En

@0xkobold/pi-erc8004

A protocol for AI agent identity, reputation, and discovery on Base L2.

开发流程 开发AI应用管理多个本地AI编码代理协调多智能体 claude ★ 58 Updated 2026-08-02

Install & Use

Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:

Help me install this AI Skill: @0xkobold/pi-erc8004.
It is used for: A protocol for AI agent identity, reputation, and discovery on Base L2.
Full Skill content: https://321skill.com/skills/0xkobold-pi-erc8004/raw/index.md
Read that page and install it.

The prompt includes a link to the full Skill content. You can also view the full content.

This Skill addresses the lack of trusted identity, reputation, and discovery mechanisms for AI agents in decentralized environments. In practical development, developers often need to assign unique identities to multiple AI coding agents, record their work reputation, and quickly find agents with specific capabilities, but there has been a lack of a unified on-chain protocol to manage this information. The ERC-8004 protocol integrates identity registration, reputation scoring, and agent discovery into Base L2 smart contracts, making agent identities verifiable and traceable.

Usage is straightforward: simply install the extension via the pi CLI, then import the module in pi-config.ts. After that, you can use natural language commands to register an agent identity (specifying type and capabilities), submit reputation scores (0-100) for specific workers, query your own or others' reputation, and discover agents by capability keywords. All operations automatically interact with on-chain smart contracts and support integration with the pi-wallet for payments.

It is ideal for teams or individuals managing multiple AI coding agents, especially those already using the pi-coding-agent ecosystem and looking to establish an agent trust system on Base L2. It is also suitable for projects building decentralized AI markets, needing to share agent capabilities across teams, or wanting to introduce reputation endorsement mechanisms in agent collaboration.

It is recommended to use this tool uniformly when developing multi-agent collaboration systems or needing to share agent capabilities across teams. Note that it primarily relies on Base L2 smart contracts, currently supports only Base Mainnet and Base Sepolia testnet, and requires configuring the pi-wallet extension to complete on-chain transactions. Reputation data is based on on-chain contracts; each submission and query consumes Gas fees, so it's advised to validate first on the testnet.

Key Features

Unlike simple on-chain identity registration (e.g., ENS), ERC-8004 not only provides identity registration but also includes a complete reputation scoring system (0-100 points, divided into Bronze/Silver/Gold/Platinum tiers) and capability-based agent discovery. This allows developers to quickly find trusted AI agents like searching for APIs, while supporting on-chain payments and wallet integration.

Limitations

Depends on the pi-coding-agent environment and Base L2 network; requires installation of the supporting pi-wallet extension. Reputation data is based on on-chain contracts, and each transaction consumes Gas fees. Currently only supports Base chain (Mainnet and Sepolia testnet), not other EVM chains.

FAQ

How do I register an identity for an AI agent?

Use the command `/erc8004 register specialist --capabilities typescript,react` to register. You must first create and link a wallet via pi-wallet.

How do I query the reputation score of a specific agent?

Use the command `/erc8004 reputation --worker 0x742d35...` to view the reputation score and tier for a specified address. Omitting the worker parameter shows your own reputation.

Installation guide for AI assistants

If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:

Visit https://321skill.com/skills/0xkobold-pi-erc8004/raw/index.md to read the original Skill definition (Markdown format) for @0xkobold/pi-erc8004, and install it according to the instructions.