Most of us open a shiny new AI model, poke around for five minutes, then go right back to typing the same prompts we typed yesterday. That was me too. So when I saw this walkthrough from an AI professional on LinkedIn, I stopped scrolling.
The original poster wrote a click-by-click plan for the freshly released Claude Fable 5.1. The whole thing takes 30 minutes, and the creator suggests you set a timer and do each step while you read. I was genuinely impressed by how practical it is, so let me break down what this expert laid out.
⏱️ Minute 0 to 5: Get set up
The author starts simple. Install the Claude desktop app, then open it. In the prompt box, choose “Cowork”. Head into Settings and turn everything off.
Why kill all the extras? Because this whole method only leans on two things: projects and skills. The creator wants a clean workspace with no distractions before you build anything.
🎯 Minute 5 to 7: Match the model to the task
Next, this industry pro says to pick Fable 5.1 for your first two prompts, the ones where you’re still figuring out the actual problem. Set Effort = High. That’s the default, so you probably don’t have to touch a thing.
The point here: use the strongest reasoning while you’re defining the problem, not after you’ve already committed to the wrong path.
🛠️ Minute 7 to 15: Build your first skill
This is the part I loved. Think about the prompts you paste every single morning. The expert says to turn those into a reusable Skill instead of retyping them forever.
In Cowork, paste this:
“Use skill-creator to build a skill for [your task].”
Claude then writes a SKILL.md file. Here’s the workflow the creator recommends:
- Read the output it generates
- Test the skill to make sure it works
- Install it
- Go to Claude Customize, then Skills, and toggle it on
▶️ Minute 15 to 18: Actually use the skill
Now open a new chat. The post’s author says to type a forward slash, your skill name, and your context. Here’s the exact example the original poster gave:
“/client-email Project is two days late because the API keys arrived Tuesday. Next milestone Friday.”
If the output feels off, you don’t rebuild anything. You just tell it:
“Update the skill: [fix].”
And it edits itself. That self-correcting loop is what makes this stick.
📦 Minute 18 to 20: Build a second skill in one line
This trick from the creator is almost too easy. Do any task with Claude your normal way. When you like the result, type:
“Package what you just did as a skill.”
No setup, no template. You work first, then capture the win as something you can reuse forever.
🔌 Minute 20 to 23: Connect one tool
The expert keeps it to just one connector so you don’t get overwhelmed. Go to Claude Settings, then Connectors, then Browse. Add Google Drive, Slack, Granola, or Gamma, and sign in.
Back in Cowork, test it with this:
“Read my last 5 files in Drive and tell me which one I haven’t finished.”
💾 Minute 23 to 27: Save your starting prompt
The final build step is a reusable opener. This savvy professional calls it your “non-negotiable” first prompt for every Fable 5.1 chat. Here it is exactly as shared:
“Use ASD-STE100. I want to [goal, with success criteria]. Ask me clarifying questions first (use AskUserQuestion) until you have enough context to answer me.”
The clever bit is forcing the model to ask questions before it answers, so it gathers real context instead of guessing.
✦ The prompting rules for Fable 5.1
The creator closed with a set of quirks and quick fixes, and these are the ones I’ll be stealing:
- It gives results, not steps. If you want the steps, add: “Say in one line what you are about to do before you do it.”
- It stalls on long tasks. Add: “I’m not watching. Finish autonomously.”
- It copies text when summarising. Add: “Put copied text in quotation marks.”
- Long chats get bloated. Type: “Summarise this chat so I can continue in a new window,” then paste the summary into a fresh chat.
Why this matters
What makes this approach smart is the shift it teaches. Instead of treating Claude like a search box you re-explain yourself to every day, the author shows you how to bank your repeat work as skills. You define the problem once, capture the solution once, and reuse it forever.
That’s the difference between using an AI tool and actually building a system around it. And the whole thing fits in a single coffee break.
This is just my breakdown of the highlights. The original LinkedIn post walks through every click in the creator’s own words, so check out the full post for the complete play-by-play.