Claude Code Cheat Sheet: 100+ Commands, Shortcuts and Hidden Features

Command: /powerup

What it does: Runs quick interactive lessons, with animated demos, on Claude Code features.

When to use it: Your first week. It’s the fastest way to learn what’s possible.

Example:

/powerup

Command: /init, then /memory

What it does: /init creates a CLAUDE.md file for the folder you’re working in, a set of standing instructions Claude reads at the start of every session. /memory lets you edit it later, and turn Claude’s automatic memory on or off.

When to use it: As soon as you start working in a new folder. Add your tone of voice, your rules, and anything you’re sick of repeating.

Example:

/init

Then add rules like: “Always write in Australian English. Never use em dashes. Save drafts to the /drafts folder.”

Command: /plan

What it does: Puts Claude in plan mode. It looks at everything and proposes a plan, then waits for your OK before changing anything.

When to use it: Before any big or risky job, like reorganising a folder or rewriting lots of files.

Example:

/plan turn the 12 blog posts in /blogs into a 4-week content calendar, without editing the posts

Also: Press Shift+Tab to cycle into plan mode.

Command: /btw

What it does: Asks a quick question about the session without adding it to the conversation, so it doesn’t distract Claude from the main task.

When to use it: Mid-task, when you want to check something without derailing the work.

Example:

/btw which files have you changed so far?

Command: /rewind, or press Esc twice on an empty prompt

What it does: Rolls the conversation, the files Claude edited, or both back to an earlier point.

When to use it: Claude misunderstood you and changed the wrong things. Rewind instead of arguing with it.

Example:

/rewind

Shortcut: Esc

What it does: Stops Claude straight away. It keeps the work done so far, so you can redirect it.

When to use it: You can see it heading the wrong way. Don’t wait for it to finish.

Commands: /context, then /compact

What it does: /context shows what’s filling Claude’s working memory for this chat, with suggestions. /compact shrinks the conversation into a summary so Claude has room to think again.

When to use it: Long sessions where Claude starts forgetting things or getting slow.

Example:

/compact keep the brand voice decisions and the final headline list

Starting a new task?: Use /clear for a completely fresh conversation.

Command: /goal

What it does: Sets a goal. Claude keeps working across turns until the goal is met.

When to use it: Bigger jobs where you’d otherwise keep typing “keep going”.

Example:

/goal every product page in /products has a 150-word description and a meta description under 155 characters

Command: /deep-research

What it does: Runs many web searches on one question, cross-checks sources and writes a cited report.

When to use it: Competitor research, market questions, anything where you need sources.

Example:

/deep-research what are the top 5 AI tools Australian marketing agencies are using in 2026, with pricing

Commands: /slides and /design

What it does: /slides makes a presentation from your brief. /design drafts mockups, landing pages or posters.

When to use it: You’ve done the thinking in Claude Code and want something to show people.

Example:

/slides a 6-slide summary of the content calendar for our team meeting

Command: /loop

What it does: Runs a prompt again and again while your session stays open. Give it an interval, or leave it out and Claude sets the pace.

When to use it: Watching for something to change.

Example:

/loop 30m check the /inbox folder for new briefs and summarise any you find

Want it to run when your computer is off?: /schedule sets up routines that run in the cloud.

Command: /voice

What it does: Turns on voice dictation. Needs a Claude.ai account.

When to use it: Long briefs. Talking is faster than typing.

Example:

/voice

Command: /fewer-permission-prompts

What it does: Looks at what Claude has asked permission for before, and adds a list of safe, read-only actions so it stops asking every time.

When to use it: You’re clicking “yes” to the same harmless request over and over.

Example:

/fewer-permission-prompts

Command: /usage

What it does: Shows session cost, plan usage limits and activity. /cost does the same.

When to use it: You’re hitting limits and want to know why.

Pair it with: /model and /effort. A lighter model or lower effort on simple jobs makes your plan go further.

Command: /insights

What it does: Builds a report on how you use Claude Code: your projects, where things go wrong, and features to try.

When to use it: After a few weeks of use. It’s like a review of your habits.

Example:

/insights