Stop trying to remember the book. Pull its framework out into a written file that teaches your AI tool one job: apply this framework to my decisions. Feed it your own business specifics, define the moment it should fire, and name the parts of the book that do not apply here. Then point it at a real decision this week.
There is a book on your shelf you actually finished. You underlined it, told two people about it, and agreed with all of it. Then the week started again. Six months later the business runs the way it ran before you read it.
That is not a reading problem. The book did its job. It handed you a way of thinking in general form, written for every business at once. What it could not do is sit in the room on a Tuesday afternoon while you decide whether to take a client who wants something just outside what you do. That is where the framework was supposed to show up, and it never has.
The first skill I built came out of two books at once, Michael Hyatt on focus and Atomic Habits, and its whole job is planning my week. A second one came out of A Changed Mind and does something completely different. Neither is a summary of the book. Each one is the part of the book I actually use, pointed at a decision I actually make.
Why does a book you agreed with change nothing
Agreeing and applying are two different jobs, and only one fits inside your week.
Applying a framework needs three things at once. You remember it, you hold it against your own business, and you do both while the decision is in front of you. Miss one and you fall back on instinct, which built the business and is also the reason nothing changed after the book.
Owners buy books and do not finish them, and that has nothing to do with discipline. There is no space. The ones who finish land in the same place as the ones who did not, which tells you the finishing was never the variable. So stop asking yourself to remember chapter four. Put it somewhere it gets read back to you at the moment you need it.
What is a skill, in plain English
A skill is a file you write once that teaches the tool one job properly. The Made skills library is thirty-plus of them, built for client work.
Most people use an AI tool by retyping the same instructions, getting a different answer every time, and deciding the tool is unreliable. The tool is consistent. The instructions were not. A skill is those instructions, written once, in a plain text file the tool reads before it answers you.
Four things go in it: the job, what it needs to know about your business, when it should fire, and what shape the answer comes back in. No code. If you can brief a new hire on one page, you can write one.
If your tool starts from zero every time you open it, set it up to read files first.
How do you get the framework out of the book and into the file
Four moves, about an hour once the book is picked.
Pick the book you argued with
Not the most important one on the shelf. The one you finished and repeated to someone over dinner. A framework you already believe needs no persuading, so the file gets built instead of researched. Three sit on the Made resources page: Atomic Habits by James Clear, The 12 Week Year by Brian Moran, Deep Work by Cal Newport.
Write the framework in five lines, in your own words
No quotes, no copying. If five lines do not come, reread one chapter, not the book. Quoted passages make a copy of somebody’s book. Your five lines make a tool.
Hand it your business
Team size including contractors, what you sell, how a week is actually shaped, which day is already full, and the decision you make repeatedly and get wrong most. This is what makes it answer you differently from the next person.
Name what does not apply
Almost every business book was written from a context that is not yours: a large company, a solo writer, a funded product, a different decade. Applied literally, parts of it would damage a small service business. Write those parts into the file as exclusions, one line each.
Nobody does the fourth move. It is the one that makes the skill worth listening to.
A framework that is never allowed to be wrong about anything gets quietly ignored the first time it is.
Before you build anything, name the decision. One real decision you have to make this week, with a date on it. A skill tested on a hypothetical reads as clever and gets deleted by Friday.
When should it fire, and what should it hand back
Two fields, and they separate a skill you use from a file you forget.
The trigger, written as a situation. Not a topic. “When I am about to say yes to work that needs a person we do not have” beats “hiring decisions”. A situation is recognisable from inside a busy week.
The output shape, decided now. Unshaped output is an essay, and nobody reads an essay at four in the afternoon. Four parts works: the verdict in one line, what the framework says is wrong here, one alternative that fits, and the question you have not answered yet. Same shape every time, which is what turns it from a conversation into a system.
Then the book stops being something you read. It becomes a second opinion that shows up on demand and argues from the same position every time, including the times you would rather it did not. The reading half of this is how to use AI to understand what you read.
The interview that writes the file for you
Paste it into Claude or ChatGPT and answer as it asks. What comes out is the file. Name it after the book and put it where your tool reads.
I want to turn a business book I have read into a reusable skill file I point at real decisions in my business. Interview me one section at a time and wait for my answer. Do not summarise the book or reproduce passages from it. Work only from what I tell you. Where my recall is thin, ask a sharper question instead of filling the gap yourself. SECTION 1: THE FRAMEWORK Ask which book, then ask me to state its central framework in five lines in my own words. Do not improve my answer. Play it back and ask which line does the most work and which one I am least sure of. If my five lines are topics rather than a method, say so and ask for the method. SECTION 2: MY BUSINESS Ask one at a time and do not move on until you have all five: - What the business does and who for - How many people, including me, including contractors - What a normal week looks like, and which day is already full - The decision I make repeatedly and get wrong most - What I tried from this book already and abandoned, and what stopped it SECTION 3: WHAT DOES NOT APPLY This is the section everyone skips. Using Sections 1 and 2, name the parts of the framework that do not fit my business and say why, including anything that assumes a bigger team, employees rather than contractors, or hours I do not have. Give me at least three, then ask me to add the ones you missed. These become the exclusions in the file. SECTION 4: WHEN IT FIRES Ask me to describe the moment this skill should be used, as a situation I would recognise in a busy week. Push back if I answer with a category rather than a situation, and show me the difference using my own business. SECTION 5: WHAT IT HANDS BACK Propose an output shape in four parts: - THE VERDICT, one line - WHAT THE FRAMEWORK SAYS IS WRONG HERE, two or three points, each tied to a line of my five - ONE ALTERNATIVE that fits the framework - THE QUESTION I HAVE NOT ANSWERED YET Show it to me, let me change it, then lock it. SECTION 6: WRITE THE FILE Output the finished skill as one plain text block I can save, with these headings and nothing else: NAME WHEN TO USE THIS THE FRAMEWORK, my five lines MY BUSINESS, from Section 2 WHAT DOES NOT APPLY HERE, from Section 3 HOW TO ANSWER, from Section 5 HOW TO PUSH BACK, disagree with me plainly and say so when the framework has nothing useful to offer here SECTION 7: THE FIRST REAL TEST Ask me for one decision I have to make this week and run the skill on it as written. Then ask: did that tell me anything I had not already decided? If not, name the section that is too thin and rewrite it. HOW TO WORK Be direct. Do not flatter the framework and do not sell the book back to me. If what I described is a set of opinions rather than a method, say so and stop.DOWNLOAD IT AS A FILE (.MD)
Common questions
Which book should I turn into a skill first?
The one you finished and argued with, not the most important one on the shelf. A framework you already believe takes an hour to write down, because you are recording a decision rather than making one. Atomic Habits, The 12 Week Year and Deep Work sit on the resources page.
What if I cannot write the framework in five lines from memory?
Reread the one chapter that carried the idea. If five lines still do not come, the book gave you a feeling rather than a method, and a feeling has nothing to teach a file. Pick a different book and keep the feeling.
Do I need special software or any code to build a skill?
No. A skill is a plain text file your tool reads before it answers, and Claude and ChatGPT both take one. The setup that lets your tool read files at all is in how to set up AI for your service business.
How many skills should I build?
One. Use it for a month on real decisions before you build a second. Ten skills you never fire is the same shelf problem in a new folder.
Build one, then point it at something real
The test is not whether the file reads well. It is whether the next decision goes differently. Run it on the thing you have been circling this week and watch what it says about the parts of your plan you were hoping nobody would ask about.
If it tells you nothing you had not already decided, the file is thin in one place and the interview will tell you where. Fix that section only. The book is on the shelf either way. This is what takes it off.
A framework you can apply on demand is worth more than a shelf of books you agreed with. Thirty minutes, no pitch, and we will tell you which one to point at first.
BOOK A CALL →The $149 setup is where skills like this live, alongside the files that give them your business context, so a skill you build knows who you sell to before it opens its mouth.