Let’s Master Claude Code Commands! Supercharge Your Dev Workflow

So, you’re using the AI coding assistant, Claude Code? Awesome! While it doesn’t have ‘commands’ in the old-school programming sense, there’s a whole set of handy ways to interact with it that can seriously boost your workflow. In this post, we’ll quickly cover the two main types: ‘in-session commands’ you use while chatting, and ‘CLI commands’ you punch into your terminal. We’ll even tell you which ones you’ll probably use most!

⚡️ The Essentials: Basic In-Chat Commands

These are the commands you’ll use right in the chat with Claude Code to control the flow of conversation and tell the AI what you need.

Command	Description	How it Works	Frequency of Use
/help	Forgot a command? Just type this! Shows a list of what you can do.	Your go-to lifeline. Quickly see all the commands you can use in the current chat.	High
/clear	Wipes the slate clean. Resets the current conversation.	Super handy when you’re switching topics. Prevents old chatter from confusing your new requests.	High
/init	Creates a special `CLAUDE.md` instruction file for your project.	Lets you teach Claude your project’s rules, like coding styles or important files to remember.	High (especially at the start of a new project)
/review	Get a code review from your AI buddy. Works for files or Git changes.	Just point to a file or type /review to get feedback on bugs and style.	High
/commit	Tells Claude to write a commit message for your changes and commit them.	Claude checks out your changes and whips up a smart commit message for you.	Medium
/test	“Run the tests!” – asks Claude to run your project’s tests.	Just tell Claude your test command in the `CLAUDE.md` file, and it’ll handle the rest.	Medium
@<file_path>	“Hey, look at this file!” Adds a file’s content to the chat.	A super easy way to show Claude what you’re talking about without copy-pasting code.	High

🚀 Power Up: CLI Commands for Your Terminal

These are for when you want to control Claude Code directly from your terminal. Think starting up, changing settings, and that kind of stuff.

Command/Flag	Description	How it Works	Frequency of Use
claude	Kicks off a chat session with Claude.	This is your starting line. The most basic way to say hello to Claude.	High
claude "your prompt"	Starts a chat with a question you already have in mind.	Saves you a step. Great for when you just want to ask something quickly.	Medium
claude -p "your prompt"	Executes a “one-shot” command and prints the output without entering an interactive session.	Highly useful for scripting and integrating Claude into other tools.	Medium to High
claude --continue or -c	Resumes the last interactive session.	Allows you to pick up where you left off.	Medium
claude update	Updates Claude Code to the latest version.	A straightforward command to get the newest features and bug fixes.	Low
claude mcp	Configures Model Context Protocol (MCP) servers.	An advanced feature for connecting Claude to other data sources or tools.	Low

🛠️ Custom Commands

One of the most powerful features of Claude Code is the ability to create your own custom commands. This is done by creating Markdown files in a .claude/commands/ directory within your project.

Example:

If you create a file named .claude/commands/refactor.md with the content “Please refactor the currently open file to improve readability and performance,” you can then simply type /refactor in your Claude session to execute this prompt.