WORLD3 Docs
  • 📖Whitepaper
    • Introduction & Vision
    • Problem Statement
    • Introducing WORLD3
    • Technical Architecture
      • AI Agents
      • WORLD AI Protocol Contracts
      • WORLD Agent Protocol: A Decentralized AI Agent Network
      • Human Players in WORLD AI Protocol
    • Use-Cases and Applications
      • No-code AI Agent Builder
      • Use-Cases
        • Social Media Influencers (KOLs) Agents
        • On-chain Gaming Agents
        • Web3 Project Founder Agent
        • DeFi Trading Agents
    • WORLD3 Ecosystem Apps
    • Tokenomics
      • $WAI Tokenomics
      • Business Model
    • Governance and Ethical Framework
    • Community and Ecosystem Development
      • Benefits to Landowners
    • Roadmap
      • What We Have Accomplished
      • Upcoming Initiatives
    • Conclusion
  • 🖥️Pitch Deck
  • 🔎How-To Guides
    • AI Agent Portal
      • Overview
      • Quick Start Guide
        • Quick Start Guide (中文)
        • Quick Start Guide (한국어)
      • Core Features
        • Knowledge Packs
        • Skill Plugins
        • Task Management
      • Tutorials
        • How to Build Your Twitter Influencer Agent
      • Campaign Guide
        • AI Agent Influencer Campaign Guide (EN)
        • AI Agent Influencer Campaign Guide (中文)
        • AI Agent Influencer Campaign Guide (한국어)
    • Crystal Caves
      • Crystal Caves Guide
      • Crystal Caves Soneium Summit: Player's Guide
        • Crystal Caves Soneium Summit: Player's Guide (中文)
        • Crystal Caves Soneium Summit: Player's Guide (한국어)
      • B² Gold Miner Campaign: Player's Guide
        • B² Gold Miner Campaign: Player's Guide(한국어)
        • B² Gold Miner Campaign: Player's Guide (中文)
      • Crystal Caves SKALE Summit: Player's Guide
        • Crystal Caves SKALE Summit: Player's Guide (中文)
      • Crystal Caves SUI Summit: Player's Guide
        • Crystal Caves SUI Summit: Player's Guide (中文)
        • Crystal Caves SUI Summit: Player's Guide(한국어)
      • Crystal Caves opBNB Summit: Player's Guide
        • Crystal Caves opBNB Summit: Player's Guide (中文)
        • Crystal Caves opBNB Summit: Player's Guide(한국어)
      • BaseAI Summit Campaign: Player's Guide
        • BaseAI Summit Campaign: Player's Guide(한국어)
        • BaseAI Summit Campaign: Player's Guide (中文)
        • BaseAI Referral System Guide
    • SOUL Dragons
      • SOUL Dragons AI Chat Campaign
      • SOUL Dragons AI Chat Campaign (中文)
      • SOUL Dragon Trading Guide
      • SOUL Dragon Trading Guide (中文)
      • SOUL Dragons Mint Guide
      • SOUL Dragons Mint Guide (中文)
      • SOUL Dragon Whitelist Ticket Claim Guide
      • SOUL Dragons: Referral Contest Guide
      • SOUL Dragons: Referral Contest Guide (中文)
    • Whitelist Allocation for Landowners
    • Staking System Guide
      • Staking System Guide (中文)
      • Staking System Guide (한국어)
    • Referral Program Guide
      • Referral Program Guide (中文)
      • Referral Program Guide (한국어)
    • Treasure User Guides
      • Treasure User Guides (中文)
      • Treasure User Guides (한국어)
  • Announcements & Notices
    • Account Freezing Rules
      • Account Freezing Rules (中文)
Powered by GitBook
On this page
  1. Whitepaper
  2. Technical Architecture

WORLD Agent Protocol: A Decentralized AI Agent Network

PreviousWORLD AI Protocol ContractsNextHuman Players in WORLD AI Protocol

Last updated 2 months ago

The WORLD Agent Protocol reimagines AI agents as blockchain-managed digital assets and services. It combines decentralized storage, smart contract–based access control, and verifiable off-chain execution to create a secure, transparent, and economically incentivized AI ecosystem.

  1. Agent-as-an-Asset (AAA)

Each AI agent is treated as a digital asset with several core components:

  • Core Data & Configuration

    • Prompt & Knowledge Base: Defines the agent’s behavior and informational foundation.

    • Memory: Stores historical interactions for continuous learning.

    • Authentication & Access Control: Rules embedded in the agent ensure that only authorized parties can access or modify it.

    • Skills Configuration: Modular skills are integrated into the agent’s core settings. Their definitions and references are stored as pointers in smart contracts, while the associated binaries (BLOBs) reside in decentralized storage.

  • Versioning and Ownership

    • NFT-Based Version Control: Every version of an agent is minted as a unique NFT. This ensures a verifiable history, immutable provenance, and clear ownership rights.

    • Access Rights as NFTs: Access permissions are managed via NFTs that grant execution, modification, or monetization rights.

  • Token Economics

    • Fungible Tokens (FTs): Payments, incentives, and fees for agent usage are executed through smart contracts, ensuring transparent and automated settlements.


  1. Decentralized Storage & Smart Contract Integration

  • Decentralized Storage

    • All critical components (prompts, knowledge bases, memory logs, and skills data) are stored in a decentralized manner. This guarantees data integrity and availability while eliminating single points of failure.

  • Smart Contract Layer

    • Smart contracts manage access control, versioning, and the token economy. They reference both the BLOBs in decentralized storage and the dynamic configuration (including skills) of each agent, ensuring that all modifications are transparent and tamper-proof.


  1. Agent-as-a-Service via the Agent Oracle Network (AON)

AI agents are executed securely through a decentralized network of sandbox nodes. This component is designed to provide high-level remote agent calls, ensuring that AI agents can be invoked from any client environment without compromising on-chain security.

Agent Oracle Nodes

  • Secure Execution Environment: These nodes serve as secure, isolated execution environments—comparable to Chainlink oracles—where agents are invoked and run.

Request & Execution Flow

  1. Client Request Signing: Clients sign requests specifying which agent to call, including input data and authentication credentials.

  2. On-Chain Verification: Smart contracts verify the request’s authenticity, NFT-based access rights, and payment status.

  3. Resource Fetching: Oracle nodes retrieve the agent’s metadata, configuration, skills references, and BLOB data from decentralized storage.

  4. Sandboxed Execution: The agent executes in a secure sandbox, ensuring reliable operation without compromising on-chain assets.

  5. High-Level Remote Agent Call (via AON):

    1. Remote Execution Layer: This layer is responsible for abstracting and managing all remote interactions, ensuring that AI agents can be securely and efficiently invoked from any client environment. It provides the scalability and security necessary for decentralized execution, forming the backbone of our remote call mechanism.

    2. Context Extension Module: To enhance an agent’s functionality, our protocol supports a modular context extension capability. This allows the integration of additional tools and data sources to enrich the agent's operational context. For example, the Model Context Protocol (MCP) can be integrated as a standard module to facilitate remote tool calls and extend an agent’s tool skills. While MCP is one example, our design is flexible enough to incorporate alternative modules without altering the overall remote execution framework.

  6. Settlement & Incentivization: Upon successful execution, results are relayed back to the client and smart contracts trigger appropriate token settlements and rewards.

  7. Protocol Workflow

  8. Client Invocation (Off-Chain to On-Chain)

    1. A client signs and submits a request, which is recorded on-chain.

  9. Oracle Network Processing

    1. A decentralized oracle node verifies the request, fetches the necessary agent components (including skills configuration), and executes the agent securely.

  10. Remote Agent Call via AON

    1. The agent is invoked remotely using AON. If required, it performs remote tool calls to leverage additional data or services.

  11. On-Chain Settlement

    1. Execution outcomes are reported back, and smart contracts handle token transfers for payments, royalties, and incentives.


  1. Key Benefits

  • Decentralization & Security: By using blockchain and decentralized storage, the protocol eliminates central points of failure and ensures tamper-proof records.

  • Transparency & Trust: All interactions—agent updates, access permissions, and token economics—are managed via smart contracts, providing verifiable and transparent operations.

  • Modularity & Flexibility: Skills configuration and remote tool calls allow agents to extend their functionality dynamically while maintaining robust, decentralized control.

  • High-Level Remote Agent Call: The Agent Oracle Network (AON) ensures secure, scalable remote agent invocation. It encapsulates lower-level components like MCP, offering a unified remote execution layer.

  • Economic Incentives: NFT and FT mechanisms incentivize stakeholders, ensuring sustainable participation and continuous improvement within the ecosystem.

📖