Beginner

Paste this guide into Claude Code and it will walk you through every step interactively.

The Mental Model

What this gives you: A clear understanding of what Claude Code can actually do for your business. This isn't a chatbot - it reads your files, runs your commands, connects to your tools, and works in parallel. Once you get this, everything else clicks.

The biggest mistake people make with Claude Code is treating it like a chatbot. It is not. Understanding what it actually is - and what it can actually do - is what separates people who get 10x leverage from people who get marginal value.

Stop Thinking Chatbot. Start Thinking Junior Developer.

ChatGPT and Claude.ai are chatbots. You type, they respond with text. The conversation happens entirely in a message thread. They cannot touch your filesystem, run commands, make API calls, or deploy anything. They can only write text that you then go and execute yourself.

Claude Code is fundamentally different. It is a software agent that runs on your machine with access to your actual environment. The correct mental model is not "AI chatbot" - it's "a skilled junior developer sitting at your keyboard who can work autonomously on tasks you delegate."

That junior developer can:

The key shift: you are not using Claude Code to get answers. You are using it to get work done.

What Claude Code Can Actually Do

Here are concrete examples of tasks you can hand off entirely - not just get help with, but fully delegate:

File Operations

# Claude can read, create, edit, delete, and reorganise files "Refactor the entire auth module into separate files for login, register, and token refresh" "Find all hardcoded strings in the codebase and extract them to a constants file" "Rename all components from PascalCase files to kebab-case files and update all imports"

Shell Commands

# Claude runs commands and sees the output, then acts on it "Run the test suite, identify every failing test, and fix them one by one" "Run npm audit, then update or patch all vulnerabilities it finds" "Build the project, check for TypeScript errors, and fix every one of them"

Git Operations

# Full git workflow - stage, commit, push, branch, merge "Create a feature branch, implement the user dashboard feature, commit with a proper message, and push" "Look at the last 10 commits and write a changelog entry summarising what changed" "Revert the changes from the last commit without destroying the commit history"

Remote Server Access

# SSH into remote machines and run commands there "SSH into my VPS, check the pm2 process list, and restart any processes that have crashed" "Check the nginx error logs on the server and tell me what errors occurred in the last hour" "Deploy the latest version of the app to the VPS and verify it's running"

The Three Layers of Claude Code

Understanding the architecture of Claude Code helps you know where to configure things and why certain setups are more powerful than others. There are three layers:

Layer 1: CLAUDE.md - The Brain

This is the persistent memory and instruction set. It's what Claude reads at the start of every session to understand who you are, what tools are available, what mistakes to avoid, and how to behave. Without a CLAUDE.md, every session starts from zero. With one, every session starts with full context.

Think of this layer as the onboarding document. Everything about your working style, your environment, your rules, and your history lives here.

Layer 2: MCP Servers - The Hands

MCP (Model Context Protocol) servers are plugins that extend what Claude Code can interact with. Without MCP servers, Claude can only work with your local filesystem and shell. With them, it can reach out to external services as if they were part of your local environment.

# Without MCP: Claude writes the Trello API call code and asks you to run it # With MCP Trello: Claude calls Trello directly, reads the board, updates cards # Without MCP: Claude drafts a Google Doc and gives you the text to paste # With MCP Google Docs: Claude creates the document, formats it, sets permissions # Without MCP: Claude writes the Stripe API code for you to execute # With MCP Stripe: Claude pulls live revenue data, checks subscriptions, creates invoices

Common MCP servers worth setting up: Google Calendar, Notion, Trello, GitHub, Stripe, Slack. Each one multiplies what Claude can accomplish autonomously.

Layer 3: Skills - The Muscle Memory

Skills are reusable instruction files stored in ~/.claude/skills/. A skill is a detailed playbook for a complex, repeatable task. Instead of explaining the same multi-step workflow every time, you write it once as a skill and invoke it with a slash command.

# Instead of this every time: "Pull my Stripe data, check this month's revenue, compare to last month, check Notion for active projects, pull Calendar for this week's calls, then write a morning brief with all of that information formatted clearly..." # You write it once as a skill, then just do: /morning-brief

Skills handle the complexity so your prompts stay simple. They're the difference between knowing how to do something and having the muscle memory to do it instantly.

The Context Window and How It Works

The context window is the working memory of the model - the total amount of text (your conversation + all file contents Claude has read + all command outputs) it can hold at once. Claude Opus 4 has a 200,000 token context window. That sounds large, but it fills up faster than you expect when Claude is reading multiple large files and running several commands.

Practical implications:

PracticeWhy It Matters
Read specific files, not entire directoriesIf you ask Claude to "look at my codebase," it will try to read everything. Instead, direct it: "look at src/auth/login.ts."
Use /compact when sessions get longThe /compact command summarises the conversation history to free up context space without losing the thread.
Fresh sessions are sometimes better than long onesFor a new, unrelated task, starting a fresh claude session gives Claude a clean slate and maximum available context.
One-shot mode keeps context minimalclaude -p "task" uses only what's needed for that single task - no session overhead.
# Check context usage inside a session /status # Output shows something like: Model: claude-opus-4-5 Context used: 47,832 / 200,000 tokens Cost so far: $0.23
Tip: If a session is getting slow or Claude seems to be losing track of earlier context, run /compact to compress the history. Claude retains the key facts but frees up token space for the rest of the work.

Interactive vs One-Shot Mode: When to Use Which

This is one of the most practical decisions you'll make repeatedly. Here's a clear framework:

Use Interactive Mode (claude) when:

Use One-Shot Mode (claude -p) when:

# Interactive: open-ended exploration claude > I'm getting a weird bug where users can't log in after password reset, help me debug it # One-shot: defined, automatable task claude -p "Generate a summary of all TODO comments in this repo" > todo-report.md # One-shot in a script git diff HEAD~1 | claude -p "Review this git diff and suggest any improvements"

Claude Code vs ChatGPT / Regular Claude: The Real Difference

FeatureClaude.ai / ChatGPTClaude Code
Reads your filesNoYes
Runs shell commandsNoYes
Writes filesNoYes
Makes API callsNoYes
Accesses git historyNoYes
Deploys codeNoYes
SSHs into serversNoYes
Persistent memoryNo (per session)Yes (CLAUDE.md)
Works autonomouslyNoYes
Costs per tokenYesYes

The other AI tools are conversational. Claude Code is operational. If you find yourself copy-pasting code from Claude.ai into your editor, running it, then copy-pasting the error back - you're doing it the slow way. Claude Code eliminates that loop entirely.

How Claude Code Reads Your Codebase

When you start a Claude Code session in a project directory, Claude automatically reads several things before you even type your first message:

What Claude ReadsPurpose
CLAUDE.md (global, then project-level)The instruction set - your rules, preferences, and context
Directory structureA high-level map of what's in the project
package.json / pyproject.toml / go.modDependencies and scripts
README.md (if present)Project overview

It does not read every file immediately (that would burn through your context window). Instead, it reads files on demand as it needs them to complete your task. When you ask it to "fix the login bug," it reasons about which files are likely relevant and reads those specifically.

# You can explicitly tell Claude what to read "Before we start, read src/auth/, src/middleware/, and the .env.example file" # Or tell it to search "Find all files that import from the database module and read them"

Ready to set this up for your workflow?

Book a call and get a personalised Claude Code setup tailored to your tech stack and business.

Book a Call

Sub-Agents and Parallel Work

One of the most powerful (and least-used) features of Claude Code is the ability to spawn sub-agents - separate Claude instances that work on specific sub-tasks in parallel. This is particularly valuable for research, QA, and complex multi-part builds.

The core pattern is "fan out, fan in":

# Fan out: give each sub-agent a different angle "Spawn three sub-agents: 1. Research how Stripe handles subscription upgrades 2. Research how our current billing code works 3. Research common migration pitfalls Then synthesise their findings into a migration plan."

Practical use cases for sub-agents:

PatternHow It Works
Developer + QAOne agent builds the feature, a fresh agent with no context reviews and tests it. The fresh agent catches assumptions the builder missed.
Research parallelisationResearch 5 different APIs simultaneously instead of sequentially - each agent goes deep on its specific topic.
Stochastic consensusFor strategy decisions, spawn multiple agents with slightly different framings. Where 3+ agents agree, you have high confidence in the answer.
Tip: Sub-agents each consume their own tokens, so use them for genuinely parallel work where the speed and quality gain justifies the cost - not for tasks that naturally flow sequentially.

Token Costs and How to Think About Them

Claude Code uses the Claude API, which charges per token. Here's how to think about this practically:

# Approximate costs (Opus, as of mid-2025) Input: $15 per million tokens Output: $75 per million tokens # What that means in practice A 1-hour complex coding session: ~$2-8 A simple one-shot command: ~$0.02-0.10 Reading a 1,000-line file: ~$0.015 Running a sub-agent task: ~$0.50-2.00

The right way to think about cost is not "this costs money" but "what is this worth to me?" If a 2-hour task that would cost you $150 of your time can be done by Claude Code in 10 minutes for $3, that's a 50x return on the spend.

Strategies to keep costs reasonable without sacrificing quality:

StrategyHow It Reduces Cost
Use one-shot mode for scripted, repeatable tasksCheaper than interactive sessions because there's no conversation overhead.
Use /compact in long sessionsCompresses history and reduces input tokens on all subsequent messages in the session.
Be specific about what to read"Read src/auth.ts" is cheaper than "look at my codebase." Targeted reads keep context lean.
Use /status to monitor spendBuilds awareness of what different task types cost so you can make informed trade-offs.
Reserve sub-agents for tasks that genuinely benefitDon't spin up 5 agents when 1 will do. Parallelization has real cost - justify it.
# Monitor session costs in real time /status # Output: Model: claude-opus-4-5 Tokens used: 23,421 (input) / 4,832 (output) Session cost: $0.71 Context remaining: 171,747 tokens

Once you internalise these numbers, you stop thinking about Claude Code as an expense and start thinking about it as the highest-leverage employee you've ever hired - one that works continuously and gets better as your CLAUDE.md grows.

Next Guide
Building Your First Skill