Claude Code Documentation
Documentation from Anthropic for Claude Code, an AI coding assistant that can read, edit, and reason about code in your project.
Claude Code by Anthropic
Official DocsOfficial docs for Claude Code, a terminal-based AI coding assistant developed by Anthropic that can read, edit, and reason about the files in your current project.
What Is This Resource?
This resource contains documentation for Claude Code by Anthropic, a terminal-based AI coding assistant. It differs from the more commonly known Claude in that it does not involve any web UI and instead uses slash commands to perform common coding tasks in your terminal. Once installed, it can be invoked from inside the terminal of your existing projects, reading from and changing files within them.
The official documentation is hosted at code.claude.com and describes how to install Claude Code, basic commands to invoke it, and some tips for interacting with it in more sophisticated ways. If you are familiar with using terminals and are used to their workflow, the learning curve is rather flat since much of it resembles regular terminal commands.
Content & Scope
The documentation consists of several sections, including:
- Installation guide (package install, authentication, and basic configurations).
- Core CLI commands and flag parameters to use with Claude Code.
- Basic slash commands, such as
/edit,/create,/review,/explain, and/test. - Managing context windows: giving Claude Code access to your files.
- Optimal workflow when working with large code bases.
- Working in agentic mode, where Claude Code can plan out and run tasks autonomously.
- Handling conversations (memory) between sessions.
- Securing your interactions with autonomous Claude Code.
Relevance for CS Students
Unlike many other tools in this guide, Claude Code excels in handling existing code and modifying it in ways that make sense. Since you rarely need to write a program from scratch, having such capabilities is particularly useful for student programmers. Claude Code works by reading the project before it changes anything, ensuring that the generated code complies with project conventions.
Being able to work with your existing terminal means that Claude Code easily integrates into your workflows. Slash commands allow for quick invocation of frequently needed services (writing tests, generating docstrings, editing particular parts of the code), while the underlying Claude 3.5 Sonnet model is consistently ranked among the best in SWE-bench, a test of software engineering skills.
Tips for Effective Use
In order to become productive with Claude Code quickly:
- Work through the Quickstart section. In under 10 minutes, you will be able to install and start using Claude Code.
- Learn slash commands early on. You will mostly need
/edit(changes in the file),/review(reviews and comments), and/test(generates tests for your functions). - Read the "Working with Large Codebases" part before embarking on any significant project. It describes how to give Claude enough context without exhausting token limits.
- Try using
/explainon particularly challenging pieces of code. Claude Code understands code very well and can break down its meaning concisely and understandably. - Be careful when doing agentic tasks. Start with something small, like "add type hints to all functions in utils.py".