Claude Code, the complete guide. From installation to your own skills and plugins
All our articles on Claude Code in one place, in the order you need them. Installation, the first day, everyday commands, settings, editors, Cowork, skills and plugins, plus three paths to get started.
In short
Claude Code is an AI agent from Anthropic that works on the files in your project, runs commands and checks the result itself. This guide collects all our articles on it, from installation to your own skills and plugins, and suggests where to start whether you do not code, code a little or code for a living. Each link comes with one sentence on what you will find there.
Someone hears from a friend that Claude Code fixed his website on its own, and that evening types the name into a search engine. They get dozens of documentation pages, videos from half a year ago and posts where the commands no longer work. This Claude Code guide exists to sort that out. Each article in the series answers one question, and we checked all of them against the official documentation, as of 10 November 2026.
We work in Claude Code every day ourselves, often in several sessions at once. What that looks like is described in how we use AI every day. How much AI really speeds up developers and how we check code from AI is in AI in programming.
What Claude Code is
Claude Code is Anthropic's tool for working with code and files. Unlike a chat, it does not only answer, it carries out the task. It reads project files, changes them, runs commands, checks what happened and takes the next step until it is done or needs to ask you something. This way of working is called an agent, and the difference between a chat and an agent is explained in what is an AI agent.
A full description of the tool, including the permission modes, is in what is Claude Code. According to Anthropic's documentation, most ways of using Claude Code require a paid Claude plan or an Anthropic Console account.
Where Claude Code runs
Claude Code is one engine with several ways in. The same CLAUDE.md files and the same settings apply whether you open it in a terminal, an editor, the desktop app or a browser.
Where Claude Code runs
The terminal gives you the most options, and most tutorials start there. The VS Code extension shows changes next to your code, and the Claude desktop app has a Code tab that needs nothing extra installed. Google Antigravity is a separate product with its own agent, and you can run Claude Code inside its terminal.
Where to start
If you run a business and do not code, do not start with the terminal. First read what an LLM is, so you know what to expect from the answers. Then look at Claude Cowork setup, the mode for working on files in the regular app, and five office tasks. Finish with official Claude skills, which let Claude hand back finished Word and Excel files.
If you code a little, for example you maintain your own website, start with installing on your system and the first run. Then the VS Code extension, so you see changes before you accept them, and permissions, so you know what Claude may do without asking. The fourth step is a CLAUDE.md file with your project's rules.
If you code for a living, go through the essential commands and slash commands, then hooks, MCP and headless mode in CI. Finish with how to write a skill and plugins, so the team shares one set of tools.
Every article in the series
Installation and the first day
Installing on macOS shows the one Terminal command, Homebrew and the most common PATH error. Installing on Windows explains why PowerShell and CMD have separate commands and when to choose WSL. Installing on Linux covers the script, the apt, dnf and apk repositories and signing in on a server without a browser. The first run walks you through signing in, the folder trust prompt and a calm first day of work.
Everyday work
Essential commands collects the flags and keyboard shortcuts that really save time, in tables. Slash commands explains the difference between a built-in command and a skill and shows a custom command in one file. Subagents shows how to hand log reading to a helper so that only the summary comes back. The article on the Claude usage limit explains why a short question in a long chat costs so much. How to write prompts helps with agent tasks too.
Settings and extensions
Permissions covers allow, ask and deny rules in settings.json files and which file wins. CLAUDE.md explained shows what Claude reads at the start of every session, and CLAUDE.md examples gives four templates to copy. Hooks shows a command that runs every time, not only when Claude remembers. MCP in Claude Code is about connecting outside tools with one command and choosing who gets access. Headless mode and CI covers the -p flag, JSON output and Claude in GitHub Actions.
Editors
Claude Code in VS Code is the extension with a diff view next to your code and file mentions. Claude Code in Antigravity separates what certainly works from what Anthropic does not confirm. Terminal, VS Code or Antigravity helps you pick a way in depending on how much you code.
Instruction files and concepts
What is AGENTS.md describes the instruction file read by Codex, Cursor and Copilot. AGENTS.md and CLAUDE.md shows how to keep one source of rules for every tool. What is an LLM explains where a model's confident mistakes come from. Tokens and the context window explains what fills the model's working memory before you type anything. The article on connecting AI to your calendar shows how to link AI to your calendar and other tools.
Claude Cowork
Cowork setup says which plans and systems it runs on and how to start it. Cowork in the office gives five tasks with a ready prompt and a list of things to check. Cowork limits and safety explains why it uses up the limit quickly and where files are processed. Cowork in practice suggests giving it a copy of a folder for the first task.
Skills and plugins
What is a Claude skill explains the folder with a SKILL.md file and how Claude loads it in stages. Installing skills shows three routes, a ZIP in the app, a folder in Claude Code and a plugin. How to write a skill is a working example from scratch with a test and Anthropic's rules. Claude Code plugins covers packages of skills and hooks, catalogs and install commands. Official Claude skills reviews Anthropic's skills and plugins. Impeccable is a popular skill from outside Anthropic for interface design.
You do not need to read everything. Pick your path from the section on where to start, and come back here when a new question comes up. If you want to know where AI can take work off your team, try our free AI audit.
What would you like to change in your business?
Leave your number. We will talk about your process and see whether we can help. The first call is free.
The service this text is about
Automation and AI / OCRAutomation of document re-typing, OCR and checking rules, and your data stays with you.