# echo-agent 知识库
本目录包含 echo-agent 项目涉及的核心概念和技术知识文档。
---
## 文档索引
| [Agent 模式](./agent-patterns.md) | ReAct、Plan-and-Execute、Self-Reflection、LangGraph 工作流等 AI Agent 核心模式 |
| [MCP 协议](./mcp-protocol.md) | Model Context Protocol 规范与 echo-agent 集成实现 |
| [Skill 系统](./skill-system.md) | agentskills.io 规范、渐进式披露、代码型/文件型技能 |
| [A2A 协议](./a2a-protocol.md) | Agent-to-Agent 协议、Agent Card、任务状态机 |
---
## 核心概念速查
### Agent 执行模式
| **ReAct** | Thought → Action → Observation | 工具编排、开放式问答 |
| **Plan-and-Execute** | Plan → Execute (DAG) → Summary | 结构化多步任务 |
| **Self-Reflection** | Generate → Critique → Refine | 高质量输出保证 |
| **Graph Workflow** | State → Node → State | 多 Agent 协作编排 |
### 协议对比
| **MCP** | 工具/资源访问 | Tool 层 |
| **A2A** | Agent 间通信 | Agent 层 |
| **OpenAI Functions** | LLM 工具调用格式 | API 层 |
### 抽象层级
```
┌─────────────────────────────────────────┐
│ Application │
├─────────────────────────────────────────┤
│ A2A Protocol (Agent-to-Agent) │ ← Agent 互操作
├─────────────────────────────────────────┤
│ Graph Workflow / Multi-Agent │ ← 编排层
├─────────────────────────────────────────┤
│ Agent (ReAct / Plan-Exec / Reflection) │ ← Agent 实现
├─────────────────────────────────────────┤
│ Skills (Code-based / File-based) │ ← 能力包
├─────────────────────────────────────────┤
│ Tools + MCP │ ← 工具层
├─────────────────────────────────────────┤
│ LLM Provider (OpenAI / Anthropic / ...) │ ← 基础层
└─────────────────────────────────────────┘
```
---
## 扩展阅读
### 学术论文
1. Yao, S., et al. "ReAct: Synergizing Reasoning and Acting in Language Models." ICLR 2023.
2. Wei, J., et al. "Chain-of-Thought Prompting Elicits Reasoning in Large Language Models." NeurIPS 2022.
3. Shinn, N., et al. "Reflexion: Language Agents with Verbal Reinforcement Learning." NeurIPS 2023.
4. Gou, Z., et al. "CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing." ICLR 2024.
### 技术规范
- [LangGraph Documentation](https://langchain-ai.github.io/langgraph/)
- [MCP Specification](https://modelcontextprotocol.io/)
- [agentskills.io Specification](https://agentskills.io/specification)
- [A2A Protocol](https://github.com/google/A2A)
### 相关框架
- [LangChain](https://www.langchain.com/) - Python/JS Agent 框架
- [CrewAI](https://www.crewai.com/) - 多 Agent 协作框架
- [AutoGen](https://microsoft.github.io/autogen/) - 微软多 Agent 框架
- [LangGraph](https://langchain-ai.github.io/langgraph/) - 图工作流框架