Claude Code series: 15 hands-on guides for a real project

Claude Code is Anthropic's coding agent. It works in your terminal, your editor and your repository: it reads the code, runs commands, edits files and proposes changes you review. This series collects 15 short guides to set it up properly on a real project, from the first session to continuous integration.

It is written for developers who already use Claude Code, or are about to, on client work such as a Shopify store or a Next.js app. Every article solves one thing with complete, tested files: a CLAUDE.md, a skill, a hook, a subagent, a GitHub Actions workflow. Commands and setting names are checked against the official Claude Code documentation and the anthropics/skills repository on the day each article is published, and every article links to the page it relied on.

Each guide comes with a step-by-step PDF and a ZIP of files ready to drop into your repository, with bash and PowerShell versions where the operating system changes something.

How to read it: if you are starting out, follow Setup in order (CLAUDE.md, a first skill, and the difference between skills, subagents, hooks and plugins). Hooks then turn rules that today depend on Claude remembering them into rules that always hold, such as keeping .env out of reach and running the linter after every edit. Agents and workflow covers code review with a subagent, plan mode and parallel work with git worktree. Official skills walks through the skills Anthropic publishes for design, Playwright testing, MCP servers and writing skills. Integrations takes Claude Code into your CI and your own APIs.

What you will not find here: productivity promises or numbers without a source. When an article uses an example figure, it says so. The articles ship in batches; the ones without a link are still being written.

If you would rather start with the overall method, read my 4-step workflow with Claude Code.