How Claude Code Works
A plain-English guide to Claude Code and Cowork: the .claude folder, CLAUDE.md, skills, slash commands, hooks, subagents, and MCP — from your first install to genuine mastery.
Claude Code & Cowork: From Zero to Mastery
A plain-language tutorial that takes you from “never used it” to “knows what’s happening under the hood.” Each module is its own file. Read straight through, or jump to what you need.
How it’s organized
Stage 1 — Getting started
- Module 0 — What is this thing, really?
- Module 1 — Installing and your first session
- Module 2 — The
.claudefolder: Claude’s home - Module 3 — CLAUDE.md: the project briefing
Stage 2 — The core mechanics
- Module 4 — Context: how Claude “remembers”
- Module 5 — Skills: teaching Claude a procedure
- Module 6 — Slash commands: reusable shortcuts
- Module 7 — Hooks: automatic actions
- Module 8 — Subagents: a team of specialists
Stage 3 — Power features
- Module 9 — MCP and connectors
- Module 10 — Putting the pieces together
- Module 11 — Cowork power features
- Module 12 — Mastery: habits and checklist
Stage 4 — Going deeper
- Module 13 — Best practices for token optimization
- Module 14 — Best practices for context engineering
A quick note on names
Claude Code is a developer tool you run in your terminal. Cowork is a friendlier desktop version of the same engine, built for people who aren’t programmers — it works with your files and apps instead of code. They share the same foundations (skills, hooks, agents, memory), so almost everything here applies to both. Where they differ, the module says so.