# Module 1 — The AI Practitioner's Toolkit

> The right tool at the right moment is the difference between a 10-minute task and a 3-hour one.

---

## The Four Tools

### 1. VS Code
**What it is:** The industry-standard code editor. Free, extensible, runs everywhere.

**When to use it:**
- Writing, editing, and reviewing code or structured text
- Working with local files without cloud dependency
- Reviewing diffs and version history

**Key AI features:**
- GitHub Copilot integration (inline suggestions, chat)
- Copilot Edits (multi-file changes from a single prompt)
- Terminal Copilot (explains errors, suggests commands)

**Install:** [code.visualstudio.com](https://code.visualstudio.com)

---

### 2. Cursor
**What it is:** An AI-first code editor built on VS Code. Thinks in terms of your whole codebase, not just the current file.

**When to use it:**
- Building or modifying complex projects where context across files matters
- Large refactors ("change how all API calls handle errors")
- When you want the AI to *propose changes*, not just suggest completions

**Key AI features:**
- Composer: describe a multi-file change, review before applying
- Codebase-aware chat (asks about your whole project, not just the open file)
- `.cursorrules` file for persistent project-level instructions

**Best for:** Builders and developers making substantive changes to a codebase.

**Install:** [cursor.com](https://cursor.com)

---

### 3. GitHub Copilot
**What it is:** AI coding assistant embedded across your workflow — VS Code, JetBrains, GitHub.com, and the CLI.

**When to use it:**
- Inline code completion as you type
- Quick explanations of unfamiliar code
- Writing unit tests and documentation automatically

**Key AI features:**
- Ghost text completions (Tab to accept)
- Copilot Chat (ask questions about your code)
- Copilot in pull requests (PR summaries, suggested fixes)

**Pricing:** $10/mo individual, free for students and open source.

**Install:** [github.com/features/copilot](https://github.com/features/copilot)

---

### 4. Claude Code (Anthropic)
**What it is:** A terminal-native AI coding agent. Runs from your command line and operates directly on your files.

**When to use it:**
- Autonomous multi-step tasks ("build this feature end-to-end")
- When you want the agent to *execute*, not just suggest
- Complex debugging across files and logs

**Key AI features:**
- Full file system access (reads, writes, runs commands)
- Long-context reasoning — handles large codebases
- Can run shell commands, install packages, run tests

**Best for:** Practitioners comfortable with the terminal who want an agent that acts.

**Install:** `npm install -g @anthropic-ai/claude-code`

---

## Choosing the Right Tool

| Situation | Use |
|-----------|-----|
| Quick edit + inline suggestions | VS Code + Copilot |
| Multi-file project changes | Cursor |
| PR review + GitHub workflow | GitHub Copilot |
| Autonomous end-to-end tasks | Claude Code |
| Writing / documents / analysis | Any model via browser or API |

---

## Setup Checklist

- [ ] Install VS Code
- [ ] Install Cursor
- [ ] Enable GitHub Copilot (VS Code extension)
- [ ] Install Claude Code (`npm install -g @anthropic-ai/claude-code`)
- [ ] Sign into each with your credentials
- [ ] Open the same project in both VS Code and Cursor — feel the difference

---

*Next: [Module 2 — Choosing the Right Model](models.md)*