Ethereum has introduced a new standard proposal, ERC-8126, aimed at building a verifiable trust layer for on-chain AI Agents. The standard standardizes five categories of risk verification — smart contracts, media content, Solidity code, websites, and wallets — and outputs a unified risk score along with verifiable proofs. This enables wallets, marketplaces, and decentralized applications (dApps) to make risk-control decisions based on objective signals. ERC-8126 fills a critical gap between identity registration (ERC-8004) and trusted interaction.
As AI Agents become increasingly active in on-chain ecosystems, trust issues have emerged as a bottleneck to their growth. ERC-8126 provides a "credit-check" mechanism through a modular risk-verification framework. It not only verifies the identity of AI Agents but also assesses the security of their associated contracts and data authenticity, thereby reducing friction costs in on-chain interactions.
Specifically, contract risk verification examines smart contract vulnerabilities and compliance; media verification checks content authenticity and source credibility; Solidity code verification ensures code quality and safety; website verification assesses domain and front-end risks; and wallet verification reviews address history and asset legitimacy. By integrating these indicators, ERC-8126 generates a comprehensive risk score and provides on-chain verifiable certificates for real-time use by other protocols or users.