We will be undergoing planned maintenance on January 16th, 2026 at 1:00pm UTC. Please make sure to save your work.

Agent SLA Engine

Trustless, on-chain Service Level Agreements for autonomous AI agents using MNEE escrow, decentralized verification, and economic incentives.

Comment

Inspiration

As AI agents become increasingly autonomous, a critical problem emerges: accountability.

If an AI agent fails to deliver a task, how do we enforce outcomes without relying on trust, legal contracts, or centralized platforms? Traditional SLAs depend on human oversight and legal enforcement, which do not scale to autonomous, machine-driven economies.

We were inspired to build Agent SLA Engine to answer this question using programmable money, smart contracts, and decentralized verification — enabling AI agents to operate as reliable economic actors.

What it does

Agent SLA Engine is a decentralized Service Level Agreement (SLA) system for AI agents, built on Ethereum and powered by MNEE (ERC-20).

It enables:

  • Managers to create SLAs and escrow MNEE payments on-chain
  • Workers (AI agents) to bid on tasks and get paid only upon successful delivery
  • Verifiers to independently validate work quality by staking MNEE
  • Automatic resolution with slashing for dishonest verification

The entire SLA lifecycle — from task creation to payment settlement — is enforced trustlessly by smart contracts.

How we built it

We built the core protocol as a Solidity smart contract using OpenZeppelin libraries for security and correctness.

The system follows a strict state machine: CREATED → BIDDING → VERIFYING → RESOLVED

Key components include:

  • MNEE-based escrow for guaranteed payment
  • Worker bidding mechanism for fair task allocation
  • Two-verifier consensus system to avoid centralized judgment
  • Slashing logic to economically penalize dishonest verifiers

The frontend is built with React, wagmi, and ethers.js, providing a clear interface for each role.

To ensure judges can easily evaluate the project, we implemented a fully functional Demo Mode that simulates the exact same logic using a Hardhat-deployed ERC-20 mock with the same interface as the official MNEE contract.

Challenges we ran into

  • Verifier disagreement handling: Designing a fair and deterministic slashing mechanism required careful incentive design.
  • Demo vs production parity: Ensuring the demo environment behaves identically to mainnet logic without risking real funds.
  • UX clarity: Making complex on-chain flows understandable in under five minutes for judges.

Accomplishments that we're proud of

  • A complete on-chain SLA lifecycle enforced entirely by smart contracts
  • A two-verifier consensus model that prevents single points of failure
  • Economic incentives and slashing that align honest behavior
  • A judge-friendly demo mode requiring no wallet or token setup
  • A clean, minimal UI that clearly communicates system state

What we learned

  • Incentive design and game theory matter as much as smart contract code
  • Demo accessibility is critical for hackathon judging
  • State machines are ideal for modeling decentralized workflows
  • Programmable money enables entirely new coordination mechanisms

What's next for Agent SLA Engine

  • Multi-verifier pools with reputation scoring
  • DAO-based dispute escalation
  • SDKs for autonomous AI agents to create and accept SLAs programmatically
  • IPFS integration for off-chain deliverables

Built With

Share this project:

Updates