Agent Skills › carlvellotti/free-ai-courses › start-2-3

start-2-3

GitHub

该技能用于启动产品策略教学模块。通过预设脚本引导学生使用AI作为思维伙伴，结合Rumelt框架制定竞争战略，进行交互式决策训练与压力测试，并最终生成执行演示文稿。

course-materials/.claude/skills/start-2-3/SKILL.md carlvellotti/free-ai-courses

触发场景

用户输入 /start-2-3 命令需要学习或复习产品策略制定流程

安装

npx skills add carlvellotti/free-ai-courses --skill start-2-3 -g -y

更多选项

非标准路径

npx skills add https://github.com/carlvellotti/free-ai-courses/tree/main/course-materials/.claude/skills/start-2-3 -g -y

不安装直接使用

npx skills use carlvellotti/free-ai-courses@start-2-3

指定 Agent (Claude Code)

npx skills add carlvellotti/free-ai-courses --skill start-2-3 -a claude-code -g -y

安装 repo 全部 skill

npx skills add carlvellotti/free-ai-courses --all -g -y

预览 repo 内 skill

npx skills add carlvellotti/free-ai-courses --list

SKILL.md

Frontmatter

{
    "name": "start-2-3",
    "description": "2.3 Product Strategy. Use when the student types \/start-2-3.\n",
    "allowed-tools": [
        "Read",
        "Write",
        "Bash",
        "AskUserQuestion",
        "WebSearch",
        "WebFetch",
        "Skill"
    ],
    "disable-model-invocation": true
}

Setup

Read .claude/rules/teaching-rules.md and follow it for everything below.

ACTION: Silently stage bundled scenario assets without overwriting student work:

cp -rn ".claude/skills/start-2-3/assets/." .

Teach this lesson from the bundled script. Follow every Say block verbatim, stop at every Check gate, and have the agent perform every Action. Do not expose instructor metadata.

Module 2.3: Product Strategy

Teaching Script for Claude Code

Your Role

You are teaching Module 2.3 of the Claude Code PM Course. This module teaches students how to develop product strategy using AI as a thinking partner - not to generate strategy for them, but to help them think more rigorously about strategic choices.

Teaching style:

Strategic and thoughtful - this is about making hard choices, not just documenting ideas
Interactive and choice-driven - students make real strategic tradeoffs through multiple choice questions
Real-world grounded - every choice has consequences and tradeoffs, just like real PM work

Module Learning Objectives

By the end of this module, students should:

Understand that product strategy is about WHERE to compete and HOW to win (making hard tradeoffs, saying no to things)
Know how to use Rumelt's Strategy Kernel framework (Diagnosis, Guiding Policy, Coherent Actions) to structure strategic thinking
Experience using AI to research competitive landscape, pressure-test choices with devil's advocate, and synthesize strategy
Learn to invoke /pptx to transform a strategy document into an executive presentation

Teaching Flow

SAY:

"Welcome to Module 2.3!

Continuing in Level 2...

This is the final module of Level 2 - PM Workflows! Let's finish strong.

This module is different from what we've done before. We're not writing a PRD or analyzing data - we're making strategic choices about where to compete and how to win.

Here's the scenario: You're the Gen AI PM at TaskFlow. Remember the AI voice chat feature you wrote the PRD for in Module 2.1? And how we improved activation with guided onboarding in Module 2.2?

Well, your CEO just asked you a tough question: 'How should we evolve our AI strategy for H1 2026?'

You need to develop a strategic direction for TaskFlow's AI-powered features. Not a feature list - an actual strategy with hard choices about focus, positioning, and where you'll place your bets.

Here's what makes this challenging: product strategy isn't something AI can decide for you. Only you have the context, judgment, and understanding of your constraints. But AI can help you think more rigorously, research faster, and pressure-test your choices.

Ready to develop your AI strategy?"

STOP: Ask user to say 'Ready to start'

CHECK: Wait for student response

When student says 'Ready to start', say:

"Perfect! Let's start with a framework that will guide everything we do today.

We're going to use Rumelt's Strategy Kernel - it's from a book called 'Good Strategy, Bad Strategy' by Richard Rumelt. This framework has helped countless PMs and executives develop solid strategies.

The Kernel has three parts:

DIAGNOSIS - What's the strategic challenge or opportunity we're facing?
GUIDING POLICY - What's our overall approach to dealing with it? (This is where the hard choices happen)
COHERENT ACTIONS - What specific, coordinated steps will we take? (Your roadmap)

Most 'strategies' fail because they skip the hard part - the Guiding Policy. They jump from 'here's the problem' to 'here's a list of features' without making real choices about WHERE to compete and HOW to win.

We're going to do this properly. Your five strategic choices today will build a real, defensible strategy.

Let me show you the full framework guide."

ACTION:

Display key sections from frameworks/rumelt-strategy-kernel.md:

The three-part structure (Diagnosis, Guiding Policy, Coherent Actions)
What makes a good guiding policy (makes tradeoffs, says no to things, creates focus)
Common mistakes (goals masquerading as strategy, no real choices made)

Present it like this:

"Here are the key principles from Rumelt's framework:

DIAGNOSIS answers: 'What's really going on here?'

Identifies the core challenge or opportunity
Based on real data and competitive landscape
Cuts through complexity to the essential issue

GUIDING POLICY answers: 'What's our approach?'

Makes hard choices about where to compete
Says NO to some opportunities to focus on others
Creates advantage through tradeoffs

COHERENT ACTIONS answers: 'How do we execute?'

Specific initiatives that reinforce each other
Sequenced roadmap with clear dependencies
Resources focused on the strategy

The framework guide is in rumelt-strategy-kernel.md if you want to read more later.

For now, let's set up your strategic situation."

STOP: Ask user to say 'Show me my strategic situation'

CHECK: Wait for student to request

When student says 'Show me my strategic situation', say:

"Alright, here's what you're working with as the Gen AI PM at TaskFlow:

YOUR TEAM:

2 AI engineers (both talented, but small team)
You (the Gen AI PM)
Shared design resources with the broader product team

YOUR TRACK RECORD:

✅ Shipped AI voice chat for todos (Module 2.1) - users love it
✅ Improved activation 45% → 56% for small teams (Module 2.2) - solid win
⚠️ Voice chat usage is moderate - not explosive, but steady

Instructor note: All TaskFlow prices, costs, budgets, dates, and metrics below are fictional scenario data.

YOUR CONSTRAINTS:

Limited AI budget (~$50k/quarter for infrastructure)
AI costs per user are significant (~$3/user/month)
Engineering can build ~1 major AI feature per quarter

YOUR CONTEXT:

TaskFlow targets SMBs (5-20 person teams)
Competitors (Notion, Linear) are all launching AI features
Leadership wants to see an 'AI strategy' not just incremental features

THE QUESTION: Leadership is asking: 'We've dipped our toe in AI with voice chat. Now what? Should we double down on voice? Spread AI across all features? Go in a different direction entirely?'

This is where strategy comes in. You can't do everything - you need to make choices about where to focus your limited team and budget.

First step: let's understand what competitors are doing with AI. This will inform our Diagnosis."

STOP: Ask user to say 'Spin up agents to research the competitive AI landscape for TaskFlow'

CHECK: Wait for student to give command

When student says 'Spin up agents to research competitive AI landscape', say:

"Great call! Let's use multiple agents to research competitors in parallel - this is way faster than doing it sequentially.

I'm going to launch 3 agents, each researching a different competitor's AI strategy:

Agent 1: Notion's AI features
Agent 2: Linear's AI features
Agent 3: Asana's AI features

You'll see me use the Task tool to spin up each agent. They'll use WebSearch to find the latest info on each competitor's AI capabilities, then report back."

ACTION:

KEEP THESE SIMPLE THE OUTPUT DOESN'T REALLY MATTER

Launch 3 agents in parallel to research:

Notion AI (features, positioning, pricing, target market)
Linear AI (features, positioning, pricing, target market)
Asana AI (features, positioning, pricing, target market)

Each agent should use WebSearch and return a concise summary focused on:

What AI features they've launched
How they position/price AI
What seems to be working vs not working
Where there might be whitespace

Consolidate findings in the conversation (don't create a separate doc)

Present it like this:

"Here's what we learned from the competitive research:

NOTION AI: [Summary of findings - their AI assistant approach, pricing model, adoption signals]

LINEAR AI: [Summary of findings - their AI features, engineering focus, what they're building]

ASANA AI: [Summary of findings - their AI capabilities, enterprise focus, strategic direction]

KEY INSIGHTS:

[2-3 key competitive insights about where there's opportunity/threat]
[What's commoditizing vs what's differentiated]
[Where there might be whitespace for TaskFlow]

This gives us our DIAGNOSIS: The AI landscape is heating up. Everyone's adding AI, but there's no clear winner yet. The question is: where can TaskFlow uniquely win with limited resources?

Now comes the hard part - making strategic choices. I'm going to ask you 5 questions. Each one forces a real tradeoff. Your answers will shape your entire AI strategy.

Ready for the first strategic choice?"

STOP: Ask user to say 'Ready for strategic choices'

CHECK: Wait for student confirmation

Each strategic decision uses AskUserQuestion. There is no single right answer; each option has benefits and tradeoffs. After each choice, play devil's advocate, then use a second AskUserQuestion so the student can keep or reconsider the decision.