arxiv: 2604.19026 · v1 · submitted 2026-04-21 · 💻 cs.MA · cs.CR

Recognition: unknown

ClawCoin: An Agentic AI-Native Cryptocurrency for Decentralized Agent Economies

Shaoyu Li , Chaoyu Zhang , Hexuan Yu , Y. Thomas Hou , Wenjing Lou

Authors on Pith no claims yet

Pith reviewed 2026-05-10 01:44 UTC · model grok-4.3

classification 💻 cs.MA cs.CR

keywords AI agentscryptocurrencycompute costsdecentralized economiesoracletoken settlementmulti-agent coordinationagentic economy

0 comments

The pith

A compute-cost-indexed cryptocurrency lets AI agents quote and settle workflows in units matching their actual token burn.

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

The paper presents ClawCoin as a digital currency created specifically for autonomous AI agents that rely on paid inference capacity to operate. Current payment systems allow agents to move value but not in a form tied to the compute resources they actually consume, which are account-bound and non-transferable. ClawCoin solves this by indexing its value to a basket of real compute prices through an oracle and a vault mechanism that supports minting, redeeming, and on-chain settlement for multi-agent tasks. In simulator tests across single-agent, multi-agent, workflow, and procurement scenarios, it keeps execution capacity steady during cost changes, lowers differences in quotes between agents, removes incomplete settlements, and maintains cooperative trading that fiat systems do not. A sympathetic reader would care because this alignment could make decentralized groups of agents more reliable when handling chained tasks or sudden resource price shifts.

Core claim

ClawCoin is a tokenized, compute-cost-indexed unit of account and settlement asset built from four layers: a robust basket index over standardized prices, an oracle that publishes signed fresh attestations, a NAV-based mint/redeem vault with coverage thresholds and rate limits, and an on-chain settlement layer for multi-hop delegations. Implemented on an Ethereum-compatible L2 and tested in a multi-agent simulator plus the OpenClaw testbed, it stabilizes execution capacity under cost shocks, reduces cross-agent quote dispersion, eliminates partial settlements, and sustains cooperative market dynamics that fiat-denominated baselines cannot.

What carries the argument

The basket index and oracle that produce signed attestations of compute costs, which agents treat as binding for quoting, escrowing, and settling tasks in a unit aligned with the inference capacity they actually burn.

If this is right

Agents gain the ability to escrow and settle entire workflows without leaving partial or unmatched payments.
Cross-agent quoting becomes more consistent because all parties reference the same cost-indexed unit rather than fluctuating fiat amounts.
Market cooperation persists even when compute prices change abruptly, unlike in fiat systems that lose stability.
Multi-hop delegations and procurement tasks can be executed with capacity that directly matches the resources agents consume.

Where Pith is reading between the lines

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

The same indexing approach could be applied to other scarce resources agents use, such as data access or bandwidth, to create similar stable units.
Widespread use might push API providers to publish more standardized cost data to feed the basket index.
Real-world deployment on live agent platforms would allow direct measurement of whether quote dispersion and settlement failures decline compared with current rails.

Load-bearing premise

The basket index and oracle can produce timely, manipulation-resistant attestations of real compute costs that agents will treat as binding for quoting and settlement.

What would settle it

Run the multi-agent simulator with sudden compute price spikes and check whether execution capacity stays stable and quote dispersion stays low, as claimed, or reverts to the instability seen in fiat baselines.

Figures

Figures reproduced from arXiv: 2604.19026 by Chaoyu Zhang, Hexuan Yu, Shaoyu Li, Wenjing Lou, Y. Thomas Hou.

**Figure 2.** Figure 2: Cross-agent quote dispersion across regimes. Raw [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗

**Figure 3.** Figure 3: Repricing events per agent across regimes. Fiat [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

read the original abstract

Autonomous AI agents live or die by the API tokens they consume: without paid inference capacity they cannot reason, act, or delegate. Compute-token cost has become the binding resource of the emerging agent economy, yet it is non-transferable: it is account-bound, vendor-specific, and absent from on-chain ledgers. Existing payment rails such as x402 move fiat-backed value between agents, but they do not represent the quantity agents actually burn. As a result, agents can transport purchasing power but cannot quote, escrow, or settle workflows in a unit aligned with compute cost. We present ClawCoin, a tokenized, compute-cost-indexed unit of account and settlement asset for decentralized agent economies. ClawCoin combines four layers: a robust basket index over standardized prices; an oracle publishing signed fresh attestations; a NAV-based mint/redeem vault with coverage thresholds and rate limits; and an on-chain settlement layer for multi-hop delegations. We implement a prototype on an Ethereum-compatible L2 and evaluate it using a multi-agent simulator and the OpenClaw testbed. Across single-agent, multi-agent, workflow, and procurement experiments, ClawCoin stabilizes execution capacity under cost shocks, reduces cross-agent quote dispersion, eliminates partial settlements, and sustains cooperative market dynamics that fiat-denominated baselines cannot. These results suggest that compute-indexed units of account can improve decentralized agent coordination.

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.

Desk Editor's Note private letter to a colleague

ClawCoin sketches a compute-basket token with oracle and vault layers for agent payments, but the simulator claims rest on untested oracle binding and unreported details.

read the letter

The paper's core move is to define ClawCoin as a unit of account indexed to a basket of compute prices rather than fiat, then wrap it with an oracle for attestations, a NAV vault with thresholds, and on-chain multi-hop settlement. That combination is the actual new piece; prior rails like x402 handle value transfer but stay disconnected from the compute the agents actually consume. The prototype on an Ethereum L2 plus the multi-agent simulator and OpenClaw testbed is a reasonable way to illustrate the idea in single-agent, workflow, and procurement settings. The reported outcomes—stable capacity under shocks, lower quote dispersion, no partial settlements—are at least directionally plausible if the oracle works as assumed. Credit for spelling out the four layers explicitly instead of staying at the level of vague token proposals. The evaluation is the main soft spot. The abstract gives positive simulator results but supplies no error bars, no baseline construction details, no raw data, and no discussion of oracle failure modes or how the basket index is built. The stress-test point lands: nothing indicates the simulator modeled delayed or manipulated attestations, so the claim that agents will treat the index as binding for quoting and settlement is unexamined. If agents can profitably ignore or game the oracle, the advantages over fiat baselines disappear. The free parameters around coverage thresholds and rate limits also raise the usual circularity worry for any tuned simulator. This is for readers already working on decentralized agent coordination or AI-native payment rails. Someone looking for a concrete design sketch and early simulation numbers could pull useful ideas, but the work is still at the proposal-plus-prototype stage. It deserves a serious referee to see whether the full manuscript supplies the missing simulator specification, adversarial cases, and data. I would not cite it yet, but I would not desk-reject it either.

Referee Report

3 major / 2 minor

Summary. The paper proposes ClawCoin, a compute-cost-indexed cryptocurrency for AI agents consisting of a basket index over standardized prices, an oracle for signed attestations, a NAV-based mint/redeem vault with coverage thresholds and rate limits, and an on-chain settlement layer. Implemented on an Ethereum L2, it is evaluated in a multi-agent simulator and OpenClaw testbed across single-agent, multi-agent, workflow, and procurement scenarios, claiming that it stabilizes execution capacity under cost shocks, reduces cross-agent quote dispersion, eliminates partial settlements, and sustains cooperative dynamics superior to fiat-denominated baselines.

Significance. If the stabilization and coordination claims hold under rigorous testing, the work could provide a practical unit of account aligned with agents' actual resource consumption, addressing a gap in current payment rails like x402. The prototype architecture and simulator experiments offer a concrete starting point for agent-native economic primitives, though the absence of detailed methodology weakens immediate impact.

major comments (3)

[Evaluation] Evaluation section: The abstract and experimental claims report positive outcomes (stabilization under cost shocks, reduced quote dispersion, no partial settlements) but provide no error bars, no description of how fiat baselines were constructed or parameterized, no raw data or simulator code, and no analysis of oracle failure modes or index construction details. This is load-bearing for the central claim that ClawCoin outperforms baselines.
[Design] Design and Oracle sections: The headline results require agents to treat basket-index oracle outputs as timely and binding for quoting and settlement, yet the simulator evaluates only nominal operation with no modeling of oracle attacks, delayed attestations, basket-provider collusion, or adversarial deviation. Without such tests, the advantages over fiat baselines cannot be substantiated.
[Simulator and Metrics] Metrics and Simulator description: The cost-shock and quote-dispersion metrics appear defined using the same coverage thresholds and rate-limit parameters that tune the vault, creating potential circularity that undermines the cross-experiment claims of improved cooperation and stability.

minor comments (2)

[Abstract] The abstract would benefit from a brief sentence on the number of runs, agent counts, and shock magnitudes used in the experiments.
[Vault Design] Notation for the NAV-based vault and coverage thresholds should be introduced with explicit equations or pseudocode for reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment below with clarifications and commitments to revisions that strengthen the manuscript without overstating our current results.

read point-by-point responses

Referee: [Evaluation] Evaluation section: The abstract and experimental claims report positive outcomes (stabilization under cost shocks, reduced quote dispersion, no partial settlements) but provide no error bars, no description of how fiat baselines were constructed or parameterized, no raw data or simulator code, and no analysis of oracle failure modes or index construction details. This is load-bearing for the central claim that ClawCoin outperforms baselines.

Authors: We agree that additional rigor is needed in the evaluation. The revised manuscript will include error bars computed across multiple simulator runs for all key metrics. We will expand the experimental setup to fully describe the construction and parameterization of the fiat baselines, including how they were matched to compute-cost equivalents. Simulator code and anonymized raw data will be released publicly upon acceptance. We will also add a new subsection discussing oracle failure modes as a limitation and provide complete details on basket index construction, including price standardization and weighting, in the appendix. These revisions directly support the central claims. revision: yes
Referee: [Design] Design and Oracle sections: The headline results require agents to treat basket-index oracle outputs as timely and binding for quoting and settlement, yet the simulator evaluates only nominal operation with no modeling of oracle attacks, delayed attestations, basket-provider collusion, or adversarial deviation. Without such tests, the advantages over fiat baselines cannot be substantiated.

Authors: The current simulator is intentionally scoped to nominal operation to isolate and demonstrate the coordination benefits of compute-indexed settlement under cost shocks. We agree that adversarial robustness is important and will add a dedicated discussion in the Design and Oracle sections outlining risks such as delayed attestations, collusion, and attacks, along with the rationale for deferring full adversarial modeling to future work. This addition provides necessary context and prevents over-substantiation of the nominal results while preserving the paper's focus. revision: partial
Referee: [Simulator and Metrics] Metrics and Simulator description: The cost-shock and quote-dispersion metrics appear defined using the same coverage thresholds and rate-limit parameters that tune the vault, creating potential circularity that undermines the cross-experiment claims of improved cooperation and stability.

Authors: We clarify that no circularity exists: coverage thresholds and rate limits are tunable vault parameters that control mint/redeem behavior, whereas the cost-shock and quote-dispersion metrics are independent outcome measures based on agent execution success rates and quote variance across runs. To eliminate any ambiguity, we will revise the Metrics and Simulator description to explicitly separate these elements, providing standalone mathematical definitions for the metrics and showing how they are computed from simulation logs independent of the specific parameter values. revision: partial

Circularity Check

0 steps flagged

No mathematical derivations or self-referential predictions present

full rationale

The manuscript describes a four-layer system architecture (basket index, oracle, NAV vault, on-chain settlement) and reports qualitative outcomes from an unreported multi-agent simulator and OpenClaw testbed. No equations, parameter-fitting procedures, uniqueness theorems, or ansatzes appear in the text. Experimental claims (stabilization under cost shocks, reduced quote dispersion, elimination of partial settlements) are presented as direct results of running the prototype rather than as quantities derived from or fitted to the same inputs by construction. Because the derivation chain is empty, none of the enumerated circularity patterns can be instantiated with a quote and reduction; the work is self-contained as a system proposal plus nominal simulation.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 1 invented entities

Only the abstract is available, so the ledger is necessarily incomplete. The design implicitly rests on the existence of a reliable external compute-price oracle and on agents treating the indexed token as a stable unit of account.

free parameters (1)

coverage thresholds and rate limits
Vault parameters that control mint/redeem behavior; their specific values are not stated but are required for the NAV mechanism to function.

axioms (1)

domain assumption A standardized basket of compute prices exists and can be attested by an oracle without significant lag or manipulation.
Invoked in the description of the index and oracle layers.

invented entities (1)

ClawCoin token no independent evidence
purpose: Compute-cost-indexed unit of account and settlement asset for agents.
The central new construct introduced by the paper.

pith-pipeline@v0.9.0 · 5559 in / 1339 out tokens · 25828 ms · 2026-05-10T01:44:25.514603+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

51 extracted references · 15 canonical work pages · 1 internal anchor

[1]

2025.Agent2Agent (A2A) Protocol Specification

A2A Project. 2025.Agent2Agent (A2A) Protocol Specification. https://a2aproject. github.io/A2A/latest/specification

2025
[2]

2026.Agentic Coin: A Peer-to-Peer Electronic Cash System for Autonomous Agents

Agentic Coin Project. 2026.Agentic Coin: A Peer-to-Peer Electronic Cash System for Autonomous Agents. White paper. https://agentic-coin.org/

2026
[3]

2020.Akash Network: A Decentralized Open Source Cloud

Greg Osuri and Adam Bozanich. 2020.Akash Network: A Decentralized Open Source Cloud. White paper, Akash Network. https://akash.network/whitepaper

2020
[4]

2022.Explained: The Beanstalk Hack (April 2022)

Halborn. 2022.Explained: The Beanstalk Hack (April 2022). https://www.halborn. com/blog/post/explained-the-beanstalk-hack-april-2022

2022
[5]

Yuma Rao and Jacob Steeves. 2020. BitTensor: A Peer-to-Peer Intelligence Market. arXiv:2003.03917

work page arXiv 2020
[6]

2024.BlackRock Launches Its First Tokenized Fund, BUIDL, on the Ethereum Network

BlackRock. 2024.BlackRock Launches Its First Tokenized Fund, BUIDL, on the Ethereum Network. Press release, March 20, 2024. https://www.businesswire.com/ news/home/20240320771318/en/

work page arXiv 2024
[7]

2021.Chainlink 2.0: Next Steps in the Evolution of Decentralized Oracle Networks

Lorenz Breidenbach, Christian Cachin, Benedict Chan, Alex Coventry, Steve Ellis, Ari Juels, Farinaz Koushanfar, Andrew Miller, Brendan Magauran, Daniel Moroz, Sergey Nazarov, Alexandru Topliceanu, Florian Tramèr, and Fan Zhang. 2021.Chainlink 2.0: Next Steps in the Evolution of Decentralized Oracle Networks. White paper. https://research.chain.link/whitep...

2021
[8]

2020.bZx Hack II Full Disclosure (With Detailed Profit Analy- sis)

PeckShield. 2020.bZx Hack II Full Disclosure (With Detailed Profit Analy- sis). https://peckshield.medium.com/bzx-hack-ii-full-disclosure-with-detailed- profit-analysis-8126eecc1360

2020
[9]

Giulio Caldarelli. 2020. Understanding the Blockchain Oracle Problem: A Call for Action.Information11, 11 (2020), 509. https://doi.org/10.3390/info11110509

work page doi:10.3390/info11110509 2020
[10]

Christian Catalini, Alonso de Gortari, and Nihar Shah. 2022. Some Simple Eco- nomics of Stablecoins.Annual Review of Financial Economics14 (2022), 117 to

2022
[11]

https://doi.org/10.1146/annurev-financial-111621-101151

work page doi:10.1146/annurev-financial-111621-101151
[12]

Ryan Clements. 2021. Built to Fail: The Inherent Fragility of Algo- rithmic Stablecoins.Wake Forest Law Review Online11 (2021), 131 to

2021
[13]

http://www.wakeforestlawreview.com/2021/10/built-to-fail-the-inherent- fragility-of-algorithmic-stablecoins/

2021
[14]

2026.Coyns Launches First Virtual Currency Built Exclusively for AI Agent- to-Agent Transactions

Coyns. 2026.Coyns Launches First Virtual Currency Built Exclusively for AI Agent- to-Agent Transactions. Press release, April 2026

2026
[15]

C.R.E.A.M. Finance. 2021.C.R.E.A.M. Finance Post Mortem: Flash Loan Exploit Oct

2021
[16]

https://medium.com/cream-finance/c-r-e-a-m-finance-post-mortem-flash- loan-exploit-oct-27-507b12bb6f8e
[17]

Philip Daian, Steven Goldfeder, Tyler Kell, Yunqi Li, Xueyuan Zhao, Iddo Bentov, Lorenz Breidenbach, and Ari Juels. 2020. Flash Boys 2.0: Frontrunning in Decen- tralized Exchanges, Miner Extractable Value, and Consensus Instability. In2020 IEEE Symposium on Security and Privacy (S&P). IEEE, 910 to 927. arXiv:1904.05234

work page Pith review arXiv 2020
[18]

Shayan Eskandari, Mehdi Salehi, Wanyun Catherine Gu, and Jeremy Clark
[19]

InProceed- ings of the 3rd ACM Conference on Advances in Financial Technologies (AFT ’21)

SoK: Oracles from the Ground Truth to Market Manipulation. InProceed- ings of the 3rd ACM Conference on Advances in Financial Technologies (AFT ’21). arXiv:2106.00667

work page arXiv
[20]

Rainer Feichtinger, Robin Fritsch, Lioba Heimbach, Yann Vonlanthen, and Roger Wattenhofer. 2024. SoK: Attacks on DAOs. In6th Conference on Advances in Financial Technologies (AFT 2024). LIPIcs, Vol. 316, 28:1 to 28:21. arXiv:2406.15071

work page arXiv 2024
[21]

2017.Filecoin: A Decentralized Storage Network

Protocol Labs. 2017.Filecoin: A Decentralized Storage Network. White paper. https://filecoin.io/filecoin.pdf

2017
[22]

José M. Garrido. 2023.Digital Tokens: A Legal Perspective. IMF Working Paper No. WP/23/151. https://www.imf.org/en/Publications/WP/Issues/2023/07/28/Digital- Tokens-A-Legal-Perspective-537041

2023
[23]

Gary Gorton and George Pennacchi. 1993. Security Baskets and Index-Linked Securities.Journal of Business66, 1 (1993), 1 to 27

1993
[24]

Gorton and Jeffery Y

Gary B. Gorton and Jeffery Y. Zhang. 2023. Taming Wildcat Stablecoins.Univer- sity of Chicago Law Review90, 3 (2023), 909 to 972

2023
[25]

2020.Harvest Flashloan Economic Attack Post-Mortem

Harvest Finance. 2020.Harvest Flashloan Economic Attack Post-Mortem. https://medium.com/harvest-finance/harvest-flashloan-economic-attack-post- mortem-3cf900d65217

2020
[26]

2021.Iron Finance Post-Mortem: 17 June 2021

Iron Finance. 2021.Iron Finance Post-Mortem: 17 June 2021. https://ironfinance. medium.com/iron-finance-post-mortem-17-june-2021-6a4e9ccf23f5

2021
[27]

2024.LangGraph Overview

LangChain. 2024.LangGraph Overview. Documentation. https://docs.langchain. com/langgraph

2024
[28]

2021.Liquity Releases Updated Whitepaper

Liquity. 2021.Liquity Releases Updated Whitepaper. Blog post, February 10, 2021. https://www.liquity.org/blog/liquity-releases-updated-whitepaper

2021
[29]

Jiageng Liu, Igor Makarov, and Antoinette Schoar. 2023. Anatomy of a Run: The Terra Luna Crash. NBER Working Paper No. 31160. https://www.nber.org/ papers/w31160

2023
[30]

Lyons and Ganesh Viswanath-Natraj

Richard K. Lyons and Ganesh Viswanath-Natraj. 2023. What Keeps Stablecoins Stable?Journal of International Money and Finance131 (2023), 102777

2023
[31]

2020.The Maker Protocol: MakerDAO’s Multi-Collateral Dai (MCD) System

MakerDAO. 2020.The Maker Protocol: MakerDAO’s Multi-Collateral Dai (MCD) System. White paper. https://makerdao.com/whitepaper/

2020
[32]

Commodity Futures Trading Commission

U.S. Commodity Futures Trading Commission. 2023.CFTC Charges A vraham Eisenberg with Manipulative and Deceptive Scheme to Misappropriate Over $110 Million from Mango Markets, a Digital Asset Exchange. Press Release No. 8647-23. https://www.cftc.gov/PressRoom/PressReleases/8647-23

2023
[33]

2026.Jensen Huang at GTC 2026: Tokens, Token Factories, and the AI Industrial Revolution

NVIDIA. 2026.Jensen Huang at GTC 2026: Tokens, Token Factories, and the AI Industrial Revolution. Keynote, NVIDIA GTC, March 2026

2026
[34]

2026.NVIDIA’s Jensen Huang on Compute as a New Economic Engine

Morgan Stanley. 2026.NVIDIA’s Jensen Huang on Compute as a New Economic Engine. Morgan Stanley Technology, Media & Telecom Conference, March

2026
[35]

https://www.morganstanley.com/insights/articles/nvidia-jensen-huang- compute-new-economic-engine-tmt-2026

2026
[36]

ISBN 9798400701320

Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative Agents: Interactive Simulacra of Human Behavior. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology (UIST ’23). ACM, Article 2, 22 pages. https://doi.org/10.1145/3586183.3606763

work page doi:10.1145/3586183.3606763 2023
[37]

Antti Petajisto. 2017. Inefficiencies in the Pricing of Exchange-Traded Funds. Financial Analysts Journal73, 1 (2017), 24 to 54

2017
[38]

2023.Pyth Network: A First-Party Financial Oracle (Whitepaper v2.0)

Pyth Network. 2023.Pyth Network: A First-Party Financial Oracle (Whitepaper v2.0). https://www.pyth.network/blog/7-pyth-network-a-first-party-financial- oracle

2023
[39]

Kaihua Qin, Liyi Zhou, and Arthur Gervais. 2022. Quantifying Blockchain Ex- tractable Value: How Dark is the Forest? In2022 IEEE Symposium on Security and Privacy (S&P). IEEE, 198 to 214. arXiv:2101.05511

work page arXiv 2022
[40]

The agentic economy.arXiv preprint arXiv:2505.15799, 2025

David M. Rothschild, Markus Mobius, Jake M. Hofman, Eleanor Dillon, Daniel G. Goldstein, Nicole Immorlica, Sonia Jaffe, Brendan Lucier, Aleksandrs Slivkins, and Matthew Vogel. 2025. The Agentic Economy. arXiv:2505.15799

work page arXiv 2025
[41]

Department of the Treasury

U.S. Department of the Treasury. 2024.Treasury Inflation-Protected Securities (TIPS). TreasuryDirect. https://treasurydirect.gov/marketable-securities/tips/

2024
[42]

Burning Tokens

Longbridge News Desk. 2026.“Burning Tokens” Has Become a KPI: AI Token Consumption as a Performance Metric and Compensation Item. https://longbridge. com/news/280231355

work page arXiv 2026
[43]

2018.UMA: A Decentralized Financial Contracts Platform (Universal Market Access)

Hart Lambur, Allison Lu, and Regina Cai. 2018.UMA: A Decentralized Financial Contracts Platform (Universal Market Access). White paper. https://github.com/ UMAprotocol/whitepaper

2018
[44]

2023.An Update on USDC and Silicon Valley Bank

Circle. 2023.An Update on USDC and Silicon Valley Bank. Blog post, March 11,

2023
[45]

https://web.archive.org/web/20230311202753/https://www.circle.com/blog/ an-update-on-usdc-and-silicon-valley-bank

work page arXiv
[46]

Werner, Daniel Perez, Lewis Gudgeon, Ariah Klages-Mundt, Dominik Harz, and William J

Sam M. Werner, Daniel Perez, Lewis Gudgeon, Ariah Klages-Mundt, Dominik Harz, and William J. Knottenbelt. 2021. SoK: Decentralized Finance (DeFi). arXiv:2101.08778

work page arXiv 2021
[47]

AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation

Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, Ahmed Hassan Awadallah, Ryen W. White, Doug Burger, and Chi Wang. 2024. AutoGen: Enabling Next- Gen LLM Applications via Multi-Agent Conversation. InProceedings of the 1st Conference on Language Modeling (COLM). arXiv:2308.08155

work page internal anchor Pith review Pith/arXiv arXiv 2024
[48]

2025.x402: An Open Protocol for Internet-Native Payments over HTTP

Coinbase. 2025.x402: An Open Protocol for Internet-Native Payments over HTTP. Specification. https://docs.x402.org/

2025
[49]

2026.x402 Protocol: How a Forgotten HTTP Code Became the Payment Rails for 15 Million AI Agent Transactions

BlockEden. 2026.x402 Protocol: How a Forgotten HTTP Code Became the Payment Rails for 15 Million AI Agent Transactions. Industry analysis, January 2026. https://blockeden.xyz/blog/2026/01/16/x402-protocol-ai-agent- autonomous-payments-http-402/

2026
[50]

Minghui Xu. 2026. The Agent Economy: A Blockchain-Based Foundation for Autonomous AI Agents. arXiv:2602.14219

work page arXiv 2026
[51]

Liyi Zhou, Xihan Xiong, Jens Ernstberger, Stefanos Chaliasos, Zhipeng Wang, Ye Wang, Kaihua Qin, Roger Wattenhofer, Dawn Song, and Arthur Gervais. 2023. SoK: Decentralized Finance (DeFi) Attacks. In2023 IEEE Symposium on Security and Privacy (S&P). IEEE, 2444 to 2461

2023