Your First Conversation with Claude Code

Now that Claude Code is installed, let’s learn how to interact with it effectively. Claude Code operates as a conversational agent in your terminal — you type natural language prompts and it responds with explanations, code edits, and command executions.

Starting a Session

Navigate to your project and start Claude Code:

cd ~/my-project
claude

You’ll see an interactive prompt where you can type your requests.

Basic Prompting

You can ask Claude Code anything about your codebase. Here are some good starting prompts:

# Understand the project structure
> Give me an overview of this project's architecture

# Ask about specific files
> What does the main function in app.py do?

# Get help with a task
> How would I add a new API endpoint for user profiles?

One-Shot Commands

You can also pass a single prompt directly from the command line without entering interactive mode:

claude "Explain the purpose of this project based on the README"

This is useful for quick questions or scripting.

Multi-Turn Conversations

Claude Code remembers context within a session. You can build on previous messages:

> Show me the User model

# Claude shows the User model code

> Add an email validation method to it

# Claude understands "it" refers to the User model

Slash Commands

Claude Code has built-in slash commands for common actions:

/help          - Show available commands and usage
/clear         - Clear conversation history
/compact       - Summarize and compress the conversation
/cost          - Show token usage and estimated cost
/doctor        - Check Claude Code installation health
/init          - Create a CLAUDE.md file for your project
/review        - Review a pull request
/terminal-setup - Set up terminal integration

Permission System

Claude Code asks for permission before making changes. You’ll see prompts like:

Claude wants to edit file: src/models/user.py
[Allow] [Deny] [Allow All]

You can choose to:

• Allow – Permit this specific action
• Deny – Reject this action
• Allow All – Auto-approve similar actions for this session

Tips for Effective Prompting

1. Be specific – “Add input validation to the login form” is better than “improve the form”
2. Give context – “We use Express.js with MongoDB” helps Claude make better decisions
3. Ask for explanations – “Explain why this approach is better” helps you learn
4. Iterate – If the first response isn’t perfect, refine your request

Exiting Claude Code

To end your session:

# Type exit, quit, or press Ctrl+C
> /quit

Summary

You now know how to start conversations, use slash commands, and interact effectively with Claude Code. Next, we’ll explore one of its most powerful features: generating code from scratch.

Previous: Introduction to Claude Code – Installation & Setup

Next: Code Generation with Claude Code