Attestations build trust for AI onchain coordination

A policy paper dated Aug. 6, 2026 argues that blockchain systems will need to evolve beyond recording transfers and executing smart contracts if they are to coordinate work between humans and autonomous AI agents. Its proposed answer is a “coordination layer” that uses verifiable records of contributions, known as attestations, to determine authority, reputation, rewards, and access to shared resources.

The paper, which describes a framework called the m&W Protocol, places this challenge in a setting where AI agents can be copied, deployed, and given access to digital assets at far lower cost than human organizations can verify their identities or evaluate their work. Wallet addresses and conventional online accounts, it says, are too static to establish accountability when software can create convincing activity histories at scale.

Under the proposal, an agent would receive only limited, contextual authority from a person or organization. Its actions would then be recorded with evidence and independently verifiable context, creating a history that could affect future permissions, task eligibility, governance influence, and economic rewards.

From Bitcoin’s trust model to programmable rules

The paper begins with Bitcoin’s origins during the 2008 financial crisis, citing Satoshi Nakamoto’s Oct. 31, 2008 email introducing a “new electronic cash system that’s fully peer-to-peer, with no trusted third party.”

Bitcoin’s proof-of-work system, longest-chain consensus, and economic incentives created a way for participants who did not know one another to agree on a shared transaction history. The document presents that architecture as a shift away from trust anchored primarily in banks, corporate intermediaries, and state institutions.

It describes a longer progression in how societies organize trust: kinship and local relationships in agricultural economies; contracts and centralized organizations in industrial economies; digital platforms and opaque server infrastructure during the internet era; and algorithmic consensus in blockchain networks.

Ethereum extended blockchain’s role from recording ownership and transfers to executing predefined conditions through smart contracts, according to the paper. That capability helped create decentralized finance protocols, decentralized autonomous organizations, and applications where rules could run on shared infrastructure rather than a company-controlled database.

Yet the document argues that smart contracts have a practical boundary. They can establish whether a transaction met coded conditions, whether a token changed hands, or whether an action occurred at a particular time. They cannot independently determine whether a contribution was useful, whether an agent acted in line with its creator’s goals, or whether a worker’s output had lasting value.

That gap becomes more consequential where work is collaborative and open-ended rather than a simple exchange of assets.

Attestations as records of contribution

The m&W Protocol centers on a structured contribution record the paper calls an attestation. It defines the record as:

Attestation = {Subject, Action, Context, Proof, Attester}

The subject identifies the person or agent responsible for an action. The action records what was done, while context describes the circumstances in which it occurred. Proof could include cryptographic evidence, a Git hash, or another verifiable artifact. The attester is the party or system that validates the record.

The distinction between proof and evaluation is central to the model. A Git hash, for example, can show that a particular code change was submitted. An attestation system would seek to attach additional context: who authorized the work, what task it addressed, which agent performed it, and whether a designated reviewer accepted it.

Such records could make a contribution portable across applications without relying entirely on a platform’s internal reputation system. They could also give organizations a clearer basis for assigning sensitive tasks to agents whose prior activity can be inspected and traced.

The proposal does not claim that every form of useful work can be reduced to an on-chain record. Instead, it presents attestations as a common format for documenting claims that can be checked by machines, communities, organizations, or other designated attesters.

Bounded delegation for autonomous agents

The paper describes its production model as a “builder + digital twin network,” in which human participants retain strategy, direction, and responsibility while software agents handle frequent execution across multiple tasks.

That division raises operational questions that conventional wallet ownership does not solve. An organization may need to determine whether an agent genuinely represents a specific builder, whether its actions fall within a granted mandate, what assets it can access, how long its authority lasts, and who bears responsibility if it exceeds its instructions.

The proposed identity layer would use wallet-based decentralized identity, or DID, alongside rights NFTs, contribution credentials, and soulbound tokens. Soulbound tokens are generally designed as non-transferable credentials, making them more suitable for recording affiliations or achievements than tradable assets.

A feature called “contextual delegation” would give agents limited permissions based on task scope, resource access, asset limits, time windows, and responsibility constraints. Rather than granting an agent unrestricted control over a wallet or treasury, a builder could authorize it to perform a specified job within defined limits.

That approach would place agent permissions closer to enterprise access controls while preserving blockchain-based records of who delegated authority and under what terms.

Trust graphs and contribution-based distribution

The protocol’s coordination layer would build a “computable trust graph” from accumulated attestations. The stated goal is to use verified histories to assess eligibility, match contributors or agents to tasks, and adjust permissions as their records develop.

Its ecofi layer would then route rewards, governance weight, resources, and collaboration opportunities based on documented contributions rather than token balances or short-term market performance.

That design challenges the common model in which governance power closely follows asset ownership. In a network built around agent activity, the paper argues that balance-based voting alone may offer little protection against cheaply created wallets, automated participation, or entities that can generate high transaction volumes without delivering useful work.

The paper summarizes its direction with the phrase, “AI creates intelligence, blockchain builds order.” It also uses the formula “AI + Blockchain/Web3 = 文明以止,” explaining the final character, “止,” as boundary-setting: systems for defining limits, accountability, and acceptable authority as machine actors take on a larger role in digital coordination.

To see how AI-driven automation is already reshaping real trading, explore our advanced AI copy trading guide today.

Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.