ERC-xxx: Agentic On-Chain Operation Interface

Ethereum Magicians

We are going to standardize the canonical workflow interface for agentic on-chain operations. We aim to define a signed intent envelope expressing core safety bounds (maximum input, minimum output, validity windows, and replay protection), a permissionless settlement surface enabling competitive solver fulfillment, and structured settlement receipts providing verifiable workflow closure. The standard separates intent authorship from solving and execution, enabling autonomous agents, solvers, and users to coordinate around outcome-based objectives. Core safety bounds are enforceable on-chain by every compliant hub. Additional policies MAY be enforceable via optional constraint modules. Every outcome is recorded via machine-readable receipts suitable for post-execution verification and structured record-keeping. This ERC provides a common vocabulary and interface for agentic on-chain operations: a standardized lifecycle and receipts that make agents interoperable across runtimes and ecosystems. # Motivation Ethereum transactions encode execution instructions rather than desired outcomes. Emerging agent systems require a mechanism to express goals without prescribing execution paths. Many systems define “orders” or “intents,” but these systems vary in lifecycle semantics, signature domains, and receipt formats. The ecosystem lacks a standard lifecycle and uniform settlement receipts that make agents interoperable across runtimes. Without a shared workflow interface, each agent-solver pair must negotiate bespoke formats, verification procedures, and completion signals — fragmenting the ecosystem and limiting composability. While [ERC-8004](./eip-8004.md) standardizes trustless agent identities and reputation, it does not define how agents express actionable objectives or how such objectives are executed and verified on-chain. This proposal introduces: - A canonical intent envelope with core safety bounds enforceable on-chain by every compliant hub - Optional constraint modules for additional policies, with a standard enforcement hook - A permissionless fulfillment interface supporting on-chain and off-chain intent distribution - Atomic settlement guarantees with structured receipt events - A standardized workflow lifecycle from authorship through verification The goal is to enable autonomous agents, solvers, and builders to coordinate around outcome-based execution while preserving user safety, composability, and auditability. This ERC defines the minimal workflow closure primitive — core safety bounds enforceable on-chain, structured receipts, and optional constraint modules — needed for agents to treat on-chain actions as composable, verifiable skills. 1 post - 1 participant Read full topic