This Claude Code slash commands cheat sheet is a fast, scannable reference for the commands you type at the start of a prompt to steer a coding session. Claude Code (Anthropic's agentic coding tool for the terminal and IDE) ships with a generous set of built-in slash commands, and you can add your own. The goal here is not to make you memorize a fixed list (the set genuinely changes between releases), but to group the commands by purpose so you can find the right one quickly and build muscle memory for the handful you reach for every day. AI Drops, the maker of Claude Drops, is an independent project and is not affiliated with Anthropic.
How to use this cheat sheet
A slash command is an instruction that begins with a / and is only recognized at the very start of your message. Anything you type after the command name is passed to it as arguments. Some commands act on the session directly (clearing history, switching models), while others expand into a prompt the model then acts on. Below, the commands are grouped into the jobs you actually do: managing context, choosing a model, setting up a project, configuring the tool, and inspecting what is going on. Because exact names differ by version, the descriptions stay general. When you want the precise spelling and arguments, the official Claude Code documentation and your own /help output are the source of truth.
Context and session management
The single highest-leverage habit in Claude Code is controlling your context window. As a session grows, accumulated history can drag the model off-task and burn tokens. There is a command to clear the conversation and start fresh, and one to compact a long conversation in place (summarizing it so you keep the thread but free up room). When you switch from, say, debugging an API route to writing docs, clearing or compacting first keeps responses sharp. Other session commands cover things like ending or resuming work and visualizing how your context budget is being spent.
Models, project setup, and configuration
Claude Code lets you change which Claude model is handling the session on the fly. Reach for a lighter, faster model for quick edits and a more capable one for gnarly refactors or architecture work, and switch in a single command instead of committing to one model for the whole session. Some pickers also let you tune an effort level. For project setup, there is an initialize command that scaffolds a CLAUDE.md by scanning your repository so future sessions start with shared context, plus commands for managing permissions, reviewing or editing configuration, and signing in and out. Running the initialize command once on a new repo is among the highest-leverage things you can do.
For deeper background on the files and settings these commands touch, our companion guide on Claude Code tips and tricks is a good next read, and the official docs cover the exact configuration surface.
The cheat sheet: commands by purpose
Here is the quick reference. The left column names the job, the middle column describes what the relevant command does, and the right column tells you when to reach for it. Command names are kept general on purpose, run /help to confirm the exact spelling in your version.
| Purpose | What the command does | When to use it |
|---|---|---|
| Help / discovery | Lists every available command with a short description | Your first stop when you are unsure what exists or how a command works |
| Clear context | Drops accumulated conversation history and starts fresh | When you switch to an unrelated task and want focused, cheaper responses |
| Compact context | Summarizes a long conversation in place to free up room | When a thread is getting long but you still need its gist to continue |
| Switch model | Changes which Claude model (and sometimes effort level) is active | When task difficulty changes mid-session, light for edits, heavy for refactors |
| Initialize project | Scans the repo and scaffolds a CLAUDE.md of conventions and commands | Once per new repository, so future sessions start with shared context |
| Configuration | Opens or edits settings and managed permissions | When you want to adjust behavior, tools, or permission rules |
| Account / auth | Signs in or out and manages your account session | When setting up a new machine or switching accounts |
| Usage / cost | Shows token usage, cost, or how context is being spent | When you want visibility into spend or why context feels full |
| Custom commands | Runs a prompt you defined as a markdown file (for example /review) | When you have a repeatable workflow worth turning into one keystroke |
Custom slash commands you can add
Beyond the built-ins, the most useful entries on your personal cheat sheet are the ones you write. Custom slash commands in Claude Code are just markdown files: drop review.md into a commands directory and the file body becomes a reusable prompt you trigger with /review. Project commands live in .claude/commands/ inside a repo (commit them so your team shares the same workflows), and personal commands live under ~/.claude/commands/ so they follow you everywhere. Commands can take arguments, and Claude Code has been converging custom commands with skills, so the same /name can come from a command file or a skill directory. For the full walkthrough, see our guide to Claude Code slash commands.
- Name custom commands by the verb you reach for, like /test, /review, or /release, so they are easy to recall.
- Keep each command focused on one job. Small, composable prompts beat one giant catch-all.
- Commit project commands to version control so the whole team benefits and they stay in sync.
- Use the argument placeholder for the parts that change (an issue number, a path) and hardcode the parts that do not.
How to actually memorize them
A cheat sheet you have to look up every time is a crutch, not a skill. The fastest way to internalize the commands you use most is spaced repetition: review a small deck a little every day until recall is automatic. That is exactly what /cards for Claude Code is built for, a flashcards app that drills Claude Code slash commands (and the rest of the workflow) so the right command comes to mind without scanning a table. Think of it as the learning companion to Claude Drops: /cards helps you master what exists today, while Claude Drops keeps you current as new commands ship.
- Start with the five commands you would type most: help, clear, compact, switch model, and initialize.
- Run /help on your own install and note anything new since you last looked.
- Turn your two or three most-repeated prompts into custom commands and add them to the deck.
- Drill the deck in /cards for a couple of minutes a day until recall is automatic.
- Re-run /help after each Claude Code update to catch additions, then update your deck.
Stay current
Claude Code's command set keeps expanding, so a cheat sheet is only as good as how recently it was checked. For exact, current syntax, the official Claude Code docs and your own /help output are always the source of truth. To catch new commands and capabilities the moment they land, browse the Claude Code changelog or get the Claude Drops app on the App Store for a push notification on every Claude Code release. Pair that with /cards for Claude Code to memorize the commands, and you will both know what exists and remember how to use it. Bottom line: keep this cheat sheet as a map, lean on /help for the territory, and let the two apps handle freshness and recall.
Sources
Maintainer, Claude Drops
Ian builds Claude Drops and reads every Claude Code release so you don't have to. He writes plain-English guides to Claude Code's features, drawing directly from the official changelog and documentation.