The craft · product Grows out of manual 01, Use AI well Published craft page

Product · active

The Prompt System: the brief blocks and adversary prompts, versioned and ready to load.

Steps two and six of the manual say load a real brief and run a fixed adversary against every output. This is the actual set of briefs Stan loads, written out, versioned, and packaged so you stop re-explaining yourself to the model every morning. Strategy memo, board pre-read, term-sheet stress-test, hire-letter draft, exit conversation prep.

UseWork the tool now
FormatVersioned brief library
Grows fromManual 01, steps 02 and 06
Apply for advisory Read the manual it comes from
A named brief block loaded as the first message in a chat.

What it is

Reusable briefs that end the daily re-explanation tax.

The Prompt System is a small set of named brief blocks, each one loaded as the first message in the chat where you do that kind of work. A brief carries the role, the audience, the format, what good looks like, and the lines the model must refuse to cross. Write it once, reuse it forever.

Paired with each brief is the adversary prompt that stress-tests its output, and the verdict pattern that tells you whether the result is safe to act on. Brief, example, verdict. Three parts per block.

Most operators retype the same context every time they open a chat: who they are, who the output is for, what tone to hit, what to leave out. That tax runs about an hour a week and produces slightly different results each time because the framing drifts. A named brief removes the tax and holds the framing steady.

This is not a pack of clever one-liners. It is the set of working briefs behind real operator output, with the adversary and the verdict that keep each one honest.

What is inside

Five working briefs. Each one carries the same three parts.

01 · Strategy memo

The brief for a one-page read on a live call.

Frames the situation, the options, and the recommendation in a shape a busy reader can act on. Ships with the adversary that asks which option the memo is quietly avoiding, and the verdict line that says whether it survived.

02 · Board pre-read

The brief for the document that goes out before the meeting.

Built for a reader who has eight minutes and a strong opinion. The adversary checks for the number with no source and the claim that will not survive a director's question.

03 · Term-sheet stress-test

The brief that reads a term sheet for the clause that bites later.

Pulls control, economics, and exit mechanics into plain language. The adversary names the term the draft treats as standard when it is not. Pairs with a human read before you sign anything.

04 · Hire-letter draft

The brief for the offer and the harder note around it.

Drafts the letter and the framing of expectations. The adversary checks whether the role is described as it actually is, or as the org wishes it were.

05 · Exit conversation prep

The brief for the conversation you do not want to mishandle.

Prepares the structure, the lines, and the responses to the three reactions you can predict. The adversary asks what you are hoping the other person will not say.

Who it is for

For the operator who runs the same five conversations on repeat.

If your week contains memos, board reads, deal terms, hiring, and the occasional hard conversation, these are the briefs that turn AI from a blank box into a junior who already knows the format. It assumes you have read the manual and understand why the adversary and the verdict matter as much as the brief.

It is not for someone who wants a prompt for everything. Five briefs that you actually load beat five hundred you scroll past. The system is deliberately small.

Use it now

The page is open for operators who need the tool before the work gets expensive.

Use this page to understand the shape, the trigger, and the owner-side decision before you bring the problem into advisory.

The five briefs are in daily use. What is still settling is the example output and verdict pattern attached to each one, written so a reader who has never sat in the engagement can still tell a good result from a fluent wrong one.

If you want the working set now, the route is to apply. Advisory clients run the live briefs inside the engagement, and the standalone library goes to that list first.

What this is not

A good brief sharpens the output. It does not choose the decision.

The system makes the model useful. It does not make the call for you.

A strong brief gets you a faster, cleaner draft and a stress-test that catches the obvious failures. It does not tell you whether the memo is answering the right question, or whether the term you accepted is the one that costs you the company in three years. That read belongs to a person in the situation. The manual names the line; the advisory is where it gets crossed.

Back to the manual →
Bring the situation instead when
  • The brief keeps producing clean drafts of the wrong question.
  • The decision is irreversible inside one quarter.
  • The room around it has politics no prompt can see.
  • A wrong read costs equity, talent, or trust, not minutes.

When the work is live

Apply, and you run the live briefs now.
The packaged library reaches that list first.

Application-gated. Personal reply within 48 hours. Advisory clients use the working briefs inside the engagement before the standalone set ships.

Apply for advisory

Tier 01 Outside Read: Monthly $2,500/mo · Tier 02 from $4,500/mo · All three tiers