chapter five

5 Getting started with Claude Code

 

This chapter covers

  • Agentic coding and your role as the reviewer
  • Installing and running Claude Code in your editor
  • Initializing the repository so Claude can navigate around
  • Setting permissions so Claude can work autonomously
  • Building and testing the first complete service

You’re used to driving the conversation with Claude. You take turns with the model in a chat session, and if you have nothing to add when Claude replies, the conversation ends. You’ve also seen that in Cowork, Claude’s turns can get much longer — your prompt might trigger multiple tool calls and generate output files before you see the response. Claude Code is where you’ll see just how much work Claude can do on its own. This is Anthropic’s agentic coding tool — it’s Claude running inside your repository, with access to all your code and all the tools on your machine. One prompt from you might start Claude on a turn that lasts hours and produces thousands of lines of code and tests.

5.1 What is agentic coding?

5.2 Installing and running Claude Code

5.3 Try it now — run Claude Code in your editor

5.4 Your first Claude Code session

5.5 Try it now — use Claude to build the agent documentation

5.6 Changing the permission mode to let Claude run free

5.7 Try it now — build out the application scaffolding in auto mode

5.8 Your role as reviewer

5.9 Writing code with the Superpowers skills

5.10 Try it now — build the MVP for your first service

5.11 Your turn: Complete the first service

5.12 Summary