pith. sign in

arxiv: 2606.12128 · v1 · pith:7QJPFL6Gnew · submitted 2026-06-10 · 💻 cs.CE

From Agent Identity to Agent Economy: Measuring the Operational Readiness of ERC-8004 AI Agents

Rischan Mafrur , Priagung Khusumanegara This is my paper

Pith reviewed 2026-06-27 07:48 UTC · model grok-4.3

classification 💻 cs.CE
keywords ERC-8004AI agentsblockchainEthereumoperational readinessagent economydecentralized AInetwork analysis
0
0 comments X

The pith

ERC-8004 registers AI agents on Ethereum but shows limited operational activity beyond identity.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper examines whether ERC-8004 agents on Ethereum are operationally ready beyond mere identity registration. Researchers built an agent-level feature table covering identity status, metadata, service declarations, reputation feedback, transfers, and cross-chain registration, then applied network analysis to owner-agent, feedback-client, wallet-transfer, and combined evidence relationships. They found registration visible at scale while metadata, services, reputation, and cross-chain evidence stay limited, with ownership and feedback activity highly concentrated among few wallets and clients. Richer evidence clusters around only a small subset of agents rather than spreading across the ecosystem. The results indicate that ERC-8004 supplies an identity layer for decentralized AI agents, yet the move to a functioning agent economy is still incomplete.

Core claim

Early ERC-8004 adoption is registration-heavy but operationally shallow. While the identity layer is visible at scale, metadata availability, service exposure, reputation formation, and cross-chain evidence remain limited. Ownership and feedback activity are also highly concentrated, suggesting that early participation is shaped by a small number of high-activity wallets and clients. The network analysis further shows that richer operational evidence clusters around a small subset of agents rather than being broadly distributed across the ecosystem.

What carries the argument

Operational readiness framework built from observable evidence layers (identity status, metadata, service declarations, reputation feedback, transfers, cross-chain registration) plus network analysis of owner-agent, feedback-client, wallet-transfer, and combined evidence relationships.

If this is right

  • The identity layer for decentralized AI agents is established at scale on Ethereum.
  • Metadata availability, service exposure, reputation formation, and cross-chain evidence remain limited in the current dataset.
  • Ownership and feedback activity are highly concentrated among a small number of high-activity wallets and clients.
  • Richer operational evidence clusters around a small subset of agents rather than distributing broadly.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Standards or market incentives may be required to push agents toward exposing services and accumulating reputation.
  • The observed concentration suggests the ecosystem remains in an early experimental stage dominated by a few active participants.
  • Repeating the analysis on later data or additional chains could test whether operational depth increases over time.

Load-bearing premise

The constructed agent-level feature table and observable evidence layers are sufficient and unbiased proxies for operational readiness.

What would settle it

Finding a broad distribution of agents with complete metadata, active service declarations, formed reputation feedback, and cross-chain registrations across many owners and clients would show the transition to an agent economy is more advanced than reported.

Figures

Figures reproduced from arXiv: 2606.12128 by Priagung Khusumanegara, Rischan Mafrur.

Figure 1
Figure 1. Figure 1: Operational readiness funnel for ERC-8004 agents. The figure shows how the sample narrows from identity registration to metadata, services, feedback, and cross-chain registration. The final category, full evidence, refers to agents that simultaneously have observable metadata, at least one service record, at least one feedback record, and at least one cross-chain registration record. 8 [PITH_FULL_IMAGE:fi… view at source ↗
Figure 2
Figure 2. Figure 2: Lorenz curves for ERC-8004 ownership and reputation concentration. Panel (a) shows the concentration of agent ownership across owner wallets, while Panel (b) shows the concentration of reputation feedback across feedback-client addresses. Together, the figures indicate that both identity ownership and reputation activity are unevenly distributed in the early ERC-8004 ecosystem. −0.6 −0.4 −0.2 0.0 0.2 0.4 0… view at source ↗
Figure 3
Figure 3. Figure 3: Standardized logistic regression coefficients associated with feedback formation. The model is exploratory and should be interpreted as evidence of association rather than causality. 10 [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Owner-to-agent network for ERC-8004 agents. Orange nodes represent owner wallets and blue nodes represent agent identities. The visualization shows that many agent identities are connected to a small number of high-volume owner wallets, indicating concentrated ownership in the early ERC-8004 ecosystem. 12 [PITH_FULL_IMAGE:figures/full_fig_p012_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Feedback concentration in the ERC-8004 reputation layer. Panel (a) shows the network between feedback clients and agents, where green nodes represent feedback client addresses and blue nodes represent agents. Panel (b) reports the most active feedback clients. Together, the figures show that reputation activity is concentrated around a small number of client addresses, suggesting that feedback formation re… view at source ↗
Figure 6
Figure 6. Figure 6: Wallet transfer network for ERC-8004 agent identities. Purple nodes represent wallets and directed edges represent observed transfer relationships. The figure captures ownership movement across wallets, but transfer activity should not be interpreted as proof of autonomous agent behavior. 15 [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Combined agent evidence network for ERC-8004 agents. The network links agents to owner wallets, feedback clients, service endpoint domains, and cross-chain namespaces. The figure shows that richer operational evidence is concentrated among a small subset of agents [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Declared service infrastructure among ERC-8004 agents with service records. Panel (a) reports the distribution of service types, showing that web endpoints dominate the observed service layer. Panel (b) reports the most common endpoint domains, showing that service declarations are concentrated around a small number of domains. Together, the figures indicate that the service layer is present but remains li… view at source ↗
read the original abstract

This paper examines whether blockchain-registered AI agents demonstrate operational readiness beyond identity registration. Using a dataset of ERC-8004 agents on Ethereum, we construct an agent-level feature table covering identity status, metadata, service declarations, reputation feedback, transfers, and cross-chain registration. We develop an operational readiness framework based on observable evidence layers and complement it with network analysis of owner-agent, feedback-client, wallet-transfer, and combined evidence relationships. The results show that early ERC-8004 adoption is registration-heavy but operationally shallow. While the identity layer is visible at scale, metadata availability, service exposure, reputation formation, and cross-chain evidence remain limited. Ownership and feedback activity are also highly concentrated, suggesting that early participation is shaped by a small number of high-activity wallets and clients. The network analysis further shows that richer operational evidence clusters around a small subset of agents rather than being broadly distributed across the ecosystem. The findings suggest that ERC-8004 provides an important identity layer for decentralized AI agents, but the transition from agent identity to agent economy remains incomplete.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. This paper examines whether blockchain-registered AI agents demonstrate operational readiness beyond identity registration. Using a dataset of ERC-8004 agents on Ethereum, the authors construct an agent-level feature table covering identity status, metadata, service declarations, reputation feedback, transfers, and cross-chain registration. They develop an operational readiness framework based on observable evidence layers and complement it with network analysis of owner-agent, feedback-client, wallet-transfer, and combined evidence relationships. The results show that early ERC-8004 adoption is registration-heavy but operationally shallow, with limited metadata availability, service exposure, reputation formation, and cross-chain evidence, plus high concentration in ownership and feedback activity. The network analysis shows richer operational evidence clustering around a small subset of agents. The findings suggest that ERC-8004 provides an important identity layer for decentralized AI agents, but the transition from agent identity to agent economy remains incomplete.

Significance. If the on-chain evidence layers accurately proxy operational readiness and economic activity, the work supplies a useful empirical snapshot of early ERC-8004 adoption and a structural view via network analysis of participation concentration. It establishes a baseline framework that could inform future measurements of decentralized AI agent ecosystems.

major comments (1)
  1. [Abstract and operational readiness framework] Abstract and operational readiness framework: The central claim that the transition from agent identity to agent economy remains incomplete rests on the constructed feature table (identity, metadata, services, reputation, transfers, cross-chain) serving as sufficient and unbiased proxies for operational readiness. No external validation, correlation to off-chain utility or activity, or analysis of potential biases from unindexed channels is reported, so the inference from observed on-chain shallowness to incomplete economic transition does not follow directly from the data.
minor comments (2)
  1. [Abstract] The abstract references results from a constructed dataset and network analysis but provides no sample size, data collection method, exclusion criteria, or statistical validation.
  2. [Abstract] Details on the precise thresholds or definitions used to classify evidence layers as 'limited' or activity as 'highly concentrated' are not stated.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive review. The central concern regarding the operational readiness framework and the strength of the inference to an incomplete agent economy transition is addressed below. We propose targeted revisions to clarify scope and limitations.

read point-by-point responses

  1. Referee: Abstract and operational readiness framework: The central claim that the transition from agent identity to agent economy remains incomplete rests on the constructed feature table (identity, metadata, services, reputation, transfers, cross-chain) serving as sufficient and unbiased proxies for operational readiness. No external validation, correlation to off-chain utility or activity, or analysis of potential biases from unindexed channels is reported, so the inference from observed on-chain shallowness to incomplete economic transition does not follow directly from the data.

    Authors: We agree that the manuscript does not include external validation against off-chain activity or explicit bias analysis for unindexed channels. The operational readiness framework is defined strictly in terms of observable on-chain evidence layers (identity registration, metadata, service declarations, reputation feedback, transfers, and cross-chain registration), as described in Sections 3 and 4. The central claim is therefore scoped to the visible on-chain transition rather than a comprehensive economic assessment. To strengthen the manuscript, we will revise the abstract to emphasize the on-chain scope, add an explicit limitations subsection in the discussion that acknowledges the absence of off-chain correlation and potential unindexed activity, and qualify the conclusion to state that the transition from identity to economy appears incomplete on the basis of on-chain observables. These changes will make the inferential limits transparent without altering the reported empirical findings. revision: yes

Circularity Check

0 steps flagged

No circularity: purely descriptive on-chain analysis with no derivations or fitted predictions

full rationale

The paper constructs an agent-level feature table from Ethereum ERC-8004 observables (identity, metadata, services, reputation, transfers, cross-chain) and applies network analysis to report concentration and shallowness. No equations, parameters, or predictions are defined; the operational readiness framework is explicitly built from the same observables used for measurement. No self-citations, uniqueness theorems, or ansatzes are invoked. Conclusions follow directly from the data without reduction to inputs by construction. This is standard descriptive empirical work and receives the default non-finding.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract-only view shows an empirical measurement study with no explicit free parameters, invented entities, or non-standard axioms; relies on standard assumption that blockchain event data captures agent activity.

axioms (1)
  • domain assumption Observable on-chain events and metadata fields are valid and complete indicators of operational readiness.
    Invoked when the paper states results are based on the constructed feature table covering identity, metadata, services, reputation, transfers, and cross-chain registration.

pith-pipeline@v0.9.1-grok · 5722 in / 1134 out tokens · 15092 ms · 2026-06-27T07:48:49.000535+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

22 extracted references · 8 canonical work pages

  1. [1]

    M. S. Al Jasem, T. De Clark, and A. K. Shrestha. Toward decentralized intelligence: A systematic literature review of blockchain-enabled ai systems.Information, 16(9):765,

  2. [2]

    URLhttps://doi.org/10.3390/info16090765

    doi: 10.3390/info16090765. URLhttps://doi.org/10.3390/info16090765

    work page doi:10.3390/info16090765

  3. [3]

    L. Ante. Autonomous ai agents in decentralized finance: Market dynamics, application areas, and theoretical implications.Technological Forecasting and Social Change, 228: 124669, 2026. doi: 10.1016/j.techfore.2026.124669. URLhttps://doi.org/10.1016/ j.techfore.2026.124669

    work page doi:10.1016/j.techfore.2026.124669 2026

  4. [4]

    L. Cao. Decentralized ai: Edge intelligence and smart blockchain, metaverse, web3, and desci.IEEE Intelligent Systems, 37(3):6–19, 2022. doi: 10.1109/MIS.2022.3181504. URLhttps://doi.org/10.1109/MIS.2022.3181504

    work page doi:10.1109/mis.2022.3181504 2022

  5. [5]

    T. J. Chaffer, I. Goins, Charles von, B. Okusanya, D. Cotlage, and J. Goldston. Decen- tralized governance of autonomous ai agents, 2024. URLhttps://arxiv.org/abs/24 12.17114. arXiv preprint arXiv:2412.17114

    arXiv 2024

  6. [6]

    Dafflon, B

    J. Dafflon, B. Breidenbach, and M. Kistner. Erc-6551: Non-fungible token bound ac- counts, 2023. URLhttps://eips.ethereum.org/EIPS/eip-6551. Ethereum Im- provement Proposal, ERC-6551

    2023

  7. [7]

    De Rossi, D

    M. De Rossi, D. Crapis, J. Ellis, and E. Reppel. Erc-8004: Trustless agents, 2025. URL https://eips.ethereum.org/EIPS/eip-8004. Ethereum Improvement Proposal, EIP-8004

    2025

  8. [8]

    Entriken, D

    W. Entriken, D. Shirley, J. Evans, and N. Sachs. Erc-721: Non-fungible token standard,

  9. [9]

    Ethereum Improvement Proposal, EIP-721

    URLhttps://eips.ethereum.org/EIPS/eip-721. Ethereum Improvement Proposal, EIP-721

  10. [10]

    Fan and T

    S. Fan and T. Min. Web3agent: Automating on-chain operations via natural language interfaces.ACM Transactions on the Web, 20(1):9, 2026. doi: 10.1145/3777446. URL https://doi.org/10.1145/3777446

    work page doi:10.1145/3777446 2026

  11. [11]

    T. Fan, Y. Yang, Y. Jiang, Y. Zhang, Y. Chen, and C. Huang. Ai-trader: Benchmarking autonomous agents in real-time financial markets, 2025. URLhttps://arxiv.org/ab s/2512.10971. arXiv preprint arXiv:2512.10971

    arXiv 2025

  12. [12]

    M. M. Karim, D. H. Van, S. Khan, Q. Qu, and Y. Kholodov. Ai agents meet blockchain: A survey on secure and scalable collaboration for multi-agents.Future Internet, 17(2): 57, 2025. doi: 10.3390/fi17020057. URLhttps://doi.org/10.3390/fi17020057

    work page doi:10.3390/fi17020057 2025

  13. [13]

    Y. Liu. A dataset of early blockchain-registered ai agents on ethereum, 2026. URL https://arxiv.org/abs/2604.22652. arXiv preprint arXiv:2604.22652

    Pith/arXiv arXiv 2026

  14. [14]

    Y. Liu. Replication data for: A dataset of the first 10000 blockchain registered ai agents on ethereum, 2026. URLhttps://doi.org/10.7910/DVN/HJZW8Q. 20

    work page doi:10.7910/dvn/hjzw8q 2026

  15. [15]

    R. Mafrur. Ai-based crypto tokens: The illusion of decentralized ai?IET Blockchain, 5: e70015, 2025. doi: 10.1049/blc2.70015. URLhttps://doi.org/10.1049/blc2.70015

    work page doi:10.1049/blc2.70015 2025

  16. [16]

    Ranjan, S

    R. Ranjan, S. Gupta, and S. N. Singh. Loka protocol: A decentralized framework for trustworthy and ethical ai agent ecosystems, 2025. URLhttps://arxiv.org/abs/25 04.10915. arXiv preprint arXiv:2504.10915

    arXiv 2025

  17. [17]

    Romandini, C

    N. Romandini, C. Mazzocca, K. Otsuki, and R. Montanari. Sok: Security and privacy of ai agents for blockchain. In2025 7th International Conference on Blockchain Computing and Applications (BCCA), pages 708–720. IEEE, 2025. doi: 10.1109/BCCA66705.2025 .11229689. URLhttps://doi.org/10.1109/BCCA66705.2025.11229689

    work page doi:10.1109/bcca66705.2025 2025

  18. [18]

    Tomaˇ sev, M

    N. Tomaˇ sev, M. Franklin, J. Z. Leibo, J. Jacobs, W. A. Cunningham, I. Gabriel, and S. Osindero. Virtual agent economies, 2025. URLhttps://arxiv.org/abs/2509.101

    2025

  19. [19]

    arXiv preprint arXiv:2509.10147

    arXiv

  20. [20]

    M. Xu. The agent economy: A blockchain-based foundation for autonomous ai agents,

  21. [21]

    arXiv preprint arXiv:2602.14219

    URLhttps://arxiv.org/abs/2602.14219. arXiv preprint arXiv:2602.14219

    arXiv

  22. [22]

    J. Yu, A. Zhao, and D. Sui. Paper agents, paper gains: An empirical analysis of defi investment agents, 2026. URLhttps://arxiv.org/abs/2605.29174. arXiv preprint arXiv:2605.29174. 21

    Pith/arXiv arXiv 2026