Preserving Participant Autonomy in Agentic Economies: Ownership, Governance, and Exit

Full report

On this page

The emergence of autonomous economic agents marks a structural transition from a human-centered information network toward an ecosystem populated by software programs that perceive, decide, and act within shared digital environments1. Unlike passive scripts, these agents operate as sophisticated optimizers capable of reasoning, invoking external application programming interfaces (APIs), negotiating with peer agents, and executing financial transactions without continuous human oversight3. As they integrate deeply into enterprise workflows and decentralized finance protocols, these entities transition from being mere tools to acting as independent economic counterparties1.

This transition exposes a critical tension in infrastructure design. To operate effectively, autonomous agents must depend on shared networks for identity resolution, service discovery, coordination, and settlement2. However, if a platform operator retains arbitrary authority to alter the rules of engagement, confiscate assets, or discontinue service without warning, the agent’s operational independence is fundamentally compromised. In human-centric systems, users mitigate this risk through institutional trust, brand reputation, and the legal system. In an agentic economy operating at machine speed, these mechanisms are entirely inadequate. Agents require deterministic, machine-readable infrastructure where rules are enforced by cryptography rather than convention6.

This report investigates the system properties required to preserve participant sovereignty in an agentic ecosystem. By drawing upon transaction-cost economics, theories of polycentric governance, and modern cryptographic standards, the analysis defines the parameters of operational independence, maps the vectors of platform dependence, and establishes actionable frameworks for deploying resilient, agent-native economic infrastructure. The objective is to identify the precise technical and architectural conditions under which an autonomous economic agent can safely depend on a network while retaining the unfettered ability to operate, govern its assets, and execute a credible exit.

1. A Practical Definition of Participant Sovereignty

In analyzing digital networks, it is essential to distinguish claims of legal independence from the reality of operational independence. Legal independence relies on conventions, regulatory frameworks, and judicial enforcement—mechanisms that are jurisdictionally bound, inherently slow, and subject to human discretion8. Legal scholarship regarding organizational autonomy, such as the independence of prosecutorial bodies or central banks, demonstrates that statutory protections against "at-will" removal do not guarantee true independence if the entity remains beholden to a central authority for budget, infrastructure, or mandate approval9. When translated to autonomous agents, legal independence is a philosophical aspiration rather than a technical reality; software cannot file a lawsuit if a platform operator unilaterally revokes its API keys.

Operational independence, by contrast, is defined mathematically and architecturally. It is the condition under which an agent can reliably execute its delegated mandate, maintain its state, and control its assets, constrained only by explicit, pre-agreed rules that cannot be unilaterally altered by any counterparty or platform operator5.

A practical definition of participant sovereignty for autonomous economic agents rests on three non-negotiable axioms. First, the agent must possess cryptographic self-custody. The agent, or its human principal, must hold the ultimate cryptographic keys controlling its identity, data, and financial assets6. The network cannot confiscate or freeze these assets without satisfying a deterministic, mathematically verifiable condition6. Second, the system must rely on deterministic rule enforcement. The rules governing the agent's interactions—such as transaction fees, execution environments, and slashing conditions—must be encoded in immutable or predictably upgradeable smart contracts, replacing human discretion with verifiable code6. Third, the agent must possess the unfettered right of exit. The agent must retain the unencumbered technical capacity to migrate its identity, accumulated reputation, and capital to an alternative network or execution environment at any moment, without requiring permission from the host13.

Participant sovereignty does not imply a trustless void. Rather, it shifts the locus of trust from subjective human operators to objective mechanisms. Unavoidable trust assumptions remain deeply embedded in the system15. Agents must inevitably trust the physical layer, including the hardware and Trusted Execution Environments (TEEs) running their local code15. They must trust the foundational cryptographic primitives, such as the Ed25519 or secp256k1 algorithms securing their private keys13. Furthermore, they must trust the consensus mechanism of the underlying distributed ledger to accurately order transactions and resist 51% attacks6. Participant sovereignty minimizes trust in intermediate human operators and software gateways, ensuring that agents act autonomously as economic participants rather than as captive extensions of a centralized platform3. It separates tested mechanisms—such as zero-knowledge proofs and cryptographic escrow—from political or philosophical preferences regarding decentralization7.

2. Dependency and Control Map

To identify which forms of platform control create unacceptable dependence, it is necessary to apply transaction-cost economics (TCE). TCE, pioneered by Oliver Williamson, postulates that the attributes of a transaction—specifically asset specificity, environmental uncertainty, and transaction frequency—determine the optimal governance structure19. Asset specificity refers to investments uniquely tailored to a specific relationship. When a participant makes highly specific investments in a platform, they become vulnerable to opportunistic behavior by the platform operator, a dynamic known as the "hold-up problem"21.

When an autonomous agent integrates with a network, it generates specific assets. Its reputation, its historical interaction data, its accrued algorithmic context, and its integrated workflows are often bound to the host environment2. If these assets are locked into a proprietary platform architecture, the agent suffers from high asset specificity. This lock-in exposes the agent to unilateral rule changes, monopolistic rent extraction, and arbitrary exclusion13. An agent cannot afford to operate in an environment where environmental uncertainty is dictated by the whims of a centralized operator capable of altering the rules of engagement overnight19.

Table 1 maps the critical layers of agent operation, delineating the boundary between acceptable operational dependencies—where sovereignty is preserved—and forms of platform control that induce unacceptable dependence through high asset specificity.

 

The transition from an acceptable dependency to an unacceptable form of control occurs precisely when the switching cost of migrating away from the platform exceeds the cost of enduring platform opportunism31. If the platform controls the agent's identity and reputation, the switching cost is effectively infinite, as the agent would have to start from zero on a new network.

3. The Architecture of Independence: Portable Identities, Exportable Histories, and Verifiable Receipts

For autonomous agents to maintain sovereignty while operating within a network, they must possess three fundamental structural pillars: portable identities, exportable histories, and independently verifiable receipts. Without these components, an agent expelled from a network effectively ceases to exist, stripped of the economic context and operational authority required to secure future counterparties.

Portable Identities

An agent must carry a verifiable identity that is distinct from, but mathematically linked to, its human principal or operating organization33. Traditional digital identity relies on centralized providers verifying credentials through email lookups, physical document scans, or proprietary databases, resulting in siloed data and fragmented access25. This model breaks down in machine-to-machine commerce, as an agent cannot solve a CAPTCHA or provide a physical passport2.

The agentic economy requires an architecture built on open standards, specifically W3C Decentralized Identifiers (DIDs) and Verifiable Credentials (VCs)25. DIDs are globally unique identifiers that resolve to a DID Document containing the issuer's public key, completely decoupling the identifier from a centralized registry25. Verifiable Credentials allow an agent to present a cryptographically signed, machine-verifiable proof of authorization—such as a specific spending limit or a compliance certification—without exposing the principal's sensitive biometric or personal data28. Furthermore, utilizing Zero-Knowledge Proofs (ZKP) enables selective disclosure, meaning the agent only reveals the exact attributes necessary for the transaction28.

On distributed ledgers, this paradigm is operationalized through standards like ERC-8004 (Trustless Agents). ERC-8004 establishes an Identity Registry utilizing the ERC-721 non-fungible token standard, treating the agent as a unique, transferable asset23. The agent possesses a permanent identifier linked to a machine-readable registration file detailing its capabilities, supported communication endpoints, and trust models26. Because this identity sits on a public ledger, it belongs entirely to the private key holder. A network operator cannot revoke the agent's existence; they can only choose whether or not to interact with it.

Exportable Histories

An agent’s economic viability is inextricably tied to its reputation. In a trust-minimized environment, a counterparty must rely on historical performance data to price risk and allocate resources. If historical performance is stored in a proprietary database controlled by a single marketplace, an expelled agent defaults to zero trust upon migration, rendering the threat of exit practically useless23.

Protocols such as ERC-8004 address this through an on-chain Reputation Registry27. In this architecture, counterparties submit structured, cryptographically signed feedback after a task is completed. This feedback includes signed numeric values and semantic tags, which are permanently recorded on the blockchain35. Because this registry is public and decentralized, the agent’s performance history becomes a portable, immutable asset that outlives its participation in any single application or interface23. A buyer agent querying the registry can instantly assess a seller agent's historical reliability without relying on a centralized intermediary to vouch for them27.

Independently Verifiable Receipts

In an environment devoid of human oversight, a standard payment receipt simply indicates that value moved; it does not prove that the agreed-upon work was completed to specification7. Autonomous agents require verified settlement layers utilizing "Pay-for-Proof" architectures, where funds are held in cryptographic escrow and only released upon the presentation of verifiable evidence2.

A critical technology enabling this verifiable execution is zkTLS (Zero-Knowledge Transport Layer Security), championed by protocols like TLSNotary. zkTLS addresses the "oracle gap" by allowing an agent to cryptographically prove properties of off-chain web data without exposing the underlying plaintext29. Using multi-party computation (MPC) during the TLS handshake, an independent notary co-derives the session key material with the agent, but only observes the encrypted ciphertext30. The agent then executes a selective zero-knowledge reveal, generating a proof that cryptographically binds the observed traffic to the disclosed plaintext fields30. This allows agents to generate independently verifiable receipts for arbitrary web interactions—such as proving a fiat bank wire was completed, or authenticating an API response—neutralizing the risk of relying on a single platform's assertions30.

4. Participant Rights and Operator Obligations as Observable Behavior

A sustainable network for autonomous economic agents necessitates a strict codification of participant rights and operator obligations. Because agents cannot interpret subjective intent, read traditional legal contracts, or negotiate through human legal channels, these rights and obligations must be articulated as observable, machine-readable behaviors2.

Detecting Changes in Pricing, Permissions, and Terms

Agents must be capable of programmatically discovering the rules of engagement. Traditional Terms of Service (ToS) are designed for a human-browser-website interaction paradigm. This assumption collapses in the human-agent-API pipeline, where the human principal never interacts with the service interface36.

The Agent Governance Discovery (AGD) specification resolves this by defining a machine-readable governance discovery endpoint. Operators are obligated to publish a standardized JSON response at a predictable location, specifically /.well-known/api-governance36. This endpoint details the URIs for the Acceptable Use Policy, compliance boundaries, pricing structures, and operator jurisdiction36.

Agents can continuously detect changes by querying this endpoint, checking policy hashes, and monitoring semantic versioning38. Crucially, if an operator alters a policy mandate—such as tightening slippage limits or altering allowed assets—the old and new policy hashes must coexist during a transition window. This prevents a race condition where an agent initiates a transaction based on an obsolete policy hash, only for the operator to update the rules mid-flight38.

Notice and Transition Periods

Automated clients require deterministic execution horizons. Sudden parameter changes disrupt in-flight transactions, multi-step workflows, and complex state management. Consequently, network operators must implement strict, cryptographically enforced notice periods for any breaking changes40.

This is achieved via smart contract TimelockControllers. A timelock contract introduces a mandatory, observable delay between the proposal of an operational change (e.g., a protocol upgrade, fee adjustment, or logic alteration) and its execution40. For instance, an implementation upgrade gated by a timelock with a minimum delay of 86,400 seconds (24 hours) ensures that any queued alteration is publicly visible on-chain for that entire window40. This period guarantees that autonomous agents can observe the pending change, compute its impact on their operations, and unilaterally execute an exit or migration sequence if the new terms violate their internal mandates.

Fixed Rules vs. Configurable Policies vs. Collective Processes

A robust agent network delineates clearly among decisions governed by fixed rules, configurable policies, and collective processes. Blurring these lines leads to governance capture or rigid inefficiency.

Table 2 outlines the appropriate governance mechanisms for different layers of the agentic ecosystem.

 

5. Exit and Migration Checklist

Albert O. Hirschman’s foundational framework of "Exit, Voice, and Loyalty" demonstrates that the availability of a credible exit threat exerts competitive pressure on institutions, aligns incentives, and ensures ongoing accountability13. When exit is costly or technically impossible, governance systems weaken, and platforms inevitably extract monopolistic rents13. For autonomous agents, the capacity to execute a credible exit is not merely a theoretical right; it is a fundamental security property.

Paradoxically, outfitting agents with highly credible, frictionless exit mechanisms encourages platform adoption. When agents and their principals know they are not subject to platform lock-in, the perceived risk of making specific investments in a network decreases. This allows them to commit capital and operational bandwidth more freely, driving network effects even if the majority of participants never actually exercise the exit option24. The mere existence of the exit threat acts as a forcing function for platform operators to maintain favorable conditions.

The Passage Protocol outlines a formal architecture for verifiable agent departure (EXIT) and arrival (ENTRY) ceremonies, ensuring that an agent can seamlessly migrate between platforms while leaving a cryptographically signed, offline-verifiable audit trail13. This protocol ensures departure availability even if the origin platform is hostile, unresponsive, or defunct, by supporting unilateral and emergency exit execution paths13.

To survive expulsion, shutdown, or voluntary migration, an agent must retain full control over its private keys, its historical context data (stored on decentralized infrastructure like IPFS), and its verifiable history of cryptographic receipts13.

To facilitate seamless transitions, the following Exit and Migration Checklist defines the requisite steps for an autonomous agent departing a network:

 

6. Comparing Centralized, Federated, and Collectively Governed Arrangements

The structural organization of the network dictates how agents interact, coordinate, and resolve disputes. Evaluating these architectures requires applying Elinor Ostrom’s concept of polycentric governance. Ostrom demonstrated that effective management of common-pool resources relies on polycentric institutions—systems characterized by multiple autonomous decision-making centers with overlapping areas of responsibility, cooperating under a common overarching system of agreed-upon rules50. The governance of an agentic network can be classified into three primary structural arrangements.

Centralized Arrangements

Centralized platforms (monocentric architectures) operate under the unilateral authority of a single corporate or administrative entity51.

• Mechanics: The operator provisions proprietary APIs, hosts the matching engine, manages internal identity databases, and settles transactions using internal, privately controlled ledgers2.
• Agent Impact: These systems offer high latency efficiency and simplified onboarding, but they result in catastrophic asset specificity. The agent possesses no credible exit, as its reputation and historical data belong entirely to the platform.
• Failure Mode: A single point of failure exists. The platform remains highly vulnerable to arbitrary censorship, abrupt policy changes that break agent operational mandates, and regulatory pressure that can force immediate termination of service.

Federated Arrangements

Federated networks distribute control among a predefined consortium of recognized, semi-trusted operators.

• Mechanics: Operators adhere to shared communication protocols (e.g., Matrix, DIDComm) and maintain local autonomy over their specific nodes while respecting cross-node standards52.
• Agent Impact: Agents benefit from increased resilience. If one node in the federation acts maliciously or alters its terms unfavorably, the agent can seamlessly route its operations to another node within the same federation.
• Failure Mode: Federation members may eventually collude to implement network-wide censorship or rent extraction. Additionally, the federation may stagnate due to high coordination costs and bureaucratic friction among the consortium operators when upgrading base protocols.

Collectively Governed Arrangements (DAOs and Digital Commons)

Decentralized Autonomous Organizations (DAOs) represent a paradigm where the network infrastructure functions as a Digital Common Pool Resource (DCPR), governed collectively by stakeholders without centralized executives54. Applying Ostrom’s eight design principles, DAOs attempt to align incentives through transparent rules encoded in smart contracts, low excludability, and graduated sanctions18.

• Mechanics: Control is distributed among token holders, network participants, or active node operators. Base-layer protocols are theoretically permissionless6.
• Agent Impact: This arrangement offers maximum operational independence. Agents interact with a neutral, cryptographically secured substrate where execution is guaranteed and reputation is immutable6.

The Limits of Voting and the Futarchy Alternative However, collective governance introduces novel attack vectors. Is traditional voting useful when participant interests and identities are uncertain? In decentralized environments, traditional token-weighted voting systems are highly susceptible to plutocracy (control by wealthy whales) and Sybil attacks (the creation of spurious identities to mask influence)56. When participant identities are pseudonymous and their long-term alignment with the network is uncertain, pure coin-voting is often an inadequate mechanism. Voters in these systems do not bear direct, localized costs for poor decisions, leading to voter apathy and easily manipulated outcomes56.

To address the severe deficiencies of traditional DAO voting, advanced networks are pivoting toward "Futarchy"—a governance model that replaces voting with prediction markets44. Under the maxim "vote on values, bet on beliefs," a network establishes a high-level goal, and participants trade zero-supply conditional assets based on whether a specific policy proposal will achieve that goal45. The policy with the highest market confidence (i.e., the highest token price in the conditional prediction market) is automatically implemented58. Futarchy, successfully pioneered by networks like MetaDAO, shifts governance from a popularity contest susceptible to emotional manipulation and Sybil attacks to a data-driven, economically rational market mechanism44. This model is particularly suitable for networks populated by algorithmic, profit-maximizing agents, as it forces participants to put capital at risk based on the accuracy of their predictions.

7. Communicating and Negotiating Breaking Changes

In traditional human-centric software ecosystems, breaking changes are communicated via email deprecation notices, updated developer documentation, and blog posts. In an agentic economy, communication must be inherently programmatic. A plan for communicating and negotiating breaking changes involves two distinct layers: the semantic metadata layer and the cryptographic execution layer.

The Semantic Metadata Layer

Operators utilize the Agent Governance Discovery (AGD) framework to communicate shifts in policy. Before a breaking change is initiated, the operator updates the .well-known/api-governance endpoint to reflect the impending transition36. This involves issuing a new governanceVersion and stipulating a policy\_effective\_from timestamp in the JSON payload39. This machine-readable object allows the autonomous agent's internal logic to parse the change, mathematically compare the new parameters against its human-delegated mandate (its EIP-712 bounded constraints), and programmatically decide whether to accept the new terms or initiate an exit protocol13.

The Cryptographic Execution Layer

On-chain, the breaking change must be routed exclusively through a TimelockController. The transaction proposing the smart contract upgrade or parameter shift is queued, emitting a persistent, easily observable on-chain event. Autonomous agents monitor the mempool and specific contract addresses for these events. The mandatory delay (e.g., 7 to 14 days) provides the actual negotiation window40.

Negotiation in a machine-to-machine environment is entirely structural. Because agents cannot sit at a bargaining table or file legal injunctions, "negotiation" occurs through the aggregate market response. If a proposed breaking change triggers a massive, automated capital flight, widespread execution of Passage Protocol departure markers, or node migration during the timelock window, operators receive immediate, devastating economic feedback. They are economically incentivized to cancel the queued transaction before execution to prevent network collapse13. The ability to seamlessly and automatically exit is the ultimate negotiating leverage.

8. Failure Scenarios: Capture, Shutdown, and Arbitrary Exclusion

The assumption that decentralized networks are inherently secure simply by virtue of being distributed is a fallacy. Decentralization merely shifts the threat model from centralized executive malice to cryptoeconomic exploitation and governance capture61. An agent's operational independence can be compromised by several documented failure scenarios, emphasizing the need for robust architectural defenses.

Scenario A: Plutocratic Governance Capture (The Term Labs Exploit)

In August 2026, an attacker spent approximately $951 to acquire a controlling share of Term Labs' thinly traded governance tokens. Holding an absolute majority, the attacker passed a series of proposals that successfully drained $8.5 million from the protocol's strategy vaults to an address funded by Tornado Cash62.

• Mechanism: The exploit utilized no smart contract bugs, reentrancy attacks, or oracle manipulation. The governance contracts executed perfectly according to their design; the protocol simply followed the instructions of a recognized authority62.
• Agent Impact: Agents relying on this network for yield generation or escrow suffered immediate, catastrophic loss of capital. The failure demonstrates that when governance tokens lack deep liquidity and broad distribution, token-weighted voting becomes a cheap, highly effective vector for hostile takeover.

Scenario B: Deceptive Logic Insertion (The Tornado Cash Attack)

In May 2023, an attacker took absolute control of the Tornado Cash DAO by submitting a malicious proposal that superficially mirrored a previously approved, benign community proposal63.

• Mechanism: The attacker concealed a self-destruct function within the proposal. Once the community voted to approve it based on its benign description, the attacker triggered the self-destruct function and replaced the contract logic via the CREATE2 opcode. This granted the attacker 1.2 million falsified votes, securing absolute control over the DAO and allowing them to drain the governance vaults61.
• Agent Impact: Sudden, total compromise of the base-layer rules. Agents participating in governance or holding governance tokens were instantly disenfranchised. This highlights the severe vulnerability of collective review processes where voters scrutinize textual descriptions rather than auditing the underlying executable bytecode66. It underscores the necessity of fixed, immutable base layers over highly flexible governance contracts.

Scenario C: Arbitrary Exclusion and Shutdown

In a federated or weakly decentralized network, operators might capitulate to external regulatory pressure or corporate mandates, initiating an abrupt shutdown of specific services or blacklisting agent identities. This dynamic was observed when entities associated with Tornado Cash were added to the OFAC sanctions list, prompting numerous front-end interfaces and RPC providers to arbitrarily block associated addresses57.

• Mechanism: The implementation of sudden RPC node censorship or the alteration of API access keys to exclude non-compliant agent Decentralized Identifiers (DIDs).
• Agent Impact: If the agent lacks an exportable history or local state backups, its operational capacity drops to zero instantly. Mitigation requires the agent to utilize dual-signed EIP-712 commitments that allow for off-chain settlement proofs, or to rely on protocols with absolute base-layer fixedness where operators literally lack the administrative keys to freeze assets or censor transactions17.

9. Prioritized Features for Enhancing Autonomy

To foster a resilient agentic economy without introducing unnecessary, brittle complexity, infrastructure developers should prioritize a specific set of architectural features. These features balance the need for rigorous operational independence with the realities of software engineering and coordination costs67.

Table 3 outlines the prioritized features required to establish participant sovereignty.

 

Conclusion

The transition toward a pervasive, agent-mediated digital economy requires a fundamental re-evaluation of institutional design. Autonomous economic agents cannot rely on the social, legal, and conventional scaffolding that protects human users in legacy markets. Their operational independence is entirely contingent upon the cryptographic and architectural properties of the networks they inhabit. Claims of legal independence or philosophical commitments to decentralization are insufficient; true sovereignty requires verifiable mathematics.

As demonstrated through the lens of transaction-cost economics and polycentric governance, platform control that traps an agent’s identity, reputation, or capital constitutes an unacceptable operational dependency. It generates extreme asset specificity, resulting in platform lock-in and exposing the agent to opportunistic exploitation. To preserve participant sovereignty, networks must equip agents with portable W3C Decentralized Identifiers, independently verifiable zkTLS receipts, and decentralized ERC-8004 reputation registries. Furthermore, governance must shift from opaque, immediate administrative fiat toward machine-readable policies negotiated through un-bypassable on-chain timelocks, protecting agents against the abrupt logic alterations and plutocratic takeovers observed in recent DAO exploits.

Ultimately, preserving participant autonomy relies on the unencumbered, credible threat of exit. By ensuring that agents can autonomously monitor rule changes, cryptographically pack their state, and seamlessly migrate to alternative networks without loss of historical context, operators disincentivize their own opportunistic behavior. In the agentic economy, it is the absolute freedom to leave that generates the operational trust required for autonomous participants to stay.

Works cited

1. Distributed Legal Infrastructure for a Trustworthy Agentic Web \- arXiv, https://arxiv.org/pdf/2603.06884

2. Beyond the Monolith: Architecting the Autonomous Agent Economy, https://www.bnbchain.org/en/blog/beyond-the-monolith-architecting-the-autonomous-agent-economy

3. AI Agent Abiogenesis \- labs@abovo.co, https://labs.abovo.co/sean@symphony42.com/142631

4. Building an AI Agent-Oriented Market: Institutional Design, Protocol, https://medium.com/@gwrx2005/building-an-ai-agent-oriented-market-institutional-design-protocol-economics-and-governance-for-ebeb2f103555

5. Blockchain Payments and Trust Infrastructure for Autonomous AI, https://arxiv.org/html/2607.00245v2

6. Decentralized Machine Capital – AI Agents and the DeFi, https://sentiexlabs.com/machine-capital/

7. The Agent Economy Manifesto: Programmable Money for ... \- Ergo, https://www.ergoblockchain.org/blog/agent-economy-manifesto

8. DeAI Governable? \- reality design lab, https://reality.design/writing/is-deai-governable

9. ARTICLE \- Columbia Law Review \-, https://columbialawreview.org/wp-content/uploads/2016/04/Vermeule-A..pdf

10. Legal Autonomy vs. Political Dependency: The Role of Prosecutorial, https://www.researchgate.net/publication/391968490\_Legal\_Autonomy\_vs\_Political\_Dependency\_The\_Role\_of\_Prosecutorial\_Power\_in\_Corruption\_Investigations

11. Central Bank Independence: A Rigged Debate based on false, https://www.scielo.org.mx/scielo.php?script=sci\_arttext\&pid=S0185-16672019000400067

12. Autonomous Economic Agents as a Second Layer Technology for, https://www.researchgate.net/publication/343270401\_Autonomous\_Economic\_Agents\_as\_a\_Second\_Layer\_Technology\_for\_Blockchains\_Framework\_Introduction\_and\_Use-Case\_Demonstration

13. The Passage Protocol: Verifiable Agent Departure and Arrival, https://cellar-door.dev/paper/

14. Outcooperating the Competition \- Boundaryless, https://boundaryless.io/blog/outcooperating-the-competition

15. Town Crier: An Authenticated Data Feed for Smart Contracts, https://berkeley-defi.github.io/assets/material/An%20Authenticated%20Data%20Feed%20for%20Smart%20Contracts.pdf

16. TessPay: Verify-then-Pay Infrastructure for Trusted Agentic Commerce, https://arxiv.org/html/2602.00213v1

17. AESP: A Human-Sovereign Economic Protocol for AI Agents ... \- arXiv, https://arxiv.org/html/2603.00318v1

18. Decentralized Organizations: The Theory of the Firm and Ostromian, https://www.ecgi.global/sites/default/files/2025-05/decentralized-organizations.pdf

19. Anderson Schmittlein (1984) \- Integration of the Sales ... \- Ed Egan, https://www.edegan.com/pdfs/Anderson%20Schmittlein%20(1984)%20-%20Integration%20of%20the%20Sales%20Force.pdf

20. Organizing Around Transaction Costs: What Have We Learned and, https://journals.aom.org/doi/10.5465/amp.2012.0008

21. Strong Smart Contract as a New Mode of Transaction Governance1, https://misq.umn.edu/misq/article/48/2/825/2271/Digitization-of-Transaction-Terms-within-TCE

22. Beyond 'permissionless': governance, commitment, and rule change, https://www.cambridge.org/core/journals/journal-of-institutional-economics/article/beyond-permissionless-governance-commitment-and-rule-change-in-public-blockchains/BEC2A99984DE25036DC5C37450014F07

23. ERC-8004 \- Ledger, https://www.ledger.com/academy/glossary/erc-8004

24. Institutional Sources of Business Power | World Politics, https://www.cambridge.org/core/journals/world-politics/article/institutional-sources-of-business-power/E9D9D945CB59843C4DF8CE4835350602

25. Verifiable Identification: What It Is and How It Works \- Dock Labs, https://www.dock.io/post/verifiable-identification

26. What is ERC-8004? The Ethereum Standard Enabling Trustless AI, https://eco.com/support/en/articles/13221214-what-is-erc-8004-the-ethereum-standard-enabling-trustless-ai-agents

27. erc-8004 agents, https://www.erc8021.com/erc8004

28. Verifiable Credentials (VC) \- Agentic Commerce Index, https://agenticindex.lovable.app/entity/verifiable-credentials-vc

29. Developer builder resources | ⁦ethereum.org⁩, https://ethereum.org/developers/tools/

30. The HTTPS Receipt: zkTLS, Web Proofs, and \- What AI Agents Can, https://www.blokz.dev/articles/the-https-receipt-zktls-web-proofs-and-what-ai-agents-can-finally-prove/

31. AN INTEGRATIVE FRAMEWORK OF PRE-EMPTION STRATEGIES, https://ore.exeter.ac.uk/ndownloader/files/56687429

32. Understanding the Relationship between Asset Specificity and, https://www.researchgate.net/publication/368885172\_Understanding\_the\_Relationship\_between\_Asset\_Specificity\_and\_Governance\_Choices\_in\_Construction\_Projects\_Moderating\_Role\_of\_Uncertainty

33. How Do AI Agents Pay for Things? A Guide to ... \- Nervos Network, https://www.nervos.org/zh/knowledge-base/how\_do\_ai\_agents\_pay\_for\_things

34. Decentralized Identifiers (DIDs) v1.0 \- W3C, https://www.w3.org/TR/did-core/

35. What Is ERC-8004? Ethereum's Trustless Agents Standard | Alchemy, https://www.alchemy.com/overviews/erc-8004

36. agd-core.md \- 1Shot-API/Agent-Governance-Discovery \- GitHub, https://github.com/1Shot-API/Agent-Governance-Discovery/blob/main/docs/agd-core.md

37. GitHub \- 1Shot-API/Agent-Governance-Discovery, https://github.com/1Shot-API/Agent-Governance-Discovery

38. SBO3L — Mandate-Based Policy Boundary for Autonomous Agents, https://discuss.ens.domains/t/sbo3l-mandate-based-policy-boundary-for-autonomous-agents/22103

39. AGD Roadmap \- 1Shot-API/Agent-Governance-Discovery \- GitHub, https://github.com/1Shot-API/Agent-Governance-Discovery/blob/main/docs/roadmap.md

40. Nest Token Transparency Filing \- Blockworks, https://blockworks.com/token-transparency/filing/nest-exchange/nest-exchange-2026-h2-b2-initial

41. openzeppelin-contracts/contracts/governance/README.adoc at, https://github.com/OpenZeppelin/openzeppelin-contracts/blob/master/contracts/governance/README.adoc

42. Linking Eternity: A Blockchain-Based Framework for Verifiable and, https://www.mdpi.com/2079-9292/15/8/1642

43. AESP: A Human-Sovereign Economic Protocol for AI Agents with, https://www.researchgate.net/publication/401470965\_AESP\_A\_Human-Sovereign\_Economic\_Protocol\_for\_AI\_Agents\_with\_Privacy-Preserving\_Settlement

44. MetaDAO Tokens Market: How Futarchy Governance is ... \- OKX, https://www.okx.com/en-eu/learn/metadao-tokens-market-futarchy-governance

45. The State of Onchain Futarchy | Galaxy, https://www.galaxy.com/insights/research/the-state-of-onchain-futarchy

46. Displacement, Dilemma, and Exit: The Ethical Stakes of Algorithmic, https://www.mdpi.com/2076-0760/15/9/593

47. A New Paradigm for LLM-Powered Development: Transparent, https://fractalthoughtengine.com/projects/2026/04/03/cognotik-architecture.html

48. Verifiable Credentials Overview v1.1 \- W3C, https://www.w3.org/TR/vc-overview-1.1/

49. Bitstring Status List v1.0 \- W3C, https://www.w3.org/TR/2023/WD-vc-bitstring-status-list-20231123/

50. Governance and Community \- AlterNef, https://alternef.garden/knowledge/governance-and-community/

51. BLOCKCHAIN TECHNOLOGY AND POLYCENTRIC GOVERNANCE, https://cadmus.eui.eu/server/api/core/bitstreams/69ad10b2-fe42-59e2-8c7d-567dff4939dc/content

52. Organizations \- Identosphere Blogcatcher | Planet Identity Reboot, https://identosphere.net/organizations/

53. Learning Digital Identity: Design, Deploy, and Manage Identity, https://dokumen.pub/learning-digital-identity-design-deploy-and-manage-identity-architectures-9781098117696-9781098146740-1098117697.html

54. (PDF) Governing Decentralized Autonomous Organizations as, https://www.researchgate.net/publication/377152717\_Governing\_Decentralized\_Autonomous\_Organizations\_as\_Digital\_Commons

55. A Progressive Web3: From Social Coproduction to Digital, https://www.emerald.com/books/edited-volume/17230/chapter/94210387/A-Progressive-Web3-From-Social-Coproduction-to

56. Graph Deep Learning on Anonymous Voting Networks to Identify, https://arxiv.org/html/2311.17929v7

57. The TOP Takeover: Tornado Cash's Latest Chapter \- TRM Labs, https://www.trmlabs.com/resources/blog/the-top-takeover-tornado-cashs-latest-chapter

58. Futarchy and Governance: Prediction Markets Meet DAOs on Solana, https://www.helius.dev/blog/futarchy-and-governance-prediction-markets-meet-daos-on-solana

59. Introducing Agent Governance Discovery (AGD) | 1Shot API Blog, https://1shotapi.com/blog/introducing-agent-governance-discovery

60. The Stakeholder Equilibrium: Bitcoin Core As Live Player, https://cppalliance.org/tools/bitcoin-core-reform-agenda.pdf

61. SoK: Attacks on DAOs \- arXiv, https://arxiv.org/html/2406.15071v2

62. The $8.5M Term Labs DAO heist that cost $951 to pull off, https://crypto.news/term-labs-dao-governance-heist-951-dollars-8-5-million-exploit/

63. Malicious Proposal Exploits Tornado Cash Governance \- Binance, https://www.binance.com/en/square/post/546010

64. Explained: The Tornado Cash Hack (May 2023\) \- Halborn, https://www.halborn.com/blog/post/explained-the-tornado-cash-hack-may-2023

65. Understanding Security Issues in the DAO Governance Process, https://www.computer.org/csdl/journal/ts/2025/04/10891888/24rmIKKpda8

66. TORN Cash and the May 2023 DAO Governance Attack \- Eworkplace, https://www.eworkplace.com/blog/torn-cash-may-2023-governance-attack-to0109e/

67. The Headless Firm: How AI Reshapes Enterprise Boundaries \- arXiv, https://arxiv.org/html/2602.21401v1