Building a Trust System in the Age of AI Trading

In every cycle of business development, intermediaries that can solve trust issues always capture the greatest value.

Written by: Prathik Desai

Compiled by: Chopper, Foresight News

More than a decade ago, I made my first online purchase with a credit card. My mother was very anxious, asking why I would trust a seller hiding behind a website whom I had never met. What if the action camera was never shipped? At that time, I couldn’t answer her. Now I have an answer.

At that time, the aggregator platform Flipkart (now owned by Walmart) intervened between ordinary sellers and me, providing a layer of trust assurance for the merchants.

Every business era expands the boundaries of transactions between strangers, and each era requires new intermediaries to facilitate trust. As David Graeber described in his book “Debt,” even before the birth of currency, trade among the Sumerians and Babylonians relied on credit and social relationships. When banks are unwilling to lend to strangers, credit agencies issue credit scores for individuals; after credit cards were introduced, Visa and MasterCard provided credit endorsements for every card transaction; when the internet gave rise to a multitude of anonymous sellers, eBay, Amazon, and Flipkart managed the trading ecosystem through rating systems.

Then came the era of machines.

This article will explain how to fill the trust gap as the volume of transactions conducted by machines is about to surpass that of humans.

Non-Human Buyers

The trust systems mentioned earlier are effective only when the transaction parties are humans or companies. If a corporate entity defaults, we can trace back to specific responsible individuals.

However, this logic does not apply to autonomous intelligent agents. Nowadays, machines complete purchases according to the instructions set by human principals. At the same time, AI agents are increasingly generating sub-agents autonomously, issuing procurement tasks to achieve overall goals.

These agents are merely a string of anonymous code and software, often making it difficult to trace the actual controllers behind them at first glance.

Agents can now book itineraries, negotiate prices, settle bills, and even invoke other software to handle tasks they cannot complete themselves. The entire process is extremely fast, leaving humans unable to monitor in real-time. By the time you finish reading this sentence, AI agents have already completed hundreds of transactions. Meanwhile, the scale of funds involved in these transactions continues to grow, making this issue impossible to ignore.

So, if an agent does something that does not comply with general trading rules, whose responsibility should you pursue? The identities of these agents can be changed at will. They can switch to a new identity and restart in a matter of seconds. Cryptocurrency is helping us solve this problem.

The Transferable Identity Dilemma

In January 2026, Ethereum launched the ERC-8004 standard, which can generate a permanent identity identifier for each agent on the public chain, while recording the tasks completed by the agent, the evaluations received, and the risk warnings issued. Anyone can publicly verify this record without the need for permission from a centralized institution.

This identity serves as both a passport and a credit record. However, it has a flaw: the standard issues agent identities in the form of transferable NFT smart contracts, which are what the cryptocurrency field refers to as ERC-721 tokens. The transferable attribute brings risks: agent identities can be transferred.

If identities can be sold, and reputation is tied to identity, then reputation will also flow along with it, leading to a series of hidden dangers.

For example: someone deploys an agent that operates stably and compliantly for six months, successfully completes tasks, accumulates positive reviews, and establishes a good credit score. Afterwards, they sell this identity at a high price to someone else, and the new holder directly inherits this reputation. The original deployer and principal can cash out and exit. If the new holder has malicious intent, they will exploit the inherited reputation to engage in violations.

How do we fill this loophole?

Identity and Trust Tech Stack

First, let’s look at the traditional trust systems that intelligent agents are about to replace.

The credit of merchants and online sellers is not permanently valid after a single certification; their credit is dynamically assessed based on ongoing behavior. A few disputes or fraud reports can limit their market access and directly damage their reputation. Online aggregator platforms often impose fines on involved merchants, require additional deposits, and even have the authority to permanently delist merchants.

Agents also need similar assessment mechanisms. Currently, most people simplify trust in agents to a scoring issue, believing that it is sufficient to measure indicators well. The real challenge lies in firmly binding the assessment results to the agents themselves, preventing credit from being traded or laundered.

A core prerequisite for an agent to have credibility is that the source of funds can be traced back to human principals.

Skyfire, which has received investments from Coinbase, a16z, and Circle, has launched an agent trust tech stack to ensure that trading counterparts are verifiable and trustworthy agents. Skyfire issues signed identity tokens for agents, and its cryptographic principles are equivalent to the verification mechanisms that bank websites use to prove their real identities to users.

Skyfire’s tokens bind three-party information: the platform that developed the agent, the agent itself, and the individual or enterprise acting as the principal. When an agent initiates a transaction, the seller can verify the token, and if any issues arise, the responsible party can be clearly identified.

Skyfire also includes a payment layer; this token used for identity verification can also be used for payment settlements.

As agent payments become widespread, the settlement layer will gradually homogenize, and the exploitable value space will narrow. Value will concentrate on the identity and trust layer, and a cryptographic “credit agency” will build a system to complete agent identity verification, credit endorsement, and transaction execution.

Many startups are adopting different solutions to seize the value of this track.

Skyfire associates trust with external principals, while RNWY chooses to attach credit directly to the agents themselves. RNWY employs a binding token (soulbound token) solution, inspired by a credit system that has been in existence for thirty years.

It cites a viewpoint from a paper in the “Cambridge Journal of Knowledge Engineering”: within the ecosystem, changing identities requires incurring high costs. This can prevent agents from easily changing identities to escape punishment after violations.

Soulbound tokens impose such cost constraints. Regular ERC-721 NFTs can flow freely between different wallets, while soulbound tokens issued based on the ERC-5192 standard eliminate the transfer function, permanently binding the tokens to the corresponding wallets, just like biometric information is bound to a passport.

Therefore, for agents to erase negative records, the only way is to abandon the wallet storing that token, along with all the accumulated reputation. This is the default cost under institutional design. This token standard fundamentally increases the cost of abandoning credit.

Currently, RNWY has registered over 230,000 agents based on this mechanism, all of which have completed identity registration through the ERC-8004 contract.

If agents want to start over from scratch, they can only abandon all historical records. RNWY deliberately raises the cost of “identity switching and running away”; only by incurring this cost can they obtain non-transferable and hard-to-manipulate credible qualifications. This mechanism can also constrain long-term compliant merchants, preventing subsequent violations.

However, this binding model is a double-edged sword: if the private key corresponding to the soulbound token is lost, the agent will lose its identity, credit, and all long-accumulated behavioral records.

ChainAware takes a different approach; it does not issue passes or identity certificates but reads on-chain behavioral data from wallets to generate wallet credit scores based on this data. Its model is trained to identify addresses associated with known fraud and historically clean addresses.

Its core logic is not to retrospectively trace issues but to predict risks in advance, similar to how bank risk control departments freeze bank cards after identifying abnormal transactions. Credit scores are entirely generated from real behaviors, and high-quality reputations cannot be obtained through purchase.

A number of developers are eager to explore the value of trust and identity layers, each with different routes: Skyfire anchors trust to principals; RNWY binds credit to non-transferable tokens; ChainAware scores agents directly based on wallet behavior, skipping the identity issuance stage.

I believe there is no single solution that can address all scenarios. The insights brought by these projects indicate that for the trust layer of agents to accumulate significant value, a complete tech stack must be built. The underlying layer must be able to lock in the responsibilities of principals, the middle layer must ensure that reputation and identity are non-transferable; on top of that, a behavioral scoring system can provide important references for merchants and agents, distinguishing trading agents from their underlying principals, and assessing credit based on real behaviors that cannot be purchased. Each module collectively forms a trust stack for identity, just as bank transactions are layered with SMS verification, secondary verification, and multiple security keys.

Humans have spent a long time teaching agents to complete payments, information retrieval, and collaboration. But like all business eras, the ultimate direction of the agent economy depends on how we build the trust system.

In every cycle of business development, intermediaries that can solve trust issues always capture the greatest value. This wave of intelligent agents is unlikely to be an exception.

This content is provided for general informational purposes only and doesn’t constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.