bagualu 0.1.3

八卦炉 - Intelligent Agent Orchestration Platform for Self-Evolving Agent Clusters

BaGuaLu (八卦炉)

八卦炉 - Intelligent Agent Orchestration Platform for Self-Evolving Agent Clusters

一键部署智能体集群 | 技能自我进化 | 多工作流编排

English | 中文

---

🎯 What is BaGuaLu?

BaGuaLu (八卦炉) 是一个强大的智能体编排平台，融合了 OpenLaoKe、OpenSpace 和 OpenHarness 的优秀设计理念：

✨ 核心特性

特性	描述
🚀 一键部署	单个智能体、智能体矩阵、集群的一键部署
🧬 自我进化	智能体和技能的自我进化能力（FIX, DERIVED, CAPTURED）
🔧 多提供商支持	Ollama、LMStudio、OpenAI、Claude、CodingPlan 等
📚 技能系统	OpenCode/Claude Code 技能加载和按需知识
🌐 工作流编排	Web界面拖拽设计工作流，DAG执行
🤝 多智能体协调	Supervisor、Scheduler、Executor 协同工作
🛡️ 权限管理	多级权限控制和钩子系统
📊 资源调度	智能资源分配和负载均衡

---

🏗️ Architecture

bagualu/
├── core/              # 核心编排系统
│   ├── bagualu_core.py    # 主核心类
│   ├── orchestrator.py    # 编排器
│   └── resource_manager.py # 资源管理
├── agents/            # 智能体系统
│   ├── base.py            # 基础智能体
│   ├── executor.py        # 执行智能体
│   ├── supervisor.py      # 监工智能体
│   ├── scheduler.py       # 调度智能体
│   └── cluster.py         # 智能体集群
├── skills/            # 技能系统（参考 OpenSpace）
│   ├── skill_engine.py    # 技能引擎
│   ├── evolver.py         # 自我进化引擎
│   ├── registry.py        # 技能注册
│   └── store.py           # 持久化存储
├── config/            # 配置系统（参考 OpenLaoKe）
│   ├── config_manager.py  # 配置管理
│   ├── providers.py       # 多提供商配置
│   └── wizard.py          # 配置向导
├── workflow/          # 工作流系统
│   ├── workflow_engine.py # 工作流引擎
│   └── workflow_dag.py    # DAG定义
├── tools/             # 工具系统（参考 OpenHarness）
│   └── base.py            # 工具基类
├── web/               # Web界面
│   ├── api_server.py      # REST API
│   └── workflow_ui.py     # 工作流UI
└── entrypoints/       # 入口点
    └── cli.py             # CLI入口

---

🚀 Quick Start

Installation

# Clone repository
git clone https://github.com/cycleuser/BaGuaLu.git
cd BaGuaLu

# Install with pip
pip install -e .

# Or install from PyPI
pip install bagualu

Initialize Configuration

# Run configuration wizard
bagualu --init

Deploy an Agent

# Deploy a single agent
bagualu deploy my-agent --role executor --provider ollama --model llama2

# Or interactively
bagualu
> deploy

Execute a Workflow

# Create workflow.json
{
  "name": "my-workflow",
  "nodes": [
    {"id": "task1", "role": "executor", "instruction": "Analyze codebase"},
    {"id": "task2", "role": "supervisor", "instruction": "Review results", "dependencies": ["task1"]}
  ],
  "edges": [
    {"from": "task1", "to": "task2"}
  ]
}

# Execute workflow
bagualu run workflow.json

---

🔧 Configuration

Multi-Provider Setup

BaGuaLu 支持多种 LLM 提供商：

# ~/.bagualu/config.yaml
providers:
  ollama:
    base_url: http://localhost:11434
    model: llama2
    enabled: true
  
  openai:
    api_key: sk-xxx
    model: gpt-4
    enabled: true
  
  anthropic:
    api_key: sk-xxx
    model: claude-3-5-sonnet-20241022
    enabled: true
  
  coding_plan:
    api_key: sk-xxx
    model: default
    enabled: true

active_provider: ollama

settings:
  max_concurrent_agents: 10
  evolution_enabled: true
  quality_threshold: 0.8

---

🧬 Self-Evolution

BaGuaLu 的技能自我进化系统（参考 OpenSpace）：

Evolution Types

Type	Description
FIX	修复破损或过时的技能指令
DERIVED	从现有技能创建增强版本
CAPTURED	从成功执行中捕获新模式

Trigger Evolution

# CLI
bagualu evolve skill-name

# Python API
await core.evolve_skill("skill-name", evolution_type="auto")

---

🤝 Multi-Agent Coordination

Agent Roles

Role	Responsibility
Executor	执行具体任务
Supervisor	监督和质量控制
Scheduler	任务调度和资源分配
Coordinator	多智能体协调

Deploy a Cluster

import asyncio
from bagualu import BaGuaLuCore

async def main():
    core = BaGuaLuCore()
    await core.initialize()
    
    # Deploy cluster
    cluster_config = {
        "name": "my-cluster",
        "agents": [
            {"name": "executor1", "role": "executor"},
            {"name": "executor2", "role": "executor"},
            {"name": "supervisor1", "role": "supervisor"},
            {"name": "scheduler1", "role": "scheduler"},
        ],
        "connections": [
            {"from": "executor1", "to": "supervisor1"},
            {"from": "executor2", "to": "supervisor1"},
        ]
    }
    
    agent_ids = await core.deploy_cluster(cluster_config)
    print(f"Deployed {len(agent_ids)} agents")

asyncio.run(main())

---

🌐 Web Interface

Start Web Server

bagualu server --host 0.0.0.0 --port 8000

REST API Endpoints

• POST /agents - Deploy agent
• GET /agents - List agents
• POST /workflows - Create workflow
• POST /workflows/{id}/execute - Execute workflow
• GET /skills - List skills
• POST /skills/{name}/evolve - Evolve skill

---

📚 Skills System

Load Skills

# Load from OpenCode/Claude Code skill directory
await core.skills.import_skill_from_opencode(
    Path("~/.config/opencode/skills/my-skill")
)

Create Custom Skill

# ~/.bagualu/skills/my-skill.md

# My Custom Skill

## When to use
Use when the user asks about [specific domain].

## Workflow
1. Step one
2. Step two
3. Step three

## Triggers
- trigger keyword 1
- trigger keyword 2

---

🔌 Integration with Other Projects

BaGuaLu 兼容以下项目的技能和插件：

• ✅ OpenCode skills
• ✅ Claude Code skills
• ✅ OpenLaoKe configurations
• ✅ OpenSpace evolution patterns
• ✅ OpenHarness tool system

---

📊 Examples

See examples/ directory for:

• Agent deployment examples
• Workflow orchestration examples
• Skill evolution examples
• Multi-agent coordination examples

---

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

---

📄 License

MIT License - see LICENSE file.

---

🙏 Acknowledgments

BaGuaLu is inspired by and incorporates design patterns from:

• OpenLaoKe - Multi-provider configuration system
• OpenSpace - Self-evolving skill system
• OpenHarness - Tool and plugin architecture

Special thanks to all contributors and the open-source community!

---

八卦炉 - 锻造智能，炼化知识

Made with ❤️ by the BaGuaLu Community