Claude Code 个体创业全指南

问题验证提示词

I have a business idea: [DESCRIBE YOUR IDEA IN ONE PARAGRAPH]

Act as a brutal venture capitalist who has seen 10,000 pitches.
Your job is not to be encouraging. Your job is to find every reason 
this fails.

Tell me:
1. The three most likely reasons this does not work
2. The assumption the entire idea rests on that might be wrong
3. The existing solution customers are using that I am ignoring
4. The customer segment most likely to actually pay for this
5. The version of this idea that has the best chance of working

Be specific. Be harsh. Do not soften the feedback.

客户访谈脚本

I am interviewing potential customers for [PRODUCT DESCRIPTION].
My hypothesis about their problem: [YOUR HYPOTHESIS]

Generate 10 questions that would reveal whether my hypothesis is 
correct without leading the interviewee toward confirming it.

The questions should surface:
- What they currently do to solve this problem
- How much time and money the problem costs them
- What they have tried that did not work
- What a perfect solution would look like
- Whether they would pay for my specific approach

Never ask "would you use this product" — that question produces 
false positives. Ask about behavior, not intention.

驱动一切的CLAUDE.md

# [PROJECT NAME] — CLAUDE.md

## What We Are Building
[One clear sentence describing the product]

## The Customer
[Specific description of who this is for and what problem they have]

## MVP Scope
[List ONLY what the MVP needs to do. Nothing else.]

## Tech Stack
[Your chosen stack — keep it simple]

## Non-Negotiables
- Every feature must serve the MVP scope. No scope creep.
- Production-ready code only. No hacks that will need to be fixed later.
- Every new route and component must have error handling.
- Never store sensitive data without encryption.

## Definition of Done
[Specific description of what "finished" looks like for this MVP]

周五晚上：架构和设置（2小时）

Read my CLAUDE.md. Based on the MVP scope described, design the 
complete application architecture.

Tell me:
- The full folder structure
- Every database table needed with columns
- Every API route needed
- Every page/component needed
- The correct order to build things

Do not write any code yet. Just give me the architecture.
Wait for my confirmation before proceeding.

周六上午：核心功能（4小时）

Build the core user flow for [YOUR PRIMARY USE CASE].

Start with:
1. User authentication
2. [PRIMARY FEATURE 1]
3. [PRIMARY FEATURE 2]

Build in this exact order. Show me each piece working before moving 
to the next. Do not add anything not listed above.

周六下午：数据库和集成（3小时）

Add the following integrations in this order:
1. [PAYMENT INTEGRATION if applicable]
2. [EMAIL if applicable]
3. [ANY OTHER CRITICAL INTEGRATION]

For each integration:
- Handle all error states
- Add proper logging
- Test the happy path and the failure path

周六晚上：打磨和测试（2小时）

Do a full production readiness review.

Check for:
- Exposed environment variables
- Missing error handling
- Unhandled edge cases
- Security vulnerabilities
- Missing loading states

List every issue found ranked by severity. Fix all CRITICAL issues now.

周日：落地页和部署（3小时）

Build a complete landing page for [PRODUCT NAME].

Target customer: [DESCRIBE]
Their problem: [DESCRIBE]
Our solution: [DESCRIBE]
Proof it works: [DESCRIBE ANY EVIDENCE]
Price: [YOUR PRICE]

Include:
- A headline that names the problem
- Three specific benefits not features
- Social proof section
- How it works (3 steps)
- Pricing with a clear CTA
- FAQ (5 questions)

Write this for someone skeptical who needs to be convinced.

方法1：直接触达法

Write cold outreach for [PRODUCT] targeting [DESCRIBE CUSTOMER].

The message should:
- Open with something specific about their situation
- Name their exact problem in their language
- State what we do in one sentence
- Ask one small question to start a conversation

Under 75 words. No pitching. No links in the first message.
Do not use "I hope this finds you well" or any similar opener.

入门访谈

I just got my first 10 customers for [PRODUCT].
I need to interview each one to understand:
- Why they actually signed up (often different from why I think)
- What they were hoping to achieve
- What confused them during onboarding
- What almost made them not sign up
- What they would change about the product

Write 8 questions that would surface this information.
Make the questions feel conversational not clinical.

功能优先级系统

I have feedback from my first 10 customers for [PRODUCT].

Here is what they told me: [PASTE YOUR INTERVIEW NOTES]

Act as a product manager who has built 10 successful products.

Tell me:
1. The single highest-leverage feature I could add based on this feedback
2. The feature customers requested that I should NOT build and why
3. What the feedback reveals about who my actual customer is
4. The change to the existing product that would have the biggest impact
5. What I should stop doing based on this feedback

Be specific. Give me one recommendation per category, not a list.

预测一切的留存指标

Build a simple retention tracking system for [PRODUCT].

I need to know:
- Which users signed up in the last 30 days
- Which of them logged in again after day 7
- Which of them logged in again after day 14
- The actions that correlate with users who return vs users who churn

Use [YOUR STACK] and deposit a daily retention report in my email.

支持系统

Build a customer support system for [PRODUCT].

It should:
- Read incoming support emails
- Categorize each request as: ROUTINE, CUSTOM, or ESCALATE
- Generate a draft response for ROUTINE requests from the FAQ
- Flag CUSTOM requests for my review within 2 hours
- Always escalate BILLING and CANCELLATION requests immediately

Connect to: [YOUR EMAIL]
Use: [YOUR STACK]

收入追踪系统

Build a daily revenue dashboard for [PRODUCT].

Show me:
- MRR (monthly recurring revenue)
- New MRR today
- Churned MRR this month
- Net MRR growth
- Number of active customers
- Trial to paid conversion rate

Pull data from: [STRIPE or PAYMENT PROCESSOR]
Deliver to: [EMAIL or SLACK]

内容引擎

I have interview notes from my first 10 customers: [PASTE NOTES]

Generate 10 content ideas based on:
- Problems they had before finding my product
- Results they got after using it
- Insights they shared about their workflow
- Misconceptions they had that the product corrected

Frame each idea as something a potential customer would find 
valuable whether they buy or not.