A practical guide to using Claude properly. Covers the chat app, Claude Code, how to pick a model, skills, subagents, project setup and a bunch of other things that took me a while to figure out.
Written for people who know how to use a computer and GitHub but are new to Claude specifically.
- 01 Overview what Claude is and which tool to use for what
- 02 Choosing a model Opus vs Sonnet vs Haiku vs Fable, when to use which
- 03 The chat app claude.ai, Projects, memory, Artifacts, web search
- 04 Claude Code install, first session, how the agent loop works
- Plan mode and permissions
- CLAUDE.md give Claude standing instructions for your project
- Slash commands
- Skills
- Subagents delegate tasks to an isolated AI worker
- Hooks run shell commands on Claude Code events
- MCP connect external tools like GitHub, Slack, Drive
- 05 - VS Code vs terminal vs desktop
- 06 Skills deep dive build a skill from scratch
- 07 Project workflows CLAUDE.md + PLAN.md + PROGRESS.md, git hygiene
- 08 Tips and tricks context management, prompting, cost control
- 09 Permissions and safety the honest guide to "unrestricted mode"
- 10 Cookbook step by step recipes you can copy
- Glossary every key term with links to where it's explained
- Read the Overview first, takes 2 minutes.
- Go to claude.ai and try it. No install needed.
- When you want to use it for coding, come back for Claude Code.
License: CC BY 4.0