Skip to main content

3. VS Code

Note: The video covers material not in the guide below — please watch in full.

Action Step

Complete this before moving on.

After this training, take some time to poke around VS Code. Click around, get a feel for where things are — the explorer, the editor, the Claude Code chat panel. Try opening a file, check out the Outline view, and learn a couple of the hotkeys so you're more familiar with the environment before the next training.

Comment in Slack

Post your answer in your onboarding channel.

What was your biggest takeaway(s) from this training?

Training Guide

You just installed a bunch of things in the setup training. VS Code, Claude Code, WisprFlow, GitHub Desktop. You followed the steps, got everything running — but we didn't really explain what you're looking at.

Let's fix that. Before we go any further, you need to understand the environment you're working in.

(Let's start with what VS Code actually is)

What VS Code Is​

VS Code is an IDE — an Integrated Development Environment. Historically, it was built for developers to manage their codebases. It's where programmers write, test, and organize their code.

We've repurposed it.

The same features that make VS Code great for coding — the file explorer, the editor, the terminal, the extensions — make it great for knowledge work when you add Claude Code to the mix. Think of it as the Chrome of IDEs: free, runs on everything, and has thousands of extensions that add new capabilities.

You don't need to know how to code to use VS Code. You just need to know where things are.

(Let's walk through what you're looking at)

The Layout​

Here's what's on your screen:

  • The Explorer panel (left sidebar) — your folders, repos, and files. Basically your Google Drive, but local to your machine. Everything you cloned in the setup training lives here. Click a folder to expand it, click a file to open it.
  • The editor (center) — when you click a file, it opens here. This is where you'll read documents, review what the AI creates, and see changes. You can have multiple files open in tabs, just like browser tabs.
  • Changing Theme

That's the whole layout. Sidebar for navigation, center for viewing files, bottom for talking to the AI.

(A few shortcuts that'll make your life easier)

Key Shortcuts​

You don't need to memorize a bunch of keyboard shortcuts. But these three will save you time every single day:

  • Command + B — toggle the sidebar on and off. When you need more screen space, hide it. When you need to navigate files, bring it back.
  • Command + N — open a new tab. Useful when you want to look at a file without closing what you're currently reading.
  • Command + Shift + P — the Command Palette. This is VS Code's universal search — type any action and it'll find it. You won't need this often, but when you do, it's the fastest way to find anything.

The Outline view: When you're reading a long document, look for the Outline section at the bottom of the Explorer panel. It shows all the headers and lets you jump to any section instantly. Great for navigating long files without scrolling.

(One more thing — the extensions you installed)

Extensions​

In the setup training, you installed the Claude Code extension and MarkSharp. But what ARE extensions?

Extensions are add-ons that give VS Code new abilities. Think of them like apps on your phone — your phone does the basics out of the box, but apps make it do specific things. Same idea.

  • Claude Code extension — this is what lets you talk to Claude inside VS Code. Without it, VS Code is just a file editor. With it, you have an AI agent sitting in your workspace.
  • MarkSharp — this makes Markdown files look clean and readable. Without it, you'd see raw formatting characters everywhere. With it, documents look like actual documents.

You can browse and install more extensions later from the Extensions panel (the puzzle piece icon on the sidebar). For now, what you have is all you need.

Claude in VS Code​

Now let's talk about the thing you actually came here for — the AI.

When you open VS Code with the Claude Code extension installed, you'll see a chat interface. It might be in the bottom panel, or you might have it as a sidebar. Either way, it looks like a chat window. You type, the AI responds. Familiar territory.

But don't let the familiar interface fool you. This is not ChatGPT in a different wrapper. Let's walk through everything you're looking at.

The Chat Window​

This is where you talk to the AI. You type in natural language — plain English, not code — and the AI responds. You can paste text, file paths, even images and screenshots directly into this chat. The AI reads all of it as context.

The input box at the bottom is where your prompts go. Shift + Enter adds a new line without sending — so you can write multi-line prompts, paste file paths on separate lines, and build up your message before you hit Enter.

History and Sessions​

Here's something that trips people up: every time you start a new conversation with Claude, that's a session. Each session is its own thread — its own context, its own memory of what you've discussed.

You'll see a History panel where you can view all your past sessions. Each one is a separate conversation you've had. You can click back into old sessions to see what you discussed, what files were created, what prompts you used.

Why this matters: The AI only knows what's in the current session. If you had a great conversation yesterday where you set up a project, and you start a new session today, the AI doesn't automatically remember yesterday. Each session starts fresh. You'll learn later why this is important and how to work with it — for now, just know that sessions are separate.

You can also have multiple sessions open as tabs. Each tab is its own independent conversation. This is useful when you're working on different things and don't want to mix contexts.

Permission Modes​

This is where it gets different from any consumer AI app you've used.

Claude Code can do real things on your computer — create files, edit files, run commands. Because of that, it has a permission system that puts you in control of what it's allowed to do.

You'll see permission mode options that let you choose how much autonomy the AI gets:

  • Ask every time — the AI asks for your approval before every action. "Can I create this file?" "Can I edit this file?" You approve or deny each one. This is the safest mode and where you should start.
  • Auto-accept edits — the AI can read and write files without asking, but still asks before running terminal commands. Good once you're comfortable and trust the workflow.
  • YOLO mode — the AI does everything without asking. No guardrails. Only use this when you fully understand what you're doing and trust the task.

When the AI wants to do something that requires permission, you'll see a prompt asking you to approve or deny. Read what it's asking before you click approve. This is your control layer. The AI works for you, not the other way around.

Start with "ask every time." You'll naturally move to auto-accept as you get comfortable. There's no rush.

What's Different From Here​

You now know where everything lives. The chat window, the history, the permission system, the tabs. It looks like a chat app — and in some ways it is. You type, the AI responds.

But what the AI can actually do from this interface — reading your files, writing new ones, editing existing documents, running commands across your workspace — that's what makes it an agent platform and not just a chatbot.

The upcoming trainings will walk you through each of these capabilities one by one. For now, just get comfortable with the environment. Click around. Open a session. Look at the history. Check the permission settings. Make it feel like home.

(Now you know your workspace. Next up — the tool that makes talking to the AI even faster)