Claude Code slash commands. Custom commands are now skills
What slash commands do in Claude Code, which built-in ones are useful every day, how they differ from skills, and how to make your own command in a single file.
Part of the guideClaude Code, the complete guide. From installation to your own skills and plugins
In this article
In short
Slash commands are what you type inside a Claude Code session after a slash. We put the ones you need day to day in tables, explain the difference between a built-in command and a skill, and show how to make your own command in a single file.
Someone keeps a .claude/commands folder in their project with a few custom commands, then reads that Anthropic has merged them into skills. Nothing needs rewriting. A file at .claude/commands/deploy.md and a skill at .claude/skills/deploy/SKILL.md both create /deploy and behave the same way. That change says a lot about what Claude Code slash commands are today. Everything below was checked against the commands reference and the skills docs as of 10 November 2026.
Commands you type in a normal terminal, such as claude -c, are covered in Claude Code CLI commands. The whole series lives in the Claude Code guide.
How slash commands work in Claude Code
Type a bare slash in a session and you get a list of the commands available to you. More letters filter it. The list mixes several kinds of entries. Built-in commands with behaviour coded into the program, skills bundled with Claude Code, your own skills, and commands from plugins and MCP servers.
A command only works at the start of a message. Text after the name becomes its argument, for example /plan add a CSV export. If you send a command while Claude is still answering, it is queued and runs when the turn ends. Some, such as /status, /tasks and /usage, run immediately.
Not everyone sees the same list. Some commands depend on your platform, plan and environment. For example, /desktop only shows on macOS and x64 Windows when you are signed in with a Claude subscription.
Built-in commands for everyday work
The full list in the documentation has more than a hundred entries. These are the ones you use most.
| Command | What it does |
|---|---|
/init | Writes a first CLAUDE.md for the project |
/memory | Opens your CLAUDE.md files and memory settings |
/plan description | Enters plan mode, optionally with a task |
/model | Switches model and saves it as the default |
/effort | Sets how hard the model thinks |
/context | Shows what fills the context window |
/compact | Summarizes the conversation to free space |
/clear | Starts a new conversation with empty context |
/resume | Goes back to an earlier conversation |
/rewind | Rolls code, conversation or both back to a point |
/diff | Shows file changes, including Claude's edits |
/permissions | Manages allow, ask and deny rules |
/mcp | Manages MCP servers and their login |
/usage | Shows plan limits and stats |
/status | Shows version, model, account and connectivity |
/config | Opens settings |
/btw question | Asks a side question without adding it to the conversation |
/login, /logout | Switches account or signs out |
/help | Shows help and available commands |
/exit | Ends the session |
Several commands have more than one name. /clear also works as /reset and /new, /usage as /cost and /stats, /rewind as /checkpoint and /undo, and /permissions as /allowed-tools. A few older commands are gone. /vim was replaced by the editor mode setting in /config, and /pr-comments and /ultraplan no longer work.
One pair gets confused a lot. claude doctor in a normal terminal only reads and prints a diagnosis. /doctor inside a session is a skill that checks your install, settings and CLAUDE.md and applies fixes after you confirm.
Skills that ship with Claude Code
In the commands table, the documentation marks some entries as skills. These are ready-made instructions handed to Claude like a prompt, not features coded into the program. Examples are /code-review, which checks your current changes for bugs and with --fix applies the fixes, /simplify, which looks for cleanup, /debug, which turns on debug logging and reads the log with you, and /loop, which repeats a prompt on an interval. /update-config belongs here too. You describe a settings change, and Claude edits the right settings.json for you.
Before you ship, /security-review is worth running. It reviews the diff between your branch and the main one for security problems such as injection and broken authorization.
Custom commands are skills
You make your own command with one Markdown file. The older format is a file in .claude/commands/, the newer one a folder in .claude/skills/ with a SKILL.md inside. Both still work. Anthropic recommends skills for new work, because the folder can hold supporting files and Claude can pick the skill up on its own when it fits the task.
---
description: Review changes before a commit
---
Review the uncommitted changes in this repository.
List bugs, missing tests and places that could be simpler.
Focus on $ARGUMENTS.Once the file is saved, you type /review-changes signup form and the text after the name replaces $ARGUMENTS. A file in a subfolder, such as .claude/commands/frontend/component.md, becomes /frontend:component. The .claude folder in a project works for everyone in that repository. The ~/.claude folder in your home directory works in all your projects.
When your skill has the same name as a built-in command, your skill wins in the terminal, but it does not take over the command's aliases. A project skill called usage replaces /usage, while /cost still runs the built-in. A skill also wins over a file with the same name in .claude/commands/.
How to write good skills is covered in what a Claude skill is and how to write a skill.
To start, a few commands from the table and one of your own, for example a review before each commit, are enough. If you want to find out which repetitive tasks in your business could go to AI, our free AI audit is a good place to start.
What would you like to change in your business?
Leave your number. We will talk about your process and see whether we can help. The first call is free.
The service this text is about
Automation and AI / OCRAutomation of document re-typing, OCR and checking rules, and your data stays with you.