• Get Into AI
  • Posts
  • 27 Claude Code Shortcuts That Change How You Code In One Line

27 Claude Code Shortcuts That Change How You Code In One Line

Most people never touch this. You will after today.

Author

Diwas Basnet
December 17, 2025

In partnership with

Become the go-to AI expert in 30 days

AI keeps coming up at work, but you still don't get it?

That's exactly why 1M+ professionals working at Google, Meta, and OpenAI read Superhuman AI daily.

Here's what you get:

  • Daily AI news that matters for your career - Filtered from 1000s of sources so you know what affects your industry.

  • Step-by-step tutorials you can use immediately - Real prompts and workflows that solve actual business problems.

  • New AI tools tested and reviewed - We try everything to deliver tools that drive real results.

  • All in just 3 minutes a day

Most people use Claude Code like a chatbot. Type a request. Watch it work. Repeat.

What they don't know is that Claude Code responds very differently based on specific keywords and commands. A single word can unlock hidden thinking modes, slash your token costs, or make it work 10x faster.

So here's a 27-shortcut cheatsheet you can use in any Claude Code session to control it on demand.

These aren't hacks. They're built-in features most developers never discover.

How to use these shortcuts

  • Type the shortcut directly in your prompt or command line.

  • Some are slash commands (start with /).

  • Some are magic words you add to any prompt.

  • Some are keyboard shortcuts or flags.

Example: 

ultrathink. Refactor this authentication system.

/compact Focus on preserving the database schema.

claude --dangerously-skip-permissions

Claude Code Shortcut Cheatsheet

Thinking Keywords

Shortcut

What It Does

Example

think

Allocates 4,000 thinking tokens

think. Debug this API error.

think hard

Allocates 10,000 thinking tokens

think hard. Design this database schema.

megathink

Same as think hard (10k tokens)

megathink. Plan this migration.

think harder

Allocates 31,999 thinking tokens (max)

think harder. Architect this system.

ultrathink

Same as think harder (max budget)

ultrathink. Solve this complex bug.

think deeply

Triggers extended reasoning (10k)

think deeply. Evaluate these options.

think very hard

Max thinking budget

think very hard. Review this codebase.

Slash Commands

Shortcut

What It Does

Example

/init

Generate CLAUDE.md for your project

/init

/compact

Compress context, preserve key info

/compact Focus on auth implementation.

/clear

Wipe conversation, fresh start

/clear

/cost

Show token usage and costs

/cost

/model

Switch between Opus and Sonnet

/model opus

/memory

Edit your CLAUDE.md file

/memory

/agents

Create and manage sub-agents

/agents

/review

Trigger code review mode

/review

/help

Show all available commands

/help

Keyboard Shortcuts

Shortcut

What It Does

Example

Shift+Tab

Toggle auto-accept mode

Press once for auto-accept

Shift+Tab (x2)

Enter Plan Mode (read-only)

Press twice for planning

Esc

Stop current operation

Interrupt if going wrong

Esc Esc

Rewind to edit previous prompt

Double-tap to retry

Tab

Autocomplete file paths

@src/ then Tab

Ctrl+C

Cancel current task

Hard stop

↑ (Up arrow)

Browse previous sessions

Navigate history

CLI Flags

Shortcut

What It Does

Example

--dangerously-skip-permissions

YOLO mode: no permission prompts

claude --dangerously-skip-permissions

-c

Continue last conversation

claude -c

-p "prompt"

Headless mode for automation

claude -p "fix lint errors"

--output-format stream-json

JSON output for CI/CD

Use with -p for pipelines

Quick Memory Tricks

Shortcut

What It Does

Example

#

Add quick memory to CLAUDE.md

# Always use MUI components

@filename

Reference specific file

@src/auth/login.ts

What to try today

Here are a few fast experiments you can run right now.

  1. Add ultrathink before your next complex prompt and watch Claude reason through it step by step.

  2. Run /init on any project to auto-generate a CLAUDE.md file.

  3. Create an alias for YOLO mode: alias cc="claude --dangerously-skip-permissions"

  4. Use /compact before you hit context limits instead of waiting for auto-compaction.

  5. Double-tap Shift+Tab to enter Plan Mode before any big refactor.

Once you memorize these, Claude Code feels like a completely different tool.

Catch you tomorrow.

If this helped you use Claude Code more effectively, share this with someone who's still typing basic prompts and waiting for permission dialogs.

Did you like today's issue?

Login or Subscribe to participate in polls.

Reply

or to participate.