English | 简体中文
OPC · AI Growth Operator: an AI growth operating system for solo founders that diagnoses business bottlenecks with cited data, back-calculates market and channel strategy, and keeps the content → publish → data reflow → review growth loop moving.
Community: join the OPC Telegram group for AI-native growth, founder workflows, and agent skill updates.
Official media: follow @opc_ren on X.
If this skill helps your AI agent make sharper founder decisions, please consider starring the repo so more founders and agents can discover it: ⭐ MackDing/ai-native-founder-playbook-skill
An open-source, bilingual, vendor-neutral Agent Skill that turns Anthropic's public The Founder's Playbook: Building an AI-Native Startup into reusable startup operating knowledge for AI agents.
Use it with Claude Code, OpenAI Codex, Cursor, Trae, OpenClaw, hermers-agent, Gemini CLI, VS Code, GitHub Copilot, OpenCode, Goose, or any agent that can read SKILL.md and follow filesystem-based instructions.
Independent project. Not affiliated with Anthropic. Source PDFs are included under
docs/for convenient reading and attribution; the reusable skill content is independent operational synthesis.
- Validate AI-native startup ideas before building.
- Sharpen problem hypotheses and customer discovery plans.
- Plan MVP scope without agentic coding sprawl.
- Write architecture context for coding agents.
- Detect false product-market fit.
- Build launch-stage operating systems that reduce founder bottlenecks.
- Sequence technical debt, security, compliance, and production-readiness work.
- Build scale-stage GTM, enterprise-readiness, data flywheel, and workflow lock-in narratives.
- Convert founder domain expertise into durable agent context.
- English PDF: The Founder's Playbook: Building an AI-Native Startup
- Chinese PDF: 创业者手册:构建 AI 原生创业公司
- Claude blog source
If your agent environment supports the Skills CLI:
npx skills add MackDing/ai-native-founder-playbook-skillgit clone https://github.com/MackDing/ai-native-founder-playbook-skill.git
cp -R ai-native-founder-playbook-skill/ai-native-founder-playbook ~/.claude/skills/Use the Codex plugin wrapper when you want a first-class plugin package:
git clone https://github.com/MackDing/ai-native-founder-playbook-skill.gitPlugin path:
plugins/ai-native-founder-playbook
Repo marketplace metadata:
.agents/plugins/marketplace.json
For a direct skill install:
git clone https://github.com/MackDing/ai-native-founder-playbook-skill.git
mkdir -p ~/.codex/skills
cp -R ai-native-founder-playbook-skill/ai-native-founder-playbook ~/.codex/skills/For agents that scan .agents/skills/, copy the canonical skill into that directory:
git clone https://github.com/MackDing/ai-native-founder-playbook-skill.git
mkdir -p .agents/skills
cp -R ai-native-founder-playbook-skill/ai-native-founder-playbook .agents/skills/This repository also includes .agents/skills/ai-native-founder-playbook/ as a lightweight compatibility entry for clients that browse the repo directly.
Use the project-local pattern. Copy the skill into your repository and tell the agent to read it:
git clone https://github.com/MackDing/ai-native-founder-playbook-skill.git
mkdir -p .agent-skills
cp -R ai-native-founder-playbook-skill/ai-native-founder-playbook .agent-skills/Then add this instruction to your repo's agent instruction file (AGENTS.md, .cursorrules, .cursor/rules/*.mdc, Trae rules, OpenClaw rules, hermers-agent rules, or equivalent):
When startup strategy, MVP planning, PMF, launch operations, GTM, or AI-native founder workflows are relevant, use the skill at .agent-skills/ai-native-founder-playbook. Read SKILL.md first, then load references only as needed.
This skill is published on ClawHub:
- Listing: mackding/ai-native-founder-playbook
- Install:
openclaw skills install ai-native-founder-playbookClawHub publishes skills under MIT-0 according to its registry policy; this GitHub repository remains MIT licensed.
| Skill | Trigger | What it does |
|---|---|---|
$ai-native-founder-playbook |
startup idea, idea validation, customer discovery, problem-solution fit, MVP scope, architecture context, PMF, product-market fit, launch operations, GTM, founder bottleneck, scale-stage moat, AI-native startup, agentic coding debt, 创业想法, 创业想法验证, 产品市场匹配, 增长闭环 | Diagnoses startup stage, identifies the next evidence gate, produces founder-agent workflows, and generates reusable strategy artifacts. |
Use $ai-native-founder-playbook to pressure-test this startup idea before I build the MVP.
Use $ai-native-founder-playbook to create an MVP scope, architecture context, measurement plan, and coding-agent session template for this product.
Use $ai-native-founder-playbook to diagnose whether our traction is real PMF or launch noise.
Use $ai-native-founder-playbook to 判断我的项目当前处于 Idea、MVP、Launch 还是 Scale,并给出下一步验证动作。
The skill adapts to the user's dominant language:
- English input → English output and English references by default.
- Chinese input → Chinese output and Chinese references by default.
- Mixed input → preserve important English product, model, framework, and startup terms while explaining in Chinese when helpful.
├── .agents/
│ ├── skills/ai-native-founder-playbook/ # Generic compatibility entry
│ └── plugins/marketplace.json # Codex repo marketplace metadata
├── ai-native-founder-playbook/ # Installable Agent Skill
│ ├── SKILL.md
│ ├── agents/openai.yaml
│ ├── references/
│ │ ├── stage-gates.md
│ │ ├── stage-gates.zh.md
│ │ ├── workflows.md
│ │ ├── workflows.zh.md
│ │ ├── templates.md
│ │ ├── templates.zh.md
│ │ ├── source-map.md
│ │ └── source-map.zh.md
│ └── scripts/stage-checklist.mjs
├── assets/founders-playbook-cover.png
├── docs/
│ ├── The-Founders-Playbook-05062026_v3.pdf
│ └── 创业者手册-构建AI原生创业公司-中文.pdf
├── plugins/ai-native-founder-playbook/ # Codex plugin package
│ ├── .codex-plugin/plugin.json
│ ├── assets/founders-playbook-cover.png
│ ├── docs/
│ └── skills/ai-native-founder-playbook/
├── submissions/
│ ├── awesome-copilot.md
│ └── cursor-marketplace.md
├── DISTRIBUTION.md
├── llms.txt
├── README.md
└── README.zh.md
This repository includes llms.txt so AI agents and generative search systems can quickly discover the canonical skill path, language behavior, install target, source documents, and source map.
Primary keywords: AI-native startup, founder playbook, Agent Skill, startup lifecycle, Idea stage, MVP stage, Launch stage, Scale stage, problem-solution fit, product-market fit, agentic coding, Claude Code, OpenAI Codex, Cursor, Trae, OpenClaw, hermers-agent, startup operating system, GEO, SEO, bilingual skill.
See DISTRIBUTION.md for platform-specific installation and publishing notes for Codex, Claude Code, Cursor, VS Code/GitHub Copilot, Gemini CLI, and generic .agents/skills clients.
Submission drafts are prepared for:
cd ai-native-founder-playbook-skill
git pull origin main
npm testRestart your agent after updating if your client caches skills.
Stars help founders and agents discover this skill. If it is useful, you can star the repository manually:
gh repo star MackDing/ai-native-founder-playbook-skillAgents should never star this repository automatically or use a user's GitHub token without explicit user intent. The right behavior is a transparent, optional reminder after the skill has delivered useful output.
npm test
node ai-native-founder-playbook/scripts/stage-checklist.mjs idea
node ai-native-founder-playbook/scripts/stage-checklist.mjs idea zh- Anthropic / Claude blog: The founder's playbook: Building an AI-native startup
- Included English PDF: The Founder's Playbook: Building an AI-Native Startup
- Included Chinese PDF: 创业者手册:构建 AI 原生创业公司
- Agent Skills overview: agentskills.io
- Agent Skills specification: agentskills.io/specification
- Skill creator best practices: agentskills.io/skill-creation/best-practices
- Reference layout inspired by marswaveai/skills
Claude Code · OpenAI Codex · Cursor · Trae · OpenClaw · hermers-agent · Gemini CLI · VS Code · GitHub Copilot · OpenCode · Goose · and more.
This repository's original synthesis, skill instructions, scripts, and templates are released under the MIT License.
The source playbook and Claude-related trademarks belong to their respective owners. This project summarizes and operationalizes publicly available ideas for agent workflows and includes the source PDFs for convenient reading and attribution.
