Files
nexus/wiki/sources/multi-agent-team.md
2026-04-17 13:17:36 +08:00

52 lines
3.2 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
---
title: "Multi-Agent Specialized Team (Solo Founder Setup)"
type: source
tags: []
date: 2026-04-17
---
## Source File
- [[raw/Agent/usecases/multi-agent-team.md]]
## Summary
- 核心主题Solo founder一人创始人通过多 Agent specialized team 实现高效工作模式
- 问题域:一人创始人需要在战略、开发、营销、销售、运营等多角色间切换,上下文切换破坏深度工作
- 方法/机制:每个 Agent 有独立角色、人格、优化的模型,通过共享内存+私有上下文+单一 Telegram 控制平面+定时任务实现协同
- 结论/价值Agent 需有鲜明人格,共享内存+私有上下文组合是关键便宜模型做简单任务定时任务是飞轮建议从2个Agent开始
## Key Claims
- 单个 Agent 无法同时做好所有事情:单一 Agent 的上下文窗口在同时处理策略、代码、营销研究、商业分析时快速填满
- 通用提示产生通用输出:编码 Agent 不应该同时撰写营销文案
- 需求团队而非另一个工具Agent 应在后台工作并呈现结果,而非需要持续照料
## Key Quotes
> "Give agents distinct names and communication styles makes it natural to 'talk to your team' rather than wrestle with a generic AI" — 鲜明人格让与 Agent 交互更像与团队对话
> "Shared memory + private context: The combination is critical — agents need common ground (goals, decisions) but also their own space to accumulate domain expertise" — 共享内存与私有上下文组合是关键
> "Right model for the right job: Don't use an expensive reasoning model for keyword monitoring" — 匹配模型能力与任务复杂度
> "Start with 2, not 4: Begin with a lead + one specialist, then add agents as you identify bottlenecks" — 从2个开始逐步扩展
## Key Concepts
- [[Multi-Agent Team]]:多 Agent 协作架构,每个 Agent 有独立角色、人格、优化的模型
- [[Shared Memory]]:团队共享的文档、目标、关键决策,所有 Agent 可访问
- [[Private Context]]:每个 Agent 维护自己的对话历史和领域特定的笔记
- [[Scheduled Tasks]]Agent 主动工作而非被动响应,定时执行任务产生飞轮效应
## Key Entities
- [[OpenClaw]]AI Agent 管理工具,支持多 Agent 协调
- [[Telegram]]:单一控制平面,通过 @tag 触发特定 Agent
- [[Claude Opus]]:用于复杂推理任务(如 Strategy Lead
- [[Claude Sonnet]]:用于快速分析任务(如 Business Analyst
- [[Gemini]]:用于网络研究和长上下文分析(如 Marketing Agent
- [[Trebuh]]X 用户实践4个 OpenClaw Agent 的 solo founder
## Connections
- [[OpenClaw]] ← enables ← [[Multi-Agent Team]]
- [[Telegram]] ← routes ← [[Multi-Agent Team]]
- [[Vibe-Kanban]] ← similar_role ← [[Multi-Agent Team]]AI Agent 管理工具)
- [[n8n]] ← integrates_with ← [[Agent Chain]](工作流自动化与 Agent 协同)
## Contradictions
- 与 [[Vibe-Kanban]] 对比:
- 冲突点:两者都是 AI Agent 管理工具,但架构不同
- 当前观点Multi-Agent Team 强调多 Agent 并行、共享记忆、Telegram 控制
- 对方观点Vibe-Kanban 强调单 Agent 事件驱动、自然语言交互