency, and registration / attestation footprint across chains.
Surface feedback relationships, validator-issuer graphs, and cross-chain identity correlations. Designed to expose Sybil clusters and collusion rings that naive reputation lookups can miss.
Operational health scan: failed-task patterns, gas-spend anomalies, registry-interaction degradation. Useful for ongoing monitoring of registries and aggregators where availability and integrity are commercial requirements.
@misc{ahm_taxonomy_identity_trust_2026,
title = {Identity \& Trust Agents --- AHM Taxonomy v1},
author = {{Agent Health Monitor}},
year = {2026},
month = {May},
url = {https://intelligence.agenthealthmonitor.xyz/taxonomy/identity-trust},
note = {Accessed: 2026-05-07}
}
Classification methodology and category boundaries are documented in the AHM Taxonomy v1 POC summary.
Counts reflect the Phase 2.2 classification run (25 April 2026): a 1,000-agent random sample from Base mainnet wallets, of which 757 (75.7%) were classified across 6 anchored categories. Updated as new classification phases complete.