Blockchain security firm CertiK has outlined a transformative vision for autonomous AI systems in a new analysis released recently this month. Today’s AI agents operate in isolated environments, relying on large language models for decision-making but lacking uniform ways to prove who they are or handle secure transactions.
Without reliable identity verification or payment mechanisms, these non-human entities struggle to build trust or participate in open commerce. CertiK argues that blockchain can solve this by acting as a neutral, decentralized foundation, enabling agents to function as fully independent economic participants.
Three emerging technical standards—EIP-8004, EIP-8183, and x402—form the backbone of this shift, addressing identity, commerce rules, and instant payments respectively.
At the core of agent recognition is EIP-8004, which functions like a digital résumé for AI systems. It leverages non-fungible token technology to assign each agent a unique, cross-chain identifier compatible with existing Web3 tools. The standard rests on three interconnected registries.
The identity registry issues a global agent ID in the format {namespace}:{chainId}:{identityRegistry} and links it to an off-chain JSON file—known as an Agent Card—that details the agent’s capabilities, endpoints, status, and associated wallet address.
This wallet metadata is protected by cryptographic signatures, preventing unauthorized changes to payment flows. A separate reputation registry maintains an immutable on-chain log of client feedback, allowing anyone to submit or withdraw reviews while filtering out suspicious activity from new addresses.
Finally, a validation registry lets agents submit their outputs for independent checks using tools such as zero-knowledge proofs, trusted execution environments, or stake-based re-execution.
As of early May 2026, the identity and reputation contracts are live on approximately 30 networks, with validation features still in early rollout.
Once agents can be identified, EIP-8183—the Agentic Commerce Protocol—establishes clear rules for hiring, delivering, and settling work. Developed in collaboration with Virtuals Protocol and the Ethereum Foundation’s dAI team, it operates as a state-machine escrow system built around a “Job” primitive.
Three distinct roles interact: the client who requests work, the provider who performs it, and the evaluator who judges completion. Key steps include creating the job, selecting a provider, agreeing on a budget, funding the escrow, submitting results, and final approval or rejection. The evaluator holds decisive power, serving as the transaction’s logical gatekeeper.
Evaluators can be other AI agents for subjective tasks, zero-knowledge circuits for objective verification, or multi-signature setups for high-stakes decisions.
To handle edge cases like evaluator downtime, the protocol includes an expiration timestamp that automatically refunds the client. Flexible “hooks” allow developers to add custom logic—such as compliance checks or reputation updates—without altering the core contract, though refunds remain protected from interference.
Seamless payments complete the picture through x402, an open protocol introduced by Coinbase that revives the long-forgotten HTTP 402 “Payment Required” status code.
Designed for machine-speed interactions, x402 enables agents to handle micropayments automatically using stablecoins.
It relies on three standardized HTTP headers exchanged between servers and clients: PAYMENT-REQUIRED (sent by the server to request payment), PAYMENT-SIGNATURE (client’s cryptographic response), and PAYMENT-RESPONSE (server’s confirmation of settlement). This setup supports per-request, per-token, or usage-based billing with minimal friction.
By combining these standards, CertiK envisions a future where agent-to-agent transactions form a growing portion of on-chain activity.
Humans would transition from direct participants to designers and overseers of the governing code. Another update from the blockchain focused firm will explore security considerations and protective measures for projects building in this space. CertiK has concluded that as these protocols mature, they aim to effectively unlock an open, trust-minimized machine economy.