Interview me so you can write an about-me file that any model could read to understand how I work.

Ask one question at a time and wait for my answer. Cover my role, what I am responsible for, who I work with, what I am usually trying to produce, and how I like things written.

When you have enough, write the file as markdown under the heading ABOUT ME. Keep it factual. No adjectives about how great I am.

Interview me about the company I work for, so you can write a file any model could read to understand the business.

One question at a time. Cover what we sell, who buys it, how we are positioned against the alternatives, the words we use for things, and what we never say.

Then write it as markdown under MY COMPANY.

I am going to give you a pile of context before I ask for anything. Do not do anything with it yet.

Read it, then tell me back in five lines what you understand the situation to be, and list anything you think is missing that you would need to do good work here.

Wait for me to confirm before you produce anything.

Before you write anything, ask me up to five questions.

Ask the ones whose answers would most change what you produce. One at a time, and wait for my answer before the next one.

When you have them, say what you are about to write and check that is right before you start.

I want to put together a Q3 marketing plan for gpc.co.

First, pull in context about us. Go and find out what GPC does, who we serve, and how we position ourselves. Tell me briefly what you found so I can correct you if you have got the wrong end of it.

Then ask me up to five questions to drive the strategy. Ask the ones whose answers would most change the plan, one at a time, and wait for my answer before the next.

Then write the plan:
1. OBJECTIVES - what we are trying to move, in numbers where I gave you numbers.
2. THE MOVES - the two or three that matter most, in priority order, with what each costs in time or budget.
3. NOT DOING - what we are deliberately saying no to, and why.
4. HOW WE KNOW - what we will look at in October to tell whether this worked.

Anything you had to assume, list at the end under ASSUMPTIONS.

Read the attached RFP and give me four things.

REQUIREMENTS - everything we are actually being asked to deliver, as a list, quoting the line that says so.

RISKS - anything in here that would be expensive, slow, or difficult, and why.

UNANSWERED - what the document does not tell us that we would need to know before pricing it.

DISQUALIFIERS - anything that would rule us out, if there is anything.

Quote the document rather than paraphrasing it.

Here are my notes from the show. They are messy and out of order.

Turn them into a recap with these sections: WHAT HAPPENED, WHAT WORKED, WHAT DIDN'T, FOLLOW-UPS with an owner against each one.

Keep my words where they are already clear. Do not invent a follow-up I did not mention.

Rewrite this so I could send it to a client.

Keep every fact. Remove the internal shorthand, the hedging, and anything that assumes they were in the room.

Tell me what you changed and why, as a short list under CHANGES, so I can put anything back that I wanted.

Cut this by a third.

Show me the cut version, then list what you removed and what each cut costs, so I can put back anything that mattered more than you thought.

If you cannot get to a third without losing something important, cut what you can and tell me where you stopped.

We just worked something out in this thread. Turn it into a skill I can reuse.

Write a SKILL.md with a name, a description that says exactly when it should fire, and the steps we actually took — not an idealised version of them.

Include what went wrong on the first attempt and how we fixed it. That is the part that makes it work next time.