Claude Skill

aws-samples/sample-host-openclaw-on-amazon-bedrock-agentcore

探索在 Amazon Bedrock Agent Core 上托管 OpenClaw 技能的 Python 示例实现。了解如何使用 AWS 部署 AI 代理，构建可扩展的智能应用程序。

概览

Stars164
Forks48
语言Python
最后更新2026-05-23
最近同步2026-06-24
前往 GitHub

仓库信息

拥有者aws-samples
仓库sample-host-openclaw-on-amazon-bedrock-agentcore
完整名称aws-samples/sample-host-openclaw-on-amazon-bedrock-agentcore
Repo ID1,162,268,824
GitHub 地址https://github.com/aws-samples/sample-host-openclaw-on-amazon-bedrock-agentcore

安装这个 Skill

git clone https://github.com/aws-samples/sample-host-openclaw-on-amazon-bedrock-agentcore.git
GitHub

Registry 信息

类型openclaw_skill
质量分85/100
验证状态readme_parsed
最近验证2026-06-24
平台
ClaudeOpenClaw
能力
browserpdfmemorysearchimagevideoterminalworkflow
识别文件
README.mddocsrequirements.txttests
配置键
YOUR_TELEGRAM_BOT_TOKENURLAPI_URLWEBHOOK_SECRETTELEGRAM_TOKENYOUR_TELEGRAM_USER_IDYOUR_BOT_TOKENTOKENSECRETYOUR_MEMBER_IDUSER_IDRUNTIME_ID
安装方式
  • git clone https://github.com/aws-samples/sample-host-openclaw-on-amazon-bedrock-agentcore.git
  • pip install -r requirements.txt
  • pip install bedrock-agentcore-toolkit
  • npx promptfoo@latest view

项目简介

该仓库提供了一个示例实现，用于在 Amazon Bedrock Agent Core 上托管 OpenClaw 技能，帮助开发者利用 AWS 服务将 AI 驱动的代理能力集成到其应用程序中。

要点

  • 在 Amazon Bedrock Agent Core 上托管 OpenClaw 的示例代码
  • 基于 Python 的实现，便于集成
  • 利用 AWS Bedrock 实现可扩展的 AI 代理部署
  • 演示代理核心的配置与管理
  • 构建自定义 AI 代理的开源参考

使用场景

  • 部署用于客户支持自动化的 AI 代理
  • 为企业工作流构建智能助手
  • 将 AI 驱动的决策集成到应用程序中
  • 在 AWS 上原型设计和测试基于代理的解决方案
  • 学习如何将 Amazon Bedrock Agent Core 与自定义技能结合使用

README 摘要

# OpenClaw on AWS Bedrock AgentCore [![License: MIT-0](https://img.shields.io/badge/License-MIT--0-blue.svg)](LICENSE) [![Status: Experimental](https://img.shields.io/badge/Status-Experimental-orange.svg)]() [![AWS CDK](https://img.shields.io/badge/AWS%20CDK-v2-yellow.svg)]() > **Experimental** — This project is provided for experimentation and learning purposes only. It is **not intended for production use**. APIs, architecture, and configuration may change without notice. Deploy an AI-powered multi-channel messaging bot (Telegram, Slack) on AWS Bedrock AgentCore Runtime using CDK. ## Table of Contents - [Architecture](#architecture) - [Prerequisites](#prerequisites) - [Quick Start](#quick-start) - [Project Structure](#project-structure) - [Configuration](#configuration) - [Channel Setup](#channel-setup) - [How It Works](#how-it-works) - [Operations](#operations) - [Troubleshooting](#troubleshooting) - [Known Limitations](#known-limitations) - [Gotchas](#gotchas) - [Cleanup](#cleanup) - [Security](#security) - [Security Testing](#security-testing) - [License](#license) OpenClaw runs as **per-user serverless containers** on AgentCore Runtime. A Router Lambda handles webhook ingestion from Telegram and Slack, resolves user identity via DynamoDB, and invokes per-user AgentCore sessions. Each user gets their own microVM with workspace persistence (`.openclaw/` directory synced to S3). The agent has built-in tools (web, filesystem, runtime, sessions, automation), custom skills for file storage and cron scheduling, and **EventBridge-based cron scheduling** for recurring tasks. Users can send **text and images** — photos sent via Telegram or Slack are downloaded by the Router Lambda, stored in S3, and passed to Claude as multimodal content via Bedrock's ConverseStream

话题

暂无话题

探索更多

拥有者: aws-samples语言: Python

数据来自 GitHub，同步时间：2026-06-24