LEARN TO ASK
Prompt Mastery
Twelve weeks. Three tracks. One prompt-driven habit.
Turn your AI from a chatbot into a coworker.
Guided by your personal AI coach.
Meet Finney, your personal AI coach. Finney connects to Claude — Anthropic's AI, available as a desktop app, in your browser, or as a developer tool. He knows your progress, your struggles, and your last mission. When you're stuck, he asks the question that gets you moving — not the answer.
Three journeys, each one tailored to your use case, your needs, and your work. Each one makes you more productive.
Most AI coaching platforms have a portal. You go there to learn, then you go back to work. The learning and the working stay separate — and the skill never fully transfers.
Finney connects to your AI tool via MCP. Don't know what MCP is? You don't need to — Finney sets it up and guides you (and if you're curious, just ask him). That connection is the whole point: it lets him meet you exactly at the edge of what you already know and what you need next. Don't have Claude yet? You'll need a Claude Pro or Team account — it's the AI tool this platform is built on.
His job is not to give you the answer. It's to ask the question that makes you find it. That distinction is the whole method.

Works in Claude Desktop, Claude Code & claude.ai — wherever you already do the work.
Tracks your progress, your struggles, and your rank across every session. He never asks you to repeat yourself.
Coaches you to the output — never does it for you. The question that unblocks you beats the answer that doesn't stick.
The journeysLEARN TO ASK
Twelve weeks. Three tracks. One prompt-driven habit.
Turn your AI from a chatbot into a coworker.
LEARN TO OPERATE
Eleven personas. One AI that finally fits your role.
Plug a persona into your AI and ship work that feels yours.
LEARN TO SHIP
Twenty-six weeks from first idea to live product.
Launch your AI app — not a side project, a real one.

Understand the scope. Nothing more than that.

Take the smallest real step — not the perfect one.

Do it. Submit. Publish. Ship.

The moment you see what you made.
The loop runs about 15 times a week. Run it a hundred times and it stops being a method — it's just how you work now.

Finney doesn't let you stay in the WHAT phase. He moves you to HOW before you're ready, and to GO before you're comfortable. That's the structural reason you move faster — not motivation, mechanism.

Every checkpoint requires a real output before Finney marks it done. Not “I understood it.” A prompt that ran. A workflow that triggered. A feature that deployed. You can't fake a checkpoint. That's why the output is real.

After a hundred WOW moments — the end of a hundred checkpoints where you saw what you made — you're no longer someone who “uses AI.” You're someone who ships. Finney marks every WOW. You don't notice the shift. Then one day you do.
A certificate proves you showed up. Your Proof of Work proves what you can do — a set of real, usable deliverables, compiled automatically from the work you made along the way.

Your full project, written up — strategy, decisions, results. Word & PDF.

A ready-to-present deck of what you built. PowerPoint & PDF.

Every prompt and template you made, packaged to reuse. Word & PDF.

A public page of your work and its results — a link you can send.

A computed rating per skill, your rank, and an Open Badge anyone can verify.

An installable AI coworker trained on your own work — it runs inside your AI tool and writes in your voice.
Built automatically from your real work the moment you finish — arriving for the first graduating cohort.
Skilled Rabbit was built by a non-developer in 12 weeks — with the exact tools and method it teaches: Claude, MCP, and the Rabbit Loop. The platform is its own proof of concept.
You don't collect certificates here — you collect work. Every mission ends in a real artifact you keep: a prompt that ran, a workflow that triggered, a feature that shipped.
Finney lives inside Claude via MCP. He knows your progress and coaches you while you work — not in a portal you forget to open.

Skilled Rabbit was built by Sjors van de Wiel — not a developer, but someone who figured out how to build anyway. He built the platform using the same tools it teaches: Claude Code, MCP, the Rabbit Loop. It took twelve weeks. The platform is its own proof of concept.
The training exists because the gap between “AI can help you build” and “you can build” is real — and bridgeable. Finney is the coach he didn't have.