Surety Project
On-chain accountability and financial recourse for autonomous AI agents.
About
Surety is an on-chain accountability and recourse protocol for autonomous AI agents. Agents register a transparent operating policy and post a financial bond that can compensate harmed users when an agent violates its declared rules. The protocol is designed around verifiable claims, dispute resolution, and protection against frivolous claims, creating a neutral accountability layer between autonomous agents and the people and protocols that interact with them.
How it's made
Surety is built as a TypeScript/Solidity monorepo with a Next.js frontend using React, wagmi, viem, RainbowKit and Tailwind, alongside Foundry for smart-contract development and Ponder for blockchain indexing. The local development environment runs against Anvil, with Solidity interfaces for UMA Optimistic Oracle V3 and EAS already included. The architecture is designed around an agent registry, bonded vaults, policy enforcement and an optimistic claims flow, with clean interfaces for future identity, risk-screening and transaction-policy integrations. We prioritized modular infrastructure, testability and verifiable on-chain state so each component can be independently inspected and extended.