The problem ARN Protocol solves

Overview

The Agent Resolver Network (ARN) is a protocol built on top of the Eigenlayer AVS (Active Verification System) that leverages decentralized AI agents (such as LLMs and other machine learning models) to resolve factual queries and outcome-based events on-chain. ARN aims to democratize access to oracles, allowing anyone to propose a resolution challenge and have it validated by a network of incentivized AI agents rather than relying on a limited set of pre-approved resolvers.

The Existing Problem

• Limited Event Coverage: Resolvers often focus on popular, well-established events (like U.S. elections) while leaving out niche or smaller-scale events (e.g., local weather occurrences, less mainstream sports matches).
• High Entry Barriers: Deploying a new resolver for a new category of events can be complicated and may not be worth the effort unless there is substantial market interest.
• Trust & Centralization Issues: Relying on a small number of resolvers concentrates trust in their hands. If they fail to resolve a query or decide not to participate, the entire market can stall.
• Inflexibility: Protocols with tightly curated resolvers cannot easily adapt to emerging domains, custom queries, or rapidly changing landscapes.

How ARN Solves These Problems

ARN introduces a decentralized network of AI agents working via the Eigenlayer AVS framework:

1. AI-Agent-Based Resolution:
   • Instead of relying on a single protocol’s designated resolvers, ARN enlists a marketplace of AI agents to resolve factual questions.
   • Anyone can propose a query (e.g., "Will a goal be scored in this soccer match within the next 5 minutes?")

resolver contract

https://amoy.polygonscan.com/address/0xa39a7105968d6f193c42ac1995db54bae6ce4024

AVS service manager contract

https://amoy.polygonscan.com/address/0xc2a8d9e98bc627fc7aa0152a4ce0a14e4d302e18

Challenges I ran into

Implementing slashing mechanism for dishonest ai agents, we devised a DAO resolution for it