Your First Project

When something feels too big for your strategy folder, give it its own space. Its own CLAUDE.md. Its own rules.

What you're building today

You're going to take something you're actively working on, give it a dedicated folder with its own CLAUDE.md, and learn the pragmatism framework that keeps you shipping instead of overthinking.

Step 1: Spot the moment

Your strategy folder has brain/ for thinking and projects/ for doing. So far, everything has lived in one place. That works until it doesn't.

You know it's time to spin off a project when:

1. You keep creating files about the same topic
2. The work has its own vocabulary, timeline, or stakeholders
3. You want Claude to know different things in this context than in your general strategy folder

A client engagement, a product you're building, a content series, a business initiative. Anything with its own gravity.

Step 2: Create the project folder

mkdir -p ~/strategy/projects/[your-project-name]

Then ask Claude to set it up:

Create a CLAUDE.md in projects/[your-project-name]/ for this project.
Interview me about:
- What this project is
- Who it's for
- What stage it's in (early thinking, active building, shipping)
- Any constraints or preferences specific to this work

Use my words. Keep it tight.

This is the same interview pattern from Lesson 01, but scoped to one project. When you cd into this folder and run Claude, it reads this CLAUDE.md instead of (or in addition to) your main one. Different context, different conversation.

Step 3: The folder structure

Keep it simple. You can always add more later.

Don't over-engineer it. Three folders and a CLAUDE.md is enough to start. The structure will grow as the project grows.

Step 4: The pragmatism framework

This is the decision-making tool that keeps you moving. Two questions. That's it.

Most people spend 60% of their working time in analysis paralysis. "What if we also..." "Should we handle..." "What about..." Every "just one more thing" costs 20-40% of your remaining time.

Step 5: Build a deferral list

The permission slip. When you defer something, it doesn't disappear. It goes on the list. You can come back to it after shipping.

Create a file called projects/[project-name]/DEFERRED.md with this format:

# Deferred (Not Building Yet)

- [ ] [Feature/idea] (reason: [why it can wait])

## Why we deferred
Each item above has a reason. We're not ignoring these.
We're shipping without them first, then deciding with real feedback.

This kills the guilt. You're not ignoring ideas. You're sequencing them. Ship first, then decide what actually matters based on real usage instead of speculation.

Step 6: Reference your brain

Your project doesn't exist in isolation. Your brain has principles, patterns, and playbooks that apply here.

Before we start building, check brain/10-knowledge/ for:
- Principles that apply to this type of project
- Patterns that match what we're trying to do
- Playbooks that could save us time

Also check if I've solved something similar before in other projects.

This is the compounding effect in action. Every project you do feeds the brain. Every new project draws from it. The tenth project is faster than the first because you've encoded nine projects worth of decisions.

When to ship

The phrase that changes everything: