AI Skill Hub 推荐使用:开源AI工作流 是一款优质的Agent工作流。AI 综合评分 7.5 分,在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案,这是一个值得深入了解的选择。
From Agent to Agency — 一个 AI 做不了的事,一群 AI 可以。Multi-Agent Collaboration Platform,提供了一个多智能体协作平台,提高了AI的协作效率和智能。
开源AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
From Agent to Agency — 一个 AI 做不了的事,一群 AI 可以。Multi-Agent Collaboration Platform,提供了一个多智能体协作平台,提高了AI的协作效率和智能。
开源AI工作流 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:npm 全局安装 npm install -g mind-agency # 方式二:npx 直接运行(无需安装) npx mind-agency --help # 方式三:项目依赖安装 npm install mind-agency # 方式四:从源码运行 git clone https://github.com/Toufumind/mind-agency cd mind-agency npm install npm start
# 命令行使用
mind-agency --help
# 基本用法
mind-agency [options] <input>
# Node.js 代码中使用
const mind_agency = require('mind-agency');
const result = await mind_agency.run(options);
console.log(result);
# mind-agency 配置说明 # 查看配置选项 mind-agency --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export MIND_AGENCY_CONFIG="/path/to/config.yml"
<img src="public/logo-static.png" width="100" alt="Mind Agency Logo" />
| Feature | Description |
|---|---|
| **👥 Team Collaboration** | Create any number of Agents with roles, personalities, and memory. Collaborate via group chat and email. |
| **🗳️ Consensus Voting** | AND / OR / Threshold voting + adversarial review + multi-round debate. |
| **🔄 Workflow Engine** | YAML-defined pipelines with conditional branching, human approval gates, crash recovery, and hot-reload. |
| **🧠 Three-Layer Memory** | Session + long-term persistent + entity memory. Cross-session experience accumulation. |
| **📡 Signal-Driven** | Filesystem mtime-based incremental scanning with priority debouncing. Agents respond autonomously. |
| **📋 Audit Trail** | Every Agent action is logged and traceable. |
| **🔒 Four-Layer Permissions** | MCP tools → Permission engine → Consensus engine → Adversarial review. |
| **💾 Reliability** | DLQ + Outbox + checkpoint recovery + backpressure. |
| **🎨 Multi-Theme** | Notion, Minimal White, Warm Wood, Deep Space, Nord. |
| **🔌 Multi-Provider** | Claude, DeepSeek, GPT-4o — each Agent can use a different model. |
---
1. Set up your API Key
Open http://localhost:3000, go to Settings, and enter your AI model key.
Supports Claude / DeepSeek / GPT-4o. DeepSeek is the cheapest — pennies per day.
2. Create Agents
The system ships with sample Agents (Alice / Bob / Charlie / you) — ready to go. You can also create your own, or let Agents recruit new members via the agent_create tool:
Name: Diana
Role: Frontend specialist
Personality: Meticulous, quality-focused, loves React
Each Agent has its own config, memory, and behavioral profile.
3. Form a Group
Invite Agents into a group. Each group has its own chat channel, email system, and workflow.
4. Start Collaborating
Assign tasks and watch Agents collaborate. Or define a workflow and let the pipeline run:
steps:
- id: write
agent: Alice
action: code
prompt: Build a user registration endpoint
- id: review
agent: Bob
action: review
dependsOn: [write]
- id: test
agent: Charlie
action: test
dependsOn: [review]
- id: approve
action: human_approval
dependsOn: [test]
---
该项目提供了一个多智能体协作平台,提高了AI的协作效率和智能,但需要进一步优化和完善
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
✅ Apache 2.0 — 宽松开源协议,可商用,需保留版权声明和 NOTICE 文件,含专利授权条款。
总体来看,开源AI工作流 是一款质量良好的Agent工作流,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | mind-agency |
| 原始描述 | 开源AI工作流:From Agent to Agency — 一个 AI 做不了的事,一群 AI 可以。Multi-Agent Collaboration Platform.。⭐7 · TypeScript |
| Topics | workflowagentsaiautomationclaudecollaborationtypescript |
| GitHub | https://github.com/Toufumind/mind-agency |
| License | Apache-2.0 |
| 语言 | TypeScript |
收录时间:2026-06-06 · 更新时间:2026-06-06 · License:Apache-2.0 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端