Claude Code for Writers

# QUICKSTART: the first five things to set up

This gets you from "I installed Claude Code, now what?" to "Claude knows my
project, my voice, and my rules" in about half an hour. It assumes you have
never opened a terminal in your life. That is fine. You will type a few short
lines, press Enter, and read what comes back. That is the whole skill.

If you get stuck at any step, email bryan@becomeawritertoday.com with the step
number and what your screen says. Every message gets read.

> **Before you start:** this kit assumes Claude Code is already installed and
> you can open it. If not, follow Anthropic's own install guide first
> (https://code.claude.com/docs), then come back here. We do not cover
> installation, because Anthropic keeps their guide current and we cannot.

---

## What you are actually doing

Most people open Claude Code, type a request, get an answer, and close it. Next
time they open it, it has forgotten everything: their project, their voice,
their rules. They re-explain themselves every single session.

The five steps below fix that permanently. You are going to write Claude a
...

# CLAUDE.md: a long-writing project (book, report, or manuscript)

<!--
HOW TO USE THIS FILE
1. Copy this file into the top folder of your writing project.
2. Rename the copy to exactly CLAUDE.md (that exact name, in capitals).
3. Fill in every line marked FILL IN. Delete the guidance in brackets.
4. Delete this comment block when you are done, or leave it: text inside
   an HTML comment like this one is ignored and costs you nothing.

Claude Code reads a file named CLAUDE.md at the start of every session,
automatically, before you type anything. That is the whole trick. You are
writing a short briefing that Claude reads every single time, so you never
have to re-explain your project, your voice, or your rules again.

Keep this file short. Aim for under about 200 lines. A tight, specific
briefing is followed more reliably than a long, vague one.
-->

## What this project is

[FILL IN. One or two plain sentences. Example shape: "This is the working
folder for my book, a nonfiction guide to X for readers who are Y. Drafts
live in `chapters/`. Research and interview notes live in `research/`. The
finished manuscript is assembled in `manuscript.md`."]
...

# Example session: messy notes to a finished post

> This is a worked example so you can see the method in action before you try
> it. It is fabricated, but it is the real shape of a real session. The writer
> is a fictional freelancer, Jo Merrow, who runs a weekly newsletter called The
> Middle Draft for other freelance writers. Jo has already set up a
> `CLAUDE.md` for the newsletter (from the `newsletter.CLAUDE.md` template in
> this kit), so Claude starts the session already knowing the reader, the
> format, and Jo's voice.

Notes in italics like this explain what is happening and why. Everything else
is the session.

---

*Jo opens the terminal, changes into the newsletter folder, and runs `claude`.
Because a `CLAUDE.md` sits in that folder, Claude loads it automatically before
Jo types anything. Jo does not re-explain the newsletter. That work is already
done.*

**Jo:** Here are my messy notes from this week. I want to turn them into a
newsletter issue. Do not draft yet. First read them and tell me what the single
strongest idea is, and one structure that would earn the open.

```text
- lost a client monday. the big one. 40% of my income.
...