What is this thing, really?
Module 0 — What is this thing, really?
Imagine hiring a very capable assistant who can read your files, write new ones, run programs, search the web, and use other apps on your behalf. You talk to them in plain English; they do the work and show you the result. That is Claude Code (in a terminal) or Cowork (on your desktop).
The key mental model: Claude is an agent, not a chatbot. A chatbot only talks. An agent acts — it opens files, runs commands, checks the results, fixes its own mistakes, and repeats until the job is done. Everything else in this tutorial is about giving that agent the right tools, memory, and guardrails.
Three ideas to hold onto from the start
-
It works in a folder. Claude operates inside a specific folder on your machine (your “project” or “workspace”). It can see and change files there. It can’t wander your whole computer unless you let it.
-
It keeps notes. It creates a hidden
.claudefolder to store settings, memory, and customizations. We’ll open that folder up in Module 2. -
It can be extended. Skills, hooks, agents, and connectors all let you teach it new tricks or new manners. The rest of the tutorial is mostly about these.