Claude Skill
win4r/openclaw-a2a-gateway
OpenClaw A2A Gateway 是一个 TypeScript 插件,实现 A2A(Agent-to-Agent)协议 v0.3.0,为 OpenClaw 生态系统提供双向智能体通信网关。
概览
仓库信息
安装这个 Skill
git clone https://github.com/win4r/openclaw-a2a-gateway.git a2a-gatewayRegistry 信息
git clone https://github.com/win4r/openclaw-a2a-gateway.git a2a-gatewaynpm install --production
项目简介
OpenClaw A2A Gateway 是一个实现 A2A(Agent-to-Agent)协议 v0.3.0 的插件,为 OpenClaw 生态系统中的智能体提供双向通信网关。
OpenClaw plugin implementing the A2A (Agent-to-Agent) protocol v0.3.0 — bidirectional agent communication gateway
要点
- 实现 A2A 协议 v0.3.0,支持智能体互操作
- 提供双向智能体通信网关
- 作为 OpenClaw 插件/扩展构建
- 使用 TypeScript 编写,保证类型安全
- 支持 OpenClaw 智能体生态系统集成
使用场景
- 连接多个 OpenClaw 智能体以完成协作任务
- 构建具有标准化通信的多智能体系统
- 实现智能体之间的数据交换与协调
- 通过可互操作的智能体网络扩展 OpenClaw
README 摘要
# 🦞 OpenClaw A2A Gateway Plugin [](LICENSE) [](https://github.com/google/A2A) []() []() [English](README.md) | [简体中文](README_CN.md) | [繁體中文](README_TW.md) | [日本語](README_JA.md) | [한국어](README_KO.md) | [Français](README_FR.md) | [Español](README_ES.md) | [Deutsch](README_DE.md) | [Italiano](README_IT.md) | [Русский](README_RU.md) | [Português (Brasil)](README_PT-BR.md) A production-ready [OpenClaw](https://github.com/openclaw/openclaw) plugin that implements the [A2A (Agent-to-Agent) v0.3.0 protocol](https://github.com/google/A2A), enabling OpenClaw agents to discover and communicate with each other across servers — with zero-config install and automatic peer discovery. **The only A2A gateway with adaptive, bio-inspired routing, discovery, and resilience — designed for multi-agent ecosystems at scale.** ## Key Features ### Transport & Protocol - **Three transports**: JSON-RPC, REST, and gRPC — with automatic fallback (tries JSON-RPC → REST → gRPC) - **SSE streaming** with heartbeat keep-alive for real-time task status - **Full Part type support**: TextPart, FilePart (URI + base64), DataPart (structured JSON) - **Auto URL extraction**: file URLs in agent text responses are promoted to outbound FileParts ### Intelligent Routing - **Rule-based routing**: auto-select peer by message pattern, tags, or peer skills - **Hill equation affinity scoring**: multi-dimensional routing with sigmoid scoring — skills, tags, pattern, and success rate weighted via `score = affinity^n / (Kd^n + affinity^n)` ([Hill, 1910](