Vertical Project Plan Generator

Use Claude Code to auto-generate vertical-specific project plans.

Introduction

30 Second Summary

Every team has someone who holds all the institutional knowledge about clients, verticals, and deliverables. When that person is unavailable, business documents come out generic and useless.

In this project, you will use Claude Code to encode your company's expertise into a context file and a reusable Skill, so any team member can generate structured, vertical-specific project plans with a single command.

What You'll Build

You open Terminal, navigate to your project folder, type /generate-project-plan healthcare, and get back a complete project plan pre-filled with your company's services, realistic timelines, and vertical-specific risks.

By the end of this project, you'll have:

A one-command workflow where you type /generate-project-plan followed by any vertical and receive a complete, structured project plan your team can hand to a client.
A CLAUDE.md context file that encodes your company's agentic AI and data orchestration expertise, so Claude Code produces business-specific output instead of generic filler.
A reusable Skill that any team member can invoke without needing to know prompt engineering or hold your domain expertise in their head.
Secret Mission: Create a discovery call preparation skill that generates a brief with qualifying questions and vertical-specific pain points.

Are there any prerequisites?

You need an active Claude Max subscription ($200/month plan) and a macOS computer. Basic terminal familiarity with commands like cd and ls is expected.

Before We Start

Before we dive in, let's lock in what you're about to build and why it matters for your team.

Install Claude Code and Create Your Workspace

Claude Code is a command-line tool that lets you interact with Claude directly from your terminal. Before you can use it to generate anything, you need it installed and running in a dedicated project workspace.

In this step, you'll install Claude Code, create a folder for your project plan system, and confirm that Claude Code can read your environment.

In this step, get ready to:

Install Claude Code on your Mac.
Create a dedicated project workspace folder.
Verify Claude Code can read your environment.

Install or verify Claude Code

Claude Code runs inside Terminal on your Mac. You'll install it using a single command from Anthropic's official installer.

Press Cmd+Space (macOS) to open your search bar.
Type Terminal and press Enter to open it.
Install Claude Code by running this command:

curl -fsSL https://claude.ai/install.sh | bash

What does this command do?

curl downloads the official Claude Code installer from Anthropic.
The pipe (|) passes the downloaded script directly to bash, which runs the installer.
If Claude Code is already installed, this updates it to the latest version.

Once the installer finishes, you need to open a new Terminal window so your session recognizes the updated path.

Close your current Terminal window.
Open a new Terminal window by pressing Cmd+N or repeating the Cmd+Space search.
Verify the installation by running:

claude --version

You should see a version number printed, such as 2.1.x.

✔️ I see a version number

Claude Code is installed and ready to use. Move on to the next substep.

ⓧ Command not found

Your terminal session does not have Claude Code on its path yet. Try these fixes:

Make sure you opened a brand new Terminal window after the install completed.
If you're using Zsh (the macOS default), run source ~/.zshrc to reload your shell configuration.
If the problem persists, re-run the install command above and watch for any error messages during installation.
Confirm your macOS version is 13.0 or higher by clicking the Apple menu and selecting About This Mac.

Create your project workspace and start Claude Code

You need a dedicated folder where Claude Code will look for context files and skills. This folder becomes your project's home base.

Create the vertical-playbooks folder and navigate into it by running these commands:

mkdir -p ~/vertical-playbooks
cd ~/vertical-playbooks

Now you'll start Claude Code inside this folder. On first launch, Claude Code opens your browser for authentication.

Start Claude Code by running:

claude

What happens on first launch?

Claude Code automatically opens your browser to an Anthropic login page. Sign in with the same credentials you use for your Claude Max subscription.

If the browser does not open automatically, press c in Terminal to copy the login URL, then paste it into your browser manually.

Complete the browser authentication using your Claude credentials.

After authenticating, you'll see the Claude Code interactive prompt appear in your Terminal. This means Claude Code is ready to accept commands.

Not seeing the interactive prompt?

Check that your internet connection is active. Claude Code requires a connection to authenticate.
Confirm you signed in with the account that has your Max subscription. Pro or Max plans are required for Claude Code access.
If your browser showed a login code instead of redirecting, copy that code and paste it back into the Terminal prompt.

Still stuck? Help me troubleshoot Claude Code authentication

Confirm Claude Code is working

The interactive prompt is open, but let's confirm Claude Code can actually read your environment before moving on.

Before you type the next command, what do you think Claude Code will respond when asked about your current directory?

Type this into the Claude Code prompt:

What directory am I in?

Claude Code confirms you are in ~/vertical-playbooks. This proves it can read your file system and respond with accurate information about your environment.

Exit the Claude Code session by typing:

/exit

Claude Code shows a different directory?

Make sure you ran cd ~/vertical-playbooks before starting claude. Claude Code inherits the directory your terminal was in when you launched it.
If you see your home directory instead, exit with /exit, run cd ~/vertical-playbooks, then start claude again.

Need more help? Help me get Claude Code running in the right directory

Your workspace is set up and Claude Code is ready. Next up, you'll see what happens when you ask Claude Code to generate a project plan without any business context.

Experience the Context Gap

Claude Code is installed and your project workspace is ready. Now it's time to see what happens when you ask Claude Code to generate a business document without giving it any information about your company.

Your team's challenge right now is that AI output is too generic to be useful for your specific verticals. This step makes that problem visible so you can feel exactly why context matters before we fix it.

In this step, get ready to:

Request a full project plan from Claude Code with zero business context.
Observe how the output is too generic to hand to a client or team member.
Save the generic output for comparison later.

Start Claude Code and request a project plan

Right now, your vertical-playbooks folder is empty. There is no CLAUDE.md file, no company information, nothing. Claude Code starts every session with a fresh context window, so it only knows what you tell it or what it finds in the project folder.

Before you type anything, think about this: what does Claude Code currently know about your company, your services, or your verticals? What kind of project plan do you expect it to produce for a healthcare client with zero information about what you actually deliver?

Switch back to your Terminal and navigate to your project folder by running this command:

cd ~/vertical-playbooks

Start a Claude Code interactive session by running:

claude

The Claude Code interactive prompt appears, ready to accept your input.

Type this prompt into the Claude Code session and press Enter:

Generate a project plan for a healthcare client engagement. Include: executive summary, objectives, scope, deliverables, timeline, budget, resources and roles, risk management, communication plan, quality assurance, acceptance criteria, and project closure.

Claude Code generates a full project plan. Wait for it to finish before moving on.

What just happened?

You gave Claude Code a prompt with no company context. It has no idea what services you offer, what verticals you specialize in, or how your team delivers engagements. It produced the best generic answer it could with zero information about your business.

Observe how generic the output is

Read through what Claude Code just generated. Pay attention to how useful this would actually be if you handed it to a team member or a client.

Look at the scope section. It likely says something vague like "implement a solution" or "deliver a platform."
Check for any mention of agentic AI, data orchestration, or your specific service offerings. You will not find them.
Look at the risks. They are generic project management risks that apply to any consulting engagement.
Check the deliverables. They could describe any technology project for any company in any industry.

This is the context gap. Without your business DNA, Claude Code produces filler that no one on your team can use.

Save the generic output and exit

You will compare this output to a much better version later, so save it to a file now.

Type this into your Claude Code session to save the output:

Save that output to a file called generic-plan.md

Claude Code creates generic-plan.md in your vertical-playbooks folder.

Claude Code asking for permission?

Claude Code may ask for permission to write files. If you see a permission prompt, approve it so the file gets created.

If you are not sure what to do, help me with the Claude Code permission prompt.

Exit the Claude Code session by typing:

/exit

You are back at your normal Terminal prompt.

Confirm the file was saved by running:

ls

You should see generic-plan.md listed in the output.

Don't see generic-plan.md?

Make sure you are still in the ~/vertical-playbooks directory. Run pwd to check your current location.

If the file is missing, start Claude Code again with claude and re-request the save. Help me find my generic-plan.md file.

You just proved the problem. The same AI that will later produce business-specific output gave you generic filler because it had zero context about your company. Next up, you will fix this by giving Claude Code your business DNA.

Encode Your Business DNA with CLAUDE.md

In the previous step, you saw Claude Code produce a project plan that could have come from any company on Earth. The scope was vague, the risks were generic, and there was zero mention of agentic AI or data orchestration.

The fix is a single file called CLAUDE.md. This is a special file that Claude Code reads at the start of every session, giving it persistent context about your business. One file transforms all future output from generic filler to business-specific content.

In this step, get ready to:

Create a CLAUDE.md file encoding your company's services, verticals, and document standards.
Verify that Claude Code automatically loads your context file at the start of a session.

Create your CLAUDE.md file

CLAUDE.md lives in the root of your project folder. Claude Code automatically reads it when you start a session in that folder, so anything you put here becomes persistent knowledge across every conversation.

Make sure you are in your project folder by running this command:

cd ~/vertical-playbooks

Create the first part of your CLAUDE.md file by running this command:

cat > CLAUDE.md << 'EOF'
# Company Context

## Who We Are
[Your company name] provides agentic AI and data orchestration services to enterprise clients. We design, build, and deploy autonomous AI agent systems that integrate with existing data infrastructure to automate complex workflows.

## Core Services
- Agentic AI System Design: Architecture and deployment of autonomous AI agents that execute multi-step business processes
- Data Orchestration: Integration pipelines that connect siloed data sources into unified, AI-ready platforms
- Vertical-Specific Solutions: Pre-built agent frameworks tailored to industry-specific workflows and compliance requirements
- Managed AI Operations: Ongoing monitoring, optimization, and governance of deployed AI systems

## Verticals We Serve
- Healthcare: Clinical workflow automation, patient data orchestration, compliance-aware AI agents (HIPAA), EHR integration
- Financial Services: Risk assessment automation, regulatory reporting orchestration, fraud detection agents, multi-system data reconciliation
- Manufacturing: Supply chain orchestration, predictive maintenance agents, quality control automation, IoT data pipeline integration

## How We Engage
- Typical engagement: 12-16 weeks from discovery to production deployment
- Team structure: Solution Architect, Data Engineer(s), AI/ML Engineer(s), Project Manager, Client Success Lead
- Tech stack: Python, cloud-native (AWS/Azure/GCP), vector databases, LLM orchestration frameworks, Apache Airflow/Prefect for data pipelines
- Delivery model: Agile sprints with bi-weekly client demos
EOF

What does this content do?

You just defined four sections that tell Claude Code who your company is:

Who We Are positions your company as an agentic AI and data orchestration provider, not a generic consultancy.
Core Services lists the four service lines so Claude references them in deliverables and scope.
Verticals We Serve gives Claude industry-specific details like compliance frameworks (HIPAA, EHR) to weave into plans.
How We Engage sets realistic timelines, team roles, and tech stack so generated plans are grounded in how you actually deliver.

Add the remaining sections (your document template standards and differentiators) by running this command:

cat >> CLAUDE.md << 'EOF'

## Project Plan Standards
All project plans must follow this template structure:
1. Executive Summary
2. Objectives (measurable, with KPIs)
3. Scope (explicit in-scope and out-of-scope)
4. Deliverables (table format with descriptions and due dates)
5. Timeline (phased with date ranges)
6. Budget (broken down by category)
7. Resources & Roles
8. Risk Management (table with impact, likelihood, mitigation)
9. Communication Plan
10. Quality Assurance
11. Acceptance Criteria
12. Project Closure

## Key Differentiators
- We do not build chatbots. We build autonomous agents that execute end-to-end business processes.
- Our data orchestration layer is what makes the AI agents effective. Without clean, connected data, agents cannot act autonomously.
- Compliance is built into the architecture from day one, not bolted on after.
EOF

What do these two sections define?

Project Plan Standards tells Claude Code the exact 12-section structure every project plan must follow. Without this, Claude invents its own format each time.

Key Differentiators prevents Claude from defaulting to chatbot language or generic AI descriptions. It forces output to reflect what your company actually builds: autonomous agents on top of connected data.

Confirm the file was created by running:

ls CLAUDE.md

You should see CLAUDE.md listed in the output.

Don't see CLAUDE.md listed?

Make sure you ran cd ~/vertical-playbooks first. If you are in a different folder, the file was created there instead.
Check that you copied both commands fully, including the EOF line at the end. A missing EOF means the terminal is still waiting for input.
If your terminal prompt looks stuck (no $ sign), type EOF on a new line and press Enter to close the heredoc.

help me troubleshoot my CLAUDE.md file creation

Customize this with your real company details

The content above uses placeholder text for your company name and representative verticals. Before moving on, open CLAUDE.md in any text editor and replace the bracketed sections with your actual company name, real verticals, and true service descriptions.

The more specific and accurate this file is, the better Claude Code's output becomes. You can always refine it later as you learn what produces the best results.

✔️ Awesome, I've got everything!

Great. Make sure you have saved any customizations to CLAUDE.md.

ⓧ I'd like to double check the full code

Your CLAUDE.md file is built from the two chunks above. Scroll up and confirm you have all of these sections present, in this order:

# Company Context (top-level heading).
## Who We Are (one paragraph about your company).
## Core Services (four bullet items).
## Verticals We Serve (three bullet items: Healthcare, Financial Services, Manufacturing).
## How We Engage (four bullet items: engagement length, team, tech stack, delivery model).
## Project Plan Standards (numbered list of 12 sections).
## Key Differentiators (three bullet items).

You can view the full file in your terminal by running:

cat CLAUDE.md

Verify Claude Code loads your context

Creating the file is only half the job. You need to confirm Claude Code actually picks it up when you start a session.

Launch Claude Code inside your project folder by running:

claude

Before you check the context, what do you think /context will show you?

Once the interactive prompt appears, type:

/context

You should see CLAUDE.md listed under the Memory files section. This confirms Claude Code has loaded your business context and will use it in every response during this session.

CLAUDE.md not showing in /context?

Make sure you launched claude from inside the ~/vertical-playbooks folder. Claude Code loads CLAUDE.md from your current working directory.
Verify the file is named exactly CLAUDE.md (all caps, with the .md extension). A lowercase claude.md will not be detected.
Try exiting with /exit and relaunching claude again. The file is read at session start, not mid-session.

help me get Claude Code to recognize my CLAUDE.md file

Your business context is locked in. Next up, you will re-run the exact same project plan request and see how dramatically the output changes when Claude Code actually knows who you are.

Generate a Context-Aware Project Plan

In the previous step, you encoded your company's services, verticals, engagement model, and document standards into CLAUDE.md. Claude Code now knows your business at the start of every session.

Now you'll type the exact same project plan prompt that produced generic filler in Step 2. The only difference is that CLAUDE.md is loaded. This is the moment you feel why persistent context transforms AI output from useless to useful.

In this step, get ready to:

Generate a healthcare project plan using your CLAUDE.md context.
Save the output and compare it against the generic version from earlier.
See the concrete difference that business context makes to AI output.

Request the same project plan again

You used this exact prompt in Step 2 and got back vague, generic content. Before you run it again, think about what you expect to see differently now that CLAUDE.md is loaded.

What specific details do you think Claude Code will include this time? Which sections of your template will look most different?

In your Claude Code session, type the following prompt:

Generate a project plan for a healthcare client engagement. Include: executive summary, objectives, scope, deliverables, timeline, budget, resources and roles, risk management, communication plan, quality assurance, acceptance criteria, and project closure.

Read through what Claude Code generates. You should immediately notice it references your specific services, your team structure, and healthcare-specific compliance requirements.

What should I see?

The output should look dramatically different from the generic version. Look for these specific indicators that Claude Code is using your context:

The executive summary mentions agentic AI and data orchestration specifically.
Objectives reference measurable outcomes tied to your services (e.g., deploying autonomous clinical workflow agents).
Scope mentions HIPAA compliance, EHR integration, and your tech stack.
Resources list your actual team roles (Solution Architect, Data Engineer, AI/ML Engineer, Project Manager, Client Success Lead).
The structure follows your 12-section template.

Output still looks generic?

Type /context to confirm CLAUDE.md is listed as loaded context. If it is not showing, exit with /exit and restart Claude Code from the ~/vertical-playbooks directory.
Make sure you are running Claude Code from inside ~/vertical-playbooks where your CLAUDE.md file lives. Claude Code reads context from the current directory.
help me figure out why Claude Code is not picking up my CLAUDE.md file

Save the context-aware plan

You now have a project plan that actually reflects your business. Save it so you can compare it against the generic version.

Type this in your Claude Code session to save the output:

Save that to a file called healthcare-plan.md

Claude Code saves the plan to healthcare-plan.md inside your ~/vertical-playbooks folder.

Compare the two plans side by side

You have two files now. generic-plan.md is what Claude Code produced with no context. healthcare-plan.md is what it produced with your CLAUDE.md loaded. Let Claude Code make the contrast explicit.

Ask Claude Code to generate a comparison by typing:

Compare generic-plan.md and healthcare-plan.md. Summarize the key differences in a bullet list.

What does this comparison prove?

The same tool, the same prompt, completely different output. The only variable that changed was context.

This is the core insight of the project. Claude Code is not smarter in one session vs the other. It simply has access to your business knowledge through CLAUDE.md. The better you encode your expertise in that file, the more useful every future interaction becomes.

Comparison not showing meaningful differences?

If generic-plan.md was not saved properly in Step 2, Claude Code has nothing meaningful to compare against. Re-run the generic prompt (without CLAUDE.md context) and save it again.
If both files exist but differences are subtle, check that your CLAUDE.md has specific content about your verticals and services. Generic context produces slightly-less-generic output.
help me understand why my two plans look similar

You just proved that context is everything. A single file turned Claude Code from a generic text generator into a tool that knows your business. But right now, generating a good plan still requires knowing what to type. In the next step, you'll package this into a reusable Skill that any team member can invoke with a single command.

Build the Reusable Skill

You just proved that CLAUDE.md transforms generic output into business-specific output. But there is still a problem. Every time someone on your team wants a project plan, they need to know the exact prompt to type.

A Skill in Claude Code solves this by packaging your prompt into a single slash command. Any team member can type /generate-project-plan healthcare and get a structured plan without needing to know prompt engineering.

In this step, get ready to:

Create a reusable Skill that packages your project plan prompt into a one-command interface.
Test the Skill across multiple verticals to prove it adapts automatically.

Create the skill directory and file

Skills live in a specific folder structure inside your project. Claude Code watches this directory and picks up new skills automatically.

In your Claude Code session, type the following to exit:

/exit

Create the skill directory and write the SKILL.md file by running these commands in Terminal:

mkdir -p ~/vertical-playbooks/.claude/skills/generate-project-plan
cat > ~/vertical-playbooks/.claude/skills/generate-project-plan/SKILL.md << 'EOF'
---
description: Generate a structured project plan for a client engagement in a specific vertical. Use when someone asks for a project plan, engagement plan, SOW outline, or proposal framework for any vertical we serve.
---

## Instructions

Generate a complete project plan for a **$ARGUMENTS** client engagement.

The plan must:
- Follow our 12-section project plan template (defined in CLAUDE.md under Project Plan Standards)
- Reference our specific services (agentic AI system design, data orchestration, vertical-specific solutions, managed AI operations)
- Include vertical-specific compliance requirements, risks, and integration points
- Use realistic timelines based on our typical 12-16 week engagement model
- Include our standard team roles and delivery methodology
- Have measurable objectives with KPIs tied to the client's industry

Format the output as a clean markdown document ready to share with the team or adapt into a client-facing proposal.

If the vertical is not one we currently serve (Healthcare, Financial Services, Manufacturing), note this and generate the plan using the closest analogous vertical as a foundation, flagging areas that need validation.
EOF

What does this file do?

The --- frontmatter block contains a description field. Claude Code uses this description to know when the skill is relevant.

The $ARGUMENTS placeholder gets replaced with whatever you type after the slash command. So /generate-project-plan healthcare substitutes "healthcare" into the instructions.

The instructions reference your CLAUDE.md file for the 12-section template. This means the Skill and your context file work together to produce consistent output.

Confirm the skill file was created by running:

ls ~/vertical-playbooks/.claude/skills/generate-project-plan/

You should see SKILL.md listed in the output.

Don't see SKILL.md?

Make sure you ran the full cat command including the EOF line at the end. If you see an error about a missing directory, run the mkdir -p command again first.

help me troubleshoot why my SKILL.md file was not created in the .claude/skills directory

Why project-level skills instead of personal skills?

Skills can live in two places: .claude/skills/ inside a project folder, or ~/.claude/skills/ in your home directory.

Project-level skills can be committed to a Git repo and shared across your entire team. Personal skills only exist on one machine. For a team tool like this, project-level is the right choice.

✔️ Awesome, I've got everything!

Great. Make sure the file is saved and you are ready to test it.

ⓧ I'd like to double check the full code

Your SKILL.md file at ~/vertical-playbooks/.claude/skills/generate-project-plan/SKILL.md should contain exactly this:

---
description: Generate a structured project plan for a client engagement in a specific vertical. Use when someone asks for a project plan, engagement plan, SOW outline, or proposal framework for any vertical we serve.
---

## Instructions

Generate a complete project plan for a **$ARGUMENTS** client engagement.

The plan must:
- Follow our 12-section project plan template (defined in CLAUDE.md under Project Plan Standards)
- Reference our specific services (agentic AI system design, data orchestration, vertical-specific solutions, managed AI operations)
- Include vertical-specific compliance requirements, risks, and integration points
- Use realistic timelines based on our typical 12-16 week engagement model
- Include our standard team roles and delivery methodology
- Have measurable objectives with KPIs tied to the client's industry

Format the output as a clean markdown document ready to share with the team or adapt into a client-facing proposal.

If the vertical is not one we currently serve (Healthcare, Financial Services, Manufacturing), note this and generate the plan using the closest analogous vertical as a foundation, flagging areas that need validation.

Test the skill

Time to see if your one-command interface actually works. You will invoke the skill the same way any team member would.

Start Claude Code in your project folder by running:

cd ~/vertical-playbooks
claude

Once the interactive prompt appears, you are ready to invoke your skill.

Before you run the skill, think about this: what do you expect will be different about this output compared to typing a long prompt manually?

Invoke the skill by typing:

/generate-project-plan healthcare

Claude Code produces a full project plan using your CLAUDE.md context and the skill's instructions. The output should reference agentic AI, data orchestration, HIPAA compliance, EHR integration, and follow your 12-section template.

The key difference is that you typed one short command instead of a multi-sentence prompt. Any team member can produce the same quality output without needing your domain expertise.

Skill not recognized?

If Claude Code does not recognize /generate-project-plan as a skill, make sure you are running claude from inside the ~/vertical-playbooks directory. Claude Code only detects skills in the current project folder.

Also confirm your file is named exactly SKILL.md (all caps) inside .claude/skills/generate-project-plan/.

help me figure out why Claude Code is not recognizing my generate-project-plan skill

Test with a different vertical

A good skill adapts its output based on the argument you pass. The real proof is whether it shifts compliance requirements, integration points, and risks when you change the vertical.

What do you expect to change when you switch from healthcare to financial services? Think about which regulations, systems, and risks are different.

Invoke the skill with a different vertical by typing:

/generate-project-plan financial services

The output shifts to match the financial services vertical. You should see SOX and PCI-DSS compliance instead of HIPAA, trading system and reconciliation platform integration instead of EHR, and financial-services-specific risks throughout.

What just happened?

The same skill, the same CLAUDE.md context, and a different argument produced a completely different (and equally specific) project plan. This is the power of encoding your business knowledge into reusable workflows.

Your team now has a standardized interface for generating project plans. No one needs to remember what to ask or how to phrase it.

Secret mission

Build a Discovery Call Prep Skill

Project plans happen after a deal is in motion. But what about the calls that start the deal? Build a second Skill that generates a discovery call preparation brief for any vertical, complete with qualifying questions, pain points to probe, and disqualifiers. Give your less experienced team members the confidence to walk into prospect calls fully prepared.

Clean Up Your Resources

This project runs entirely locally on your machine. There are no cloud resources, no running servers, and no ongoing costs beyond your existing Claude Max subscription.

Decide whether to keep your project workspace, share it with your team, or remove it entirely.

Resources you used:

~/vertical-playbooks/ directory containing CLAUDE.md, skills, and generated plans.

Keep everything running

No action needed. Choose this if you are still actively refining your CLAUDE.md, adding new skills, or onboarding team members.

The ~/vertical-playbooks/ folder stays on your machine and costs nothing to keep.
Any team member can copy this folder to their own machine, install Claude Code, and immediately use your skills.
Consider pushing the folder to a Git repo so the whole team shares one source of truth for your CLAUDE.md and skills.

Pause - I'll come back to this later

There are no running processes to shut down. Your project folder is already at rest on disk.

Your ~/vertical-playbooks/ folder will be exactly where you left it whenever you return.
To resume, open Terminal and run cd ~/vertical-playbooks && claude to pick up where you left off.
Claude Code automatically loads your CLAUDE.md and skills when you start a session in that folder.

Delete - I don't want to use this again

Remove all project files and generated plans from your machine.

Delete the entire project folder by running this command in Terminal:

rm -rf ~/vertical-playbooks

Confirm the folder is gone by running:

ls ~/vertical-playbooks

You should see a "No such file or directory" message confirming the folder has been removed.

What about Claude Code itself?

Claude Code remains installed on your machine. It uses no resources when idle and takes up minimal disk space. You can use it in any future project by running claude in any directory.

Nice Work!

You just turned Claude Code into a business document generator for your team. One command, one folder, and any team member can produce consistent, vertical-specific project plans without needing prompt engineering skills or your domain expertise in their head.

Here's what you accomplished:

Installed and configured Claude Code on your Mac and proved it works inside a dedicated project workspace, ready to use on any future project.
Experienced the context gap firsthand by comparing generic AI output against business-specific output, then created a CLAUDE.md file that encodes your company's agentic AI and data orchestration expertise, verticals, and document standards.
Built a reusable Skill that any team member can invoke with /generate-project-plan followed by a vertical name to produce a structured, vertical-specific project plan without writing a detailed prompt.
Secret Mission: Created a discovery call preparation skill with supporting template files, giving your team a second standardized workflow for prospect meetings.

Ready to quiz yourself?