openclaw-starter-guide
OpenClaw 小白养成手册。从零开始搭建多 Agent AI 助手系统的完整指南,包含免费起步方案(Qwen 零成本)、进阶方案(MiniMax Coding Plan ¥49/月)、旗舰外援(SiliconFlow/NewCLI 按需调用)。涵盖模型策略、Fallback 链设计、额度管理、常见故障排查。适合首次部署 OpenClaw 或想优化模型成本的用户。当用户说"怎么开始"、"新手指南"、"省钱方案"、"模型怎么选"、"fallback 怎么配"时使用此 skill。
Install
mkdir -p .claude/skills/openclaw-starter-guide && curl -L -o skill.zip "https://mcp.directory/api/skills/download/4925" && unzip -o skill.zip -d .claude/skills/openclaw-starter-guide && rm skill.zipInstalls to .claude/skills/openclaw-starter-guide
About this skill
🦞 OpenClaw 小白养成手册
从零到多 Agent AI 系统的完整指南。
🎯 核心理念
OpenClaw 的模型策略就像组建一支球队:
🆓 免费球员(Qwen) → 保底,确保系统永远在线
💰 包月主力(MiniMax) → 日常主力,性价比最高
🚀 外援专家(旗舰模型) → 关键时刻上场,按需付费
目标:用最少的钱,让 AI 助手 24/7 在线、随时可用。
📖 目录
前置条件
- OpenClaw 已安装并运行:参考 官方文档 完成安装
- 至少一个聊天渠道:Telegram、WhatsApp、Discord 等
- Node.js 环境:用于安装 skills(
clawhubCLI)
# 安装 ClawHub CLI(用于安装 skills)
npm i -g clawhub
第一阶段:零成本起步
目标:¥0 成本让系统跑起来。
推荐免费模型
| 来源 | 模型 | 特点 | 安装 Skill |
|---|---|---|---|
| Qwen Portal | qwen-portal/coder-model | OAuth 免费,128K context | 内置 |
| SiliconFlow | Qwen/Qwen3-8B | API Key 免费,无限调用 | clawhub install add-siliconflow-provider |
| SiliconFlow | DeepSeek-R1-0528-Qwen3-8B | 免费推理模型 | 同上 |
最简配置
只需一个 provider 就能开始:
{
"models": {
"providers": {
"siliconflow": {
"baseUrl": "https://api.siliconflow.cn/v1",
"apiKey": "<YOUR_KEY>",
"api": "openai-completions",
"models": [
{
"id": "Qwen/Qwen3-8B",
"name": "Qwen3 8B (Free)",
"reasoning": false,
"input": ["text"],
"cost": {"input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0},
"contextWindow": 32768,
"maxTokens": 8192
}
]
}
}
},
"agents": {
"defaults": {
"model": {
"primary": "siliconflow/Qwen/Qwen3-8B"
}
}
}
}
注册 SiliconFlow:https://cloud.siliconflow.cn/i/ihj5inat (新用户送 ¥14)
免费阶段能做什么
- ✅ 单 Agent 对话
- ✅ 基本工具调用(文件、搜索、命令)
- ✅ 简单自动化任务
- ⚠️ 复杂推理能力有限(8B 模型)
- ⚠️ 长对话可能质量下降(32K context)
第二阶段:包月主力上场
目标:¥49/月获得企业级体验。
当免费模型不够用时,上 MiniMax Coding Plan:
| 项目 | 值 |
|---|---|
| 价格 | ¥49/月 |
| 额度 | 1500 次/5小时窗口(约 7200 次/天) |
| 模型 | MiniMax M2.1(200K context) |
| 质量 | 对标 Claude Sonnet / GPT-4o |
为什么选 MiniMax?
- 💰 性价比无敌:¥49/月 ≈ ¥0.007/次,比按量付费便宜 100 倍
- 📦 额度充足:7200 次/天,6 个 Agent 平均每个 1200 次
- 🧠 200K Context:长对话不丢失上下文
- 🔌 OpenAI 兼容:
openai-completions协议即插即用
安装
clawhub install add-minimax-provider
详细配置见:add-minimax-provider
配置策略
{
"agents": {
"defaults": {
"model": {
"primary": "minimax/MiniMax-M2.1",
"fallbacks": [
"siliconflow/Qwen/Qwen3-8B"
]
}
}
}
}
关键:免费模型放在 fallback 里!MiniMax 挂了或额度用完时,自动切到免费模型,系统不会停。
第三阶段:旗舰外援加持
目标:关键任务用最强模型,按需付费。
| 场景 | 推荐模型 | 来源 | 价格 |
|---|---|---|---|
| 复杂推理 | DeepSeek R1 | SiliconFlow | ¥4/¥16 per M |
| 代码重构 | Qwen3 Coder 480B | SiliconFlow | ¥8/¥16 per M |
| 顶级对话 | Kimi K2.5 | SiliconFlow | ¥4/¥21 per M |
| 长文档 | Claude Opus | NewCLI | 按额度 |
| 多模态 | Gemini 3 Pro | NewCLI | 按额度 |
安装
clawhub install add-siliconflow-provider # DeepSeek/Qwen/Kimi
clawhub install add-newcli-provider # Claude/GPT/Gemini
使用方式
不需要改默认配置。需要时用 /model 命令临时切换:
/model sf-kimi # 切到 Kimi K2.5
/model sf-coder-480b # 切到 Qwen3 Coder 480B
/model claude-opus # 切到 Claude Opus
/model Minimax # 切回 MiniMax(默认)
Fallback 链设计
Fallback 链是 OpenClaw 的生命线——主模型挂了,自动尝试下一个。
推荐 Fallback 策略
第1优先:minimax/MiniMax-M2.1 (API Key 包月主力)
↓ 如果额度用完或 API 故障
第2优先:minimax-portal/MiniMax-M2.1 (OAuth 免费额度)
↓ 如果 OAuth 也不可用
第3优先:siliconflow/Qwen/Qwen3-8B (免费兜底)
↓ 如果 SiliconFlow 也挂了
第4优先:qwen-portal/coder-model (OAuth 免费)
↓ 如果都挂了
第5优先:deepseek/deepseek-chat (便宜)
↓ 最后的最后
第6优先:newcli/claude-haiku (贵但稳)
配置示例
"fallbacks": [
"minimax-portal/MiniMax-M2.1",
"siliconflow/Qwen/Qwen3-8B",
"qwen-portal/coder-model",
"deepseek/deepseek-chat",
"newcli/claude-haiku-4-5-20251001"
]
设计原则
- 免费模型优先放前面:先用免费的,省钱
- 至少 2 个不同来源:避免单一供应商全挂
- 最贵的放最后:Claude/GPT 只在其他全挂时才用
- 不放推理模型:R1/Reasoner 慢且贵,不适合 fallback
额度管理策略
MiniMax Coding Plan 额度
1500 次/5小时滑动窗口
每次调用时倒算前 5 小时消耗
每天理论上限约 7200 次
⚠️ 注意:MiniMax 额度查询 API (/coding_plan/remains) 的数据可能不准确(窗口切换后不及时刷新)。判断额度是否可用,最可靠的方法是发一个真实测试请求。
均匀消耗技巧
问题:白天集中使用,凌晨浪费额度。
解决:用 cron 安排夜间自动任务:
01:00 代码质量扫描
02:00 文档评审
03:00 TODO 整理
04:00 系统健康巡检
06:00 学习材料准备
07:00 早间简报
额度监控
🔴 重要:额度查询 API 数据不可信!
MiniMax 的额度查询 API 存在惰性更新问题——窗口切换后如果没有新调用,计数器不会刷新,返回的数字可能是上一窗口的残留数据。
推荐做法:不看数字,只看真实请求能不能通。
# 判断额度是否可用的唯一可靠方法
curl -s https://api.minimaxi.com/v1/chat/completions \
-H "Authorization: Bearer <API_KEY>" \
-H "Content-Type: application/json" \
-d '{"model":"MiniMax-M2.1","messages":[{"role":"user","content":"test"}],"max_tokens":3}'
# 返回 choices → 可用 | 返回 429 → 额度耗尽
配合 OpenClaw cron 定期验证即可,无需频繁轮询 API 数字。
多 Agent 架构建议
起步(1 个 Agent)
main (默认 Agent) → 处理所有事务
进阶(3-4 个 Agent)
main → 日常对话、任务协调
coder → 代码开发(用编码专用模型)
assistant → 日程、邮件、提醒
config-bot → 系统运维(用高级模型)
模型分配原则
| Agent 类型 | 推荐模型 | 理由 |
|---|---|---|
| 日常对话 | MiniMax M2.1 | 包月,不心疼 |
| 代码开发 | MiniMax M2.1 / sf-coder-30b | 编码质量好 |
| 系统运维 | Claude Opus | 需要最高可靠性 |
| 轻量任务 | Qwen3-8B (免费) | 简单任务不浪费 |
故障排查速查
模型不响应
# 1. 检查系统状态
openclaw doctor
# 2. 查看错误日志
tail -20 ~/.openclaw/logs/gateway.err.log
# 3. 测试模型可用性
curl -s '<BASE_URL>/chat/completions' \
-H 'Authorization: Bearer <KEY>' \
-H 'Content-Type: application/json' \
-d '{"model":"<MODEL>","messages":[{"role":"user","content":"test"}],"max_tokens":5}'
Agent 不说话
- 检查 context 是否满了:
sessions_list看totalTokens/contextTokens - 超过 90% → 重置 session(删除 sessions.json 中对应条目)
- 检查 fallback 链是否配置正确
配置改错崩溃
# 恢复备份
cp ~/.openclaw/openclaw.json.backup.<TIMESTAMP> ~/.openclaw/openclaw.json
openclaw gateway restart
⚠️ 牢记:agents.defaults.models.<id> 只允许 alias 字段!一个非法字段 = 全面崩溃。
额度用完
- MiniMax Coding Plan:等 5 小时窗口重置
- SiliconFlow:充值或切到免费模型
- 临时方案:
/model sf-qwen3-8b(免费)
详细故障排查见:openclaw-troubleshooting(开发中)
💰 成本速算
| 方案 | 月费 | 每日可用次数 | 适合 |
|---|---|---|---|
| 纯免费 | ¥0 | 无限(质量有限) | 个人试玩 |
| MiniMax 包月 | ¥49 | ~7200 | 日常使用 |
| MiniMax + SF 余额 | ¥49 + ¥50 | 7200 + 按需旗舰 | 进阶用户 |
| 全家桶 | ¥49 + ¥50 + NewCLI | 全模型覆盖 | 重度用户 |
🔗 相关资源
- OpenClaw 文档:https://docs.openclaw.ai
- OpenClaw GitHub:https://github.com/openclaw/openclaw
- ClawHub 技能市场:https://clawhub.com
- 社区 Discord:https://discord.com/invite/clawd
Provider 配置技能
| 技能 | 安装命令 | 说明 |
|---|---|---|
| SiliconFlow | clawhub install add-siliconflow-provider | 98+ 模型,含免费 |
| MiniMax | clawhub install add-minimax-provider | ¥49/月包月方案 |
| NewCLI | clawhub install add-newcli-provider | Claude/GPT/Gemini |
注册链接
- SiliconFlow:https://cloud.siliconflow.cn/i/ihj5inat
- NewCLI (FoxCode):https://foxcode.rjj.cc/auth/register?aff=7WTAV8R
More by openclaw
View all skills by openclaw →You might also like
flutter-development
aj-geddes
Build beautiful cross-platform mobile apps with Flutter and Dart. Covers widgets, state management with Provider/BLoC, navigation, API integration, and material design.
drawio-diagrams-enhanced
jgtolentino
Create professional draw.io (diagrams.net) diagrams in XML format (.drawio files) with integrated PMP/PMBOK methodologies, extensive visual asset libraries, and industry-standard professional templates. Use this skill when users ask to create flowcharts, swimlane diagrams, cross-functional flowcharts, org charts, network diagrams, UML diagrams, BPMN, project management diagrams (WBS, Gantt, PERT, RACI), risk matrices, stakeholder maps, or any other visual diagram in draw.io format. This skill includes access to custom shape libraries for icons, clipart, and professional symbols.
ui-ux-pro-max
nextlevelbuilder
"UI/UX design intelligence. 50 styles, 21 palettes, 50 font pairings, 20 charts, 8 stacks (React, Next.js, Vue, Svelte, SwiftUI, React Native, Flutter, Tailwind). Actions: plan, build, create, design, implement, review, fix, improve, optimize, enhance, refactor, check UI/UX code. Projects: website, landing page, dashboard, admin panel, e-commerce, SaaS, portfolio, blog, mobile app, .html, .tsx, .vue, .svelte. Elements: button, modal, navbar, sidebar, card, table, form, chart. Styles: glassmorphism, claymorphism, minimalism, brutalism, neumorphism, bento grid, dark mode, responsive, skeuomorphism, flat design. Topics: color palette, accessibility, animation, layout, typography, font pairing, spacing, hover, shadow, gradient."
godot
bfollington
This skill should be used when working on Godot Engine projects. It provides specialized knowledge of Godot's file formats (.gd, .tscn, .tres), architecture patterns (component-based, signal-driven, resource-based), common pitfalls, validation tools, code templates, and CLI workflows. The `godot` command is available for running the game, validating scripts, importing resources, and exporting builds. Use this skill for tasks involving Godot game development, debugging scene/resource files, implementing game systems, or creating new Godot components.
nano-banana-pro
garg-aayush
Generate and edit images using Google's Nano Banana Pro (Gemini 3 Pro Image) API. Use when the user asks to generate, create, edit, modify, change, alter, or update images. Also use when user references an existing image file and asks to modify it in any way (e.g., "modify this image", "change the background", "replace X with Y"). Supports both text-to-image generation and image-to-image editing with configurable resolution (1K default, 2K, or 4K for high resolution). DO NOT read the image file first - use this skill directly with the --input-image parameter.
fastapi-templates
wshobson
Create production-ready FastAPI projects with async patterns, dependency injection, and comprehensive error handling. Use when building new FastAPI applications or setting up backend API projects.
Related MCP Servers
Browse all serversBoost your AI code assistant with Context7: inject real-time API documentation from OpenAPI specification sources into y
AI-driven control of live Chrome via Chrome DevTools: browser automation, debugging, performance analysis and network mo
Serena is a free AI code generator toolkit providing robust code editing and retrieval, turning LLMs into powerful artif
Beads — a drop-in memory upgrade for your coding agent that boosts context, speed, and reliability with zero friction.
Unlock seamless Figma to code: streamline Figma to HTML with Framelink MCP Server for fast, accurate design-to-code work
Supercharge your AI code assistant with GitMCP—get accurate, up-to-date code and API docs from any GitHub project. Free,
Stay ahead of the MCP ecosystem
Get weekly updates on new skills and servers.