Claude Command: Secret Prompt Codes That Can Change How You Work With Claude
If you use Claude for writing, coding, brainstorming, research, or everyday tasks, you probably already know that the way you phrase a prompt can dramatically change the response you receive.
But there is another interesting approach: using short command-style instructions to tell Claude exactly how you want it to handle a task.
These commands can act like shortcuts. Instead of writing a long instruction every time, you can use a short keyword such as /brief, /debug, or /critique and then explain what you want.
In this guide, we'll explore a collection of useful Claude command-style prompts, what each one is designed to do, and examples of how you can use them in real conversations.
What Are Claude Commands?
Claude commands are short prompt prefixes that describe a particular response style, workflow, or task. They are best understood as reusable prompting conventions rather than assuming that every command is an official built-in Claude feature.
For example, instead of writing:
You could use a command-style instruction:
The command gives Claude a quick signal about the kind of output you want.
Claude Secret Prompt Codes
Make the Response Sound More Human
```Use /ghost when you want a response that feels natural, conversational, and less robotic.
Rewrite Something Much Sharper
```This command-style prompt is useful when your existing copy is technically correct but feels weak, boring, or unclear.
Ask for a More Powerful Approach
```/godmode can be used as a creative instruction for Claude to approach a problem aggressively, explore stronger alternatives, and avoid giving a shallow first-pass answer.
Keep in mind that this is a prompting convention, not a special permission or hidden capability.
```Steelman the Opposition
```One of the most useful ways to test an idea is to examine the strongest argument against it.
Use /devil when you want Claude to build the strongest reasonable case against your current position.
Create a 30-Second Pitch
```/pitch is useful for entrepreneurs, freelancers, marketers, salespeople, and anyone who needs to explain an idea quickly.
Compare Two or More Options
```Use this command when you need a structured comparison instead of a general explanation.
You can also ask Claude to present the result as a table with clearly separated advantages and disadvantages.
```Find Risks and Blind Spots
```/scout can be used to turn Claude into a second pair of eyes. Ask it to search through your plan and identify potential problems you may have overlooked.
Focus on Building a Usable Artifact
```When working with code, documents, prototypes, or other structured outputs, you can use /artifacts to tell Claude that the goal is to produce something practical rather than simply discuss the idea.
Whether a particular artifact workflow is available depends on the Claude environment and features you are using.
```Work Through a Complex Problem
```The OODA concept comes from Observe, Orient, Decide, and Act. As a prompt pattern, it can encourage a more structured approach to complicated problems.
Find Problems and Improve the Work
```Don't just ask an AI to tell you that something is good. Ask it to find what is wrong with it.
Explain a Difficult Topic Simply
```This command-style prompt is useful when technical explanations become difficult to follow.
You can also combine it with a specific audience, such as a beginner, student, customer, or non-technical manager.
```Give the Shortest Useful Answer
```When you don't want a long explanation, /brief tells Claude to prioritize concise answers.
Turn Claude Into a Learning Partner
```Use /teacher when you want to learn rather than simply receive an answer.
This approach works particularly well when combined with quizzes, examples, corrections, and progressively harder exercises.
```Ask for Brutally Honest Feedback
```/roast is a deliberately aggressive feedback style. It can be useful when you want weaknesses highlighted without excessive politeness.
For professional work, it is usually better to ask for specific criticism and actionable improvements rather than insults.
```Debug Code Step by Step
```Developers can use /debug to request a systematic examination of code, including errors, likely causes, and possible fixes.
For better results, include the error message, relevant code, expected behavior, and actual behavior.
```Turn an Idea Into Actionable Steps
```Some answers are easier to use when they are converted into a checklist.
This can turn a large project into smaller tasks that are easier to track.
```Set a Specific Expert Perspective
```With /persona, you can tell Claude which professional perspective should guide the response.
The quality of the result will depend heavily on how clearly you define the role, audience, objective, and constraints.
```Condense Information Quickly
```Use /summarize when you have a long piece of content and want the essential information without reading a lengthy explanation.
How to Combine Claude Commands
You don't always have to use a single command. You can combine several instructions to create a more specific workflow.
For example:
Another example for learning:
The important part isn't the slash command itself. The real value comes from clearly describing the desired outcome.
Why These Prompt Shortcuts Are Useful
- They save time: Reusable instructions can reduce repetitive prompting.
- They clarify your goal: A command establishes the desired response style.
- They encourage structured thinking: Commands such as /scout and /critique can make you look at a problem from another angle.
- They make prompts reusable: You can build your own collection of shortcuts for recurring tasks.
- They make experimentation easier: You can quickly change the way Claude approaches the same input.
Are These Official Claude Secret Commands?
Important: These slash commands should not automatically be treated as official hidden Claude commands. Many are simply prompt conventions that tell an AI how you want a task handled.
A command such as /debug does not magically unlock a special debugging model. Likewise, /godmode does not give Claude new permissions or capabilities.
Think of these codes as shorthand for instructions. Their effectiveness comes from the instruction that follows them and the context you provide.
How to Get Better Results From Claude
A shortcut alone is rarely enough. For consistently useful results, give Claude enough context to understand the task.
A simple prompt formula:
Command + Context + Goal + Constraints + Output Format
For example:
This is much more specific than simply writing: “/critique my website.”
Build Your Own Claude Command Library
Once you discover a prompt pattern that works well, save it. You can create your own personal command library for common activities.
- Writing: /rewrite, /shorten, /expand
- SEO: /seo, /keywords, /outline
- Coding: /debug, /review, /refactor
- Research: /compare, /scout, /summarize
- Learning: /teacher, /quiz, /explainlikeim5
- Business: /pitch, /strategy, /critique
You can then customize the commands based on your own workflow instead of relying on the same generic prompt every time.
Final Thoughts
Claude command-style prompts are a simple way to make your AI conversations more structured. Whether you're writing an article, debugging code, learning a new subject, comparing products, or reviewing a business idea, a short instruction can help establish the direction of the conversation.
The most important thing to remember is that these codes are prompting shortcuts, not magic commands. The quality of your context, instructions, examples, and desired output format will usually matter more than the keyword itself.
Start with a few commands such as /brief, /critique, /debug, /teacher, /compare, and /checklist, then build your own collection as you discover what works best for your workflow.
Discussion (0)
No comments yet. Be the first to share your thoughts!