▸ TLDR
CHECK 01 10 MIN · VIDEO

What is Claude Code?

What you’ll know by the end of this check

  • The one thing that separates Claude Code from Claude.ai
  • Why “AI agent” isn’t marketing — it’s the mechanic
  • Four concrete things Claude Code can actually do to your codebase today

The shortest possible answer

Claude Code is Claude with a terminal, your files, and your shell. Instead of copying code into a chat and pasting the reply back, it opens the editor, runs the command, and checks the result itself.

It runs in your terminal, VS Code, JetBrains, the Claude Desktop app, and on the web. Same brain, different surface.

The mental model that unlocks it

Claude.ai is a chat window. Claude Code is an agent — software that runs in a loop: read, act, verify, repeat, until the goal is reached or it needs your input.

That loop is the thing. Once you’ve seen Claude Code plan a feature, edit six files, run the tests, read the failure, patch the regression, and stop — you’ll stop thinking about it as “AI chat” and start thinking about it as a coworker who never takes a break.

What it can actually do

ActionExample
Read your codebase”Trace how a user session flows from login to checkout.”
Edit files across the project”Rename getCwd to getCurrentWorkingDirectory everywhere and update every caller.”
Run terminal commands”Run the test suite. If anything fails, fix it and rerun.”
Search the web”Look up the current wrangler CLI flag for --compatibility-date and use the right one.”

Three things to remember

  1. Context is finite. Claude Code has a working memory, not an infinite one. It finds ways to get answers without loading your entire repo. (See: context management, check 07.)
  2. It asks before acting. Default behavior is ask-permission-first on every file write and every shell command. You pick how much leash.
  3. It can still be wrong. Wrong intent, subtle bug, over-engineered solution. Staying in the loop early catches it cheap.

Things to try right now (5 minutes)

  1. Watch the 3-minute lesson video at the canonical source below.
  2. Open the Claude Code product page at claude.com/product/claude-code.
  3. Pick ONE repo you touch weekly. Write a single sentence: “What would I ask Claude Code to do first in this repo?” Keep it for check 03.

The canonical version

Full official lesson is at anthropic.skilljar.com/claude-code-101/469788.

Ready to verify this check?

Finish the official lesson, then come back and mark this check cleared.