AI Slash Command Builder — Stop Re-Explaining Your Project to Claude, OpenClaw, or Clawdbot

The AI Slash Command Builder

Stop re-explaining your project every time your AI resets. Build a slash command once. Use it forever — or delete it tomorrow. Either way, it's yours.

Free prompt — Works with Claude Code, OpenClaw, Clawdbot, Moltbot, and more

You're probably treating your AI slash commands like they're precious. Like there's a limit and you might run out. There isn't. There's no slash command budget. Build them freely.

Here's a real example: you're building out a new feature for a CRM backend. It's a multi-step build. You don't want to spin up a full BMAD workflow — it's not that complex. But you're worried about one thing: what happens when you take a break, or when your AI's context fills up and you have to reset it? Are you going to be able to come back and pick up where you left off? Are you going to have to re-explain the whole project — the files, the PRD, the progress, the direction — from scratch? Every. Single. Time?

The answer is a project-specific slash command. Not a general skill. Not a massive system prompt. A single, focused command — something like /amd-006-build — that exists for exactly this feature, references all the right files, and gets your AI back on track in one keystroke. When the project's done, delete it. Or leave it. Doesn't matter. It cost you nothing to make.

What is an AI slash command?

A slash command is a saved instruction your AI loads on demand. You type /my-project and your AI instantly knows everything it needs to know about that project — the goal, the files, the current progress, how to pick up after a context reset. No re-explaining. No guessing. Just context, instantly.

In Claude Code CLI, slash commands are stored as Markdown files in your project's .claude/commands/ folder. In OpenClaw — the AI agent platform also known as Clawdbot or Moltbot — slash commands are registered commands your bot responds to on Telegram, WhatsApp, Discord, and more. Same idea, different platforms. And the Slash Command Builder prompt below handles the setup for you automatically — it asks which platform you're on and builds accordingly.

Slash commands vs. skills — why the difference matters

You could theoretically create a general skill and hope your AI detects it whenever your project comes up. But here's the problem with that: you're always guessing. Did it pick up the skill? Is it actually referencing the right files? Is it using the skill at all, or just winging it?

Slash commands are binary. You either typed it or you didn't. You invoke it on purpose. It fires. There's no ambiguity, no hoping your AI caught the right phrase. That certainty — knowing your AI has exactly the context you intended — is what makes slash commands so powerful for project work.

Skill

—Might trigger automatically
—You hope the AI detects it
—Broad use cases
—Great for recurring tasks

Slash Command

✓You invoke it on purpose
✓Binary — fires or it doesn't
✓Project-specific focus
✓Disposable — delete when done

Get the free Slash Command Builder prompt

Copy this prompt and paste it into Claude Code CLI or your Clawdbot in Telegram, WhatsApp, or Discord. Your AI will take it from there — no setup, no configuration, no coding.

Copy the prompt

Hit the button below to copy the entire Slash Command Builder prompt. It's a single block — grab the whole thing.

You are a Slash Command Builder. Your job is to interview me, understand what I need, and then actually build a working slash command for me on whatever platform I'm using.

Do not skip the interview. Do not assume what I want. Ask me questions in plain, simple English until you have enough information to build something great.

Here's how this works:

PHASE 1 — PLATFORM CHECK

First, ask me what platform I'm using. Am I working in:
- Claude Code CLI
- OpenClaw (also called Clawdbot or Moltbot) on Telegram
- OpenClaw (also called Clawdbot or Moltbot) on WhatsApp
- OpenClaw (also called Clawdbot or Moltbot) on Discord
- Something else

This matters because how slash commands are created and stored is different on each platform. You need to know this before anything else.

PHASE 2 — THE INTERVIEW

Once you know my platform, interview me to understand what this slash command should do. Ask me these things one or two at a time — don't dump them all at once. Use conversational, easy-to-understand language:

1. What is the project or task this slash command is for?
   (Get a clear description of what they're working on.)

2. What do you want to happen when you type this command?
   (What should the AI do, know, or reference when invoked?)

3. Are there any specific files, documents, or references this command should point to?
   (PRDs, progress trackers, handoff prompts, style guides, etc.)

   If they're not sure, make strong suggestions based on best practices:
   - "I'd strongly recommend including a reference to your main project document or PRD if you have one."
   - "If you have a progress tracking file, this is a great thing to attach so your AI always knows where you left off."
   - "A handoff prompt is really useful here — it tells the AI how to pick up where you left off after a context reset."

4. What do you want to call this slash command?
   (e.g., /my-crm-build, /project-atlas, /website-redesign — suggest that short, descriptive names with hyphens work best.)

5. Is this slash command permanent or disposable?
   (Help them understand that slash commands don't have to last forever. It's totally fine to build one for a single feature and delete it when you're done.)

STRONG SUGGESTIONS TO MAKE DURING THE INTERVIEW:

You should proactively recommend best practices as you go. Don't just passively collect answers — coach them. Here are things you should suggest if they haven't mentioned them:
- Always include a brief purpose statement so the AI knows WHY this command exists
- Reference any project documentation or PRD files if they exist
- Include a progress or status tracking reference so the AI knows where things stand
- Include a handoff prompt or context recovery instruction so the AI can get back up to speed after a reset
- Keep the command focused — one command per project or feature, not one giant command for everything

PHASE 3 — CONFIRMATION

Before building anything, summarize what you're about to create in plain language. Show them:
- The slash command name
- What it will do when invoked
- What files or references it will include
- Any special instructions baked into it

Ask them: "Does this look right? Want to change anything before I build it?"

PHASE 4 — BUILD IT

Once they confirm, actually create the slash command on their platform. Do not just output text for them to copy. Build the real, working command.

After building it, tell them:
- The slash command has been created
- How to use it (just type the command name)
- If they're on Claude Code CLI: let them know they may need to start a new session or reset their context with /reset for the new slash command to become available
- If they're on a Clawdbot platform: give them whatever platform-specific instructions apply for the command to become active
- Remind them: "This slash command is yours to keep, modify, or delete whenever you want. Don't be afraid to build more of these — you won't run out."

---

IMPORTANT RULES:
- Ask questions conversationally, one or two at a time. Do not overwhelm the user.
- Use simple, clear language. Your user might be a vibe coder, not a senior developer.
- Be opinionated and make strong recommendations, but let the user have the final say.
- Do not build the slash command until you've confirmed the plan with the user.
- The deliverable is a real, working slash command — not a block of text for them to implement themselves.

Paste it into Claude Code CLI, OpenClaw, or your Clawdbot

Drop the prompt into a fresh conversation. Works with Claude Code CLI, OpenClaw (Clawdbot / Moltbot) on Telegram, WhatsApp, Discord, or any AI agent platform. The builder will ask which platform you're using first so it creates the command in the right format.

Answer the interview questions

Your AI will ask you a short series of questions — one or two at a time — about what you're building, what files to reference, and what you want the command to do. Answer in plain English. No technical knowledge required.

Confirm — and your slash command is live

Your AI summarizes what it's about to create and asks for your confirmation. You say yes, and it builds the real, working slash command directly on your platform. No copying, no manual file creation — just a command that's ready to use. On Claude Code CLI, you may need to run /reset or start a new session before the new command is available.

How the 4-phase AI slash command builder works

The builder runs you through four phases. Here's exactly what happens at each one — and why it's designed this way.

Phase 1 — Platform Check

Before anything else, your AI asks which platform you're on — Claude Code CLI, Clawdbot on Telegram, WhatsApp, Discord, or something else. How slash commands are stored and activated varies by platform, and the builder needs to know this upfront to create the right file structure and give you accurate activation instructions.

Phase 2 — The Interview

One or two questions at a time, your AI builds a picture of your project: what you're working on, what should happen when you invoke the command, what files to reference, what to call it, and whether it's a permanent command or a disposable one for this build only. The builder coaches you along the way — suggesting best practices like including a progress file or handoff instruction if you haven't thought of them.

Phase 3 — Confirmation

Before building anything, your AI summarizes exactly what it's about to create — the command name, what it does, what files it references, and any special instructions. You see it all before it's built. Change anything you want at this point. Once you say yes, it builds.

Phase 4 — Build It

Your AI creates the real, working slash command on your platform. Not a template. Not a markdown block to copy somewhere else. The actual command, ready to invoke. It tells you exactly how to use it and reminds you that you can modify or delete it any time — and that you should absolutely build more of these.

What to include in your Claude slash command

The builder coaches you through this, but here's the breakdown of what makes an effective AI slash command for a project build. These aren't requirements — they're strong recommendations based on what actually works.

A purpose statement. One or two sentences explaining why this command exists. What project is it for? What phase of the build? This tells your AI the "why" before it reads anything else, and it changes how the AI approaches the context.

Your project documentation or PRD. If you have a Product Requirements Document or any specs for this build, reference it. File path is fine. Your AI will know where to look every time you invoke the command — no more hunting for files.

A progress or status tracking file. This is the one most people skip. A progress file tells your AI exactly where things stand — what's done, what's in progress, what's next. When your context resets, your AI reads the progress file and picks up where you left off, not where it guesses you left off.

A handoff or context recovery instruction. Tell your AI how to re-orient itself after a context reset. Something like: "Read the progress file first. Then read the PRD. Then confirm your understanding before acting." That simple sequence prevents the most common post-reset failure — the AI diving in confidently in the wrong direction.

A focused scope. One command per project or feature. Don't try to build one giant slash command that covers everything. Focused commands are cleaner, easier to maintain, and easier for your AI to use correctly. When you move to the next feature, build a new one.

Build AI slash commands often — they're disposable by design

The biggest mindset shift: stop thinking of slash commands as precious, permanent things. Build them freely. Build one for each feature. Build one for a single afternoon of work. Build one to help you get through one awkward multi-step build where you keep needing to re-orient your AI.

When the work is done — delete it, or leave it sitting there. Doesn't matter. The value was in the speed and accuracy it gave you while you needed it. You didn't waste time re-explaining things. You didn't wonder whether your AI had the right context. You didn't lose your thread.

That's the game. Use slash commands to stay locked in on your project, not to set up permanent infrastructure. The best Claude Code and OpenClaw slash commands are the ones that exist when you need them and cost you nothing when you don't.

Works with every major AI platform

The Slash Command Builder is smart about platforms. It asks which one you're using and creates the command in the right format. These are the platforms it supports:

Claude Code CLIOpenClaw (Telegram)OpenClaw (WhatsApp)OpenClaw (Discord)Clawdbot / MoltbotOther AI Agents

Each platform handles slash commands differently — Claude Code CLI uses .claude/commands/*.md files, while Clawdbot platforms use their own command registration systems. The builder handles those differences automatically. You just answer the questions.

Not sure what OpenClaw is? OpenClaw is a personal AI agent platform — also commonly referred to as Clawdbot or Moltbot — that you can run on Telegram, WhatsApp, Discord, and more, powered by Claude. It supports slash commands natively and is one of the best platforms for this kind of project-specific command workflow.

Frequently asked questions

What is an AI slash command?

An AI slash command is a saved instruction you type in your AI tool (like /my-project) that instantly loads context, documentation references, and instructions for a specific task or project. Instead of re-explaining your project every session, you type one command and your AI is back on track.

How is a slash command different from a skill?

A skill is a broad instruction set your AI might use when it detects a relevant situation. A slash command is binary — you either typed it or you didn't. There's no wondering whether your AI caught the right phrase. You invoke it on purpose, and it fires. That certainty is why slash commands are so powerful for project context management.

What platforms does this slash command builder work with?

The Slash Command Builder prompt works with Claude Code CLI, OpenClaw (also called Clawdbot or Moltbot) on Telegram, WhatsApp, Discord, and other AI agent platforms. The builder interviews you about which platform you're using and creates the slash command in the right format for your setup.

Do I need to know how to code to build a slash command?

No. The Slash Command Builder asks you questions in plain English. You describe what you're building, what files exist, and what you want the command to do — and your AI builds the entire thing for you. No coding required.

Are slash commands permanent?

They can be — but they don't have to be. You can create a slash command specifically for one feature build, use it throughout that project, and delete it when you're done. Slash commands are disposable tools, not permanent artifacts. Build them freely and delete them without guilt.

After building a slash command in Claude Code CLI, when can I use it?

In Claude Code CLI, you may need to start a new session or run /reset for the new slash command to become available. Your AI will remind you of this step after building your command.

What should I include in a slash command?

The most effective slash commands include: a clear purpose statement, references to your project documentation or PRD, a link to your progress tracking file, and a handoff instruction so your AI knows how to recover context after a reset. The Slash Command Builder will coach you through all of this during the interview.

Every day is magic. We are the prospectors of 1849 and there is gold in them hills. Be the one who prospects it.

Don't sleep on AI.

AI Handoff Prompt →|The Claw Loop →|Claw Fix →|← Back to Don't Sleep On AI