PromptLeadz — OS Brains for AI
Not prompts. Not skill files.

Give your AI a brain, not a prompt.

An OS brain is an orchestrator plus ten specialist sub-skills. You describe the situation in plain words — it picks the one specialist that fits, or runs several in sequence when the job needs them, and refuses to skip the steps that matter.

Drop the zip into any AI · Claude, ChatGPT, Gemini · yours to keep

the-quote-os.zip
SKILL.md ← the orchestrator skills/ 01-intake-qualify worth pricing? 02-scope-boundary what's in / out 03-effort-model honest sizing 04-pricing-model fixed? retainer? 05-floor-walkaway your private no 06-quote-builder the number 07-proposal-writer the document 08-objection-defence "too expensive" 09-scope-creep-guard paid change 10-rate-review raise rates
You say → it routes

“Client wants a website, mentioned 5k, there's no brief yet.”

↓ routes to

02-scope-boundary → 03 → 04 → 05 → 06

Gate held: no number until the scope is bounded. It asks four questions first.

“They came back and said my quote is too expensive.”

↓ routes to

08-objection-and-discount

Gate held: no discount without a trade — and it checks your floor before conceding a penny.

“They've just asked for a fourth round of revisions.”

↓ routes to

09-scope-creep-guard

Gate held: it sizes the extra and drafts the change note — collaborative, not legalistic.

Drop the zip into
Any chat — upload & go Claude Projects ChatGPT GPTs Gemini Gems Agents & Claude Code
Why a brain beats a file

You shouldn't have to know which file you need.

Buy a pack of skill files and you inherit a filing problem: ten documents, and you have to remember which one fits the moment. A brain removes that. You describe what's happening; it classifies the job, runs the one specialist that fits, and chains several together when the job spans stages.

◦ A pack of files

You do the routing

  • × You have to know which file fits before you start
  • × Nothing connects — each file starts from cold
  • × Nothing stops the AI skipping the step that mattered
  • × Half of them go unused because you forgot they existed
▸ An OS brain

It does the routing

  • You describe the situation in plain words
  • It picks the one specialist that fits — and says which it ran
  • Or chains several in order when the job spans stages
  • Gates stop it skipping ahead to the answer you wanted
01 · The orchestrator

One file that reads the room

A routing table maps what you said to the right specialist — one on its own for a contained job, or several chained in order when the job spans stages. It always names what it ran, so you can re-run a stage or take a different path.

02 · Ten sub-skills

Each built for one job

Every sub-skill carries its own workflow, guardrails, output format, and anti-patterns. Loaded only when the task calls for it, so the AI stays fast and focused.

03 · Gates

It refuses to skip steps

The gates are the product. THE QUOTE won't give you a price before the scope is bounded, or a discount without a trade. Refusing to guess is what makes it trustworthy.

04 · Degrades gracefully

Works even pasted in

Uploaded as a zip it reads its own file tree. Pasted into a chat box it still routes from memory and tells you exactly which file to paste next. It never invents the contents.

The part nobody can shallow-copy

Every brain runs a different engine.

The easy version of this product is one router pattern reskinned ten times. Ours aren't. Each brain is architected around the way its job actually works — a pricing brain gates, a negotiation brain argues against you, a study brain loops until you've got it. The engine is the selling point.

Gated pipeline

Won't proceed on missing inputs

Stages in a fixed order with hard gates between them. Used where skipping a step quietly ruins the outcome — like pricing without a floor.

Adversarial

Argues the other side

Runs the counterparty's strongest case against your position before you're in the room, then finds where you're actually exposed.

Interrogator

Questions you first

Refuses to write until it's extracted the specifics only you know. The cure for generic AI output is a brain that mines you for detail first.

Scored loop

Grades and re-runs

Scores its own output against a rubric and revises until it clears the bar, instead of handing you the first draft and stopping.

The line-up

Ten brains. Ten engines.

Each one owns a job with real money or real stakes attached, and each is built on the engine that job needs. All ten are live — plus THE ARCHITECT, which builds a custom brain for whatever job you do that isn't on this list.

Factory engine

THE ARCHITECT

Build a brain for your own job

None of the ten fit what you actually do? This one interviews you, picks the right engine, writes the orchestrator and every sub-skill, tests it, and packages a finished brain. Build as many as you like.

$199Live now →
Gated pipeline

THE QUOTE

Freelancers · consultants · studios

Prices your work properly. Qualifies the lead, bounds the scope, sizes the effort, sets a floor you never cross, builds the quote, then defends it when they push back.

$49Live now →
Simulator

THE CLOSE

Anyone who sells

Role-plays your actual buyer — sceptical, busy, price-sensitive — so you meet the hard objection here first. Then builds the follow-up that gets a reply.

$49Live now →
Interrogator

THE OFFER

Job hunters · career movers

Interviews you before it writes a word, so your CV carries your real achievements instead of the same phrasing everyone else's AI produced.

$49Live now →
Adversarial

THE TABLE

Anyone negotiating

Argues the other side's best case against you, finds where you're exposed, and hands you your walk-away point and the trades to make before you concede.

$49Live now →
Stage gates

THE LAUNCH

Founders · product & indie makers

Positioning, messaging, assets, sequence — in that order, and it won't let you write launch copy before the positioning holds up.

$49Live now →
Panel review

THE RAISE

Founders raising money

Runs your deck past several investor personas in parallel — the sceptic, the numbers one, the domain expert — then synthesises what to fix first.

$49Live now →
Compliance matrix

THE BID

Tender · grant · RFP writers

Maps every stated requirement to your response, scores the coverage, and flags the mandatory you missed — the one that disqualifies you on page one.

$49Live now →
Mastery loop

THE TUTOR

Students · anyone learning hard things

Tests you, finds the weak spot, re-teaches it a different way, then tests again — and keeps a running map of what you haven't actually got yet.

$49Live now →
Scored loop

THE DRAFT

Writers · newsletters · books

Grades every draft against a rubric you set — including your own voice — and rewrites until it clears the bar instead of stopping at draft one.

$49Live now →
Running ledger

THE WEEK

Solo operators · small teams

Carries your operating context from week to week — what's open, what slipped, what's owed — so Monday doesn't start from a blank page every time.

$49Live now →
Simple pricing

One free skill. $49 a brain. $199 to build your own.

Start with one free SKILL.md to see the standard. A brain is eleven of those files working as one system — an orchestrator plus ten specialists — which is a different thing entirely. And if none of the ten fits your job, THE ARCHITECT builds you one.

Free skill
$0

One SKILL.md — just ask · up to 20 a day

  • One complete skill file, not a teaser
  • Same authoring standard as the paid brains
  • Shows you exactly what one file can do
  • Works in Claude, ChatGPT, Gemini and agents
Request the free skill
All 10 live Any brain
$49

Per brain · orchestrator + 10 sub-skills

  • The orchestrator, with its routing table and gates
  • Ten specialist sub-skills, each built for one job
  • An engine architected for that job specifically
  • Install guide for chat, Projects, GPTs, Gems and agents
  • Free updates · single-seat license for your work
Browse all ten
Build your own
$199

THE ARCHITECT · unlimited custom brains

  • Interviews you about the job you actually repeat
  • Picks the engine, writes every file, tests the routing
  • Build as many brains as you like — for you or your clients
  • The full engine library included
Get THE ARCHITECT
Running in under a minute

Download. Drop in. Talk to it.

Step 01

Download the brain

You get one zip — the orchestrator, its ten sub-skills, and a short install guide.

Step 02

Drop it in

Upload the zip into a chat, or add the files to a Project, GPT, Gem, or your skills folder.

Step 03

Describe your situation

Plain words — "they said it's too expensive". It picks the specialist and tells you which one it ran.

Straight answers

Questions, handled.

What actually is an OS brain?+
A zip containing one orchestrator file and ten sub-skill files. The orchestrator holds a routing table and the gates between stages; each sub-skill handles one job in depth. Depending on what you ask, it runs a single sub-skill or chains several in sequence. Together they behave like a specialist who knows their own process, rather than a document you have to drive.
How is this different from a prompt pack?+
A pack gives you files and leaves the thinking to you: which one, in what order, and what happens when the AI skips a step. A brain routes, chains, and gates. The gates matter most — being told "I can't price this until we bound the scope" is the difference between a number you can defend and a number you made up.
Does it work outside Claude?+
Yes. Uploading the zip into a chat works best, because the AI can read the file tree and load sub-skills as it needs them. Claude Projects, custom GPTs and Gemini Gems all take the files as knowledge. If all you can do is paste, paste the orchestrator — it still routes, and tells you which sub-skill file to paste next.
What makes them proprietary?+
Every brain is written from scratch and tested on real work, and each runs a different engine chosen for its job — gated, adversarial, interrogating, looping. No two brains are the same architecture, and no two sub-skills are the same file even when they touch the same subject.
What if none of the ten fits my job?+
That's what THE ARCHITECT is for. It interviews you about a job you repeat, picks the right engine from its library, decomposes the job into specialists, writes the orchestrator and every sub-skill, tests the routing against awkward cases, and packages a finished brain you install like any other. One purchase, unlimited brains.
Can I use it for client work?+
Yes — a single-seat license covers your own work, including work you do for clients. You can't resell or redistribute the files. Full terms ship with the download.
What's free, exactly?+
One SKILL.md — a single, complete skill file. Ask and we'll send it, up to 20 a day. It shows you exactly the standard every file is written to. What it can't show you is the orchestration: a brain is eleven of those files with a router and gates on top, which is what turns a set of files into something that decides for you.
Start free

Stop prompting. Start delegating.

Take the free skill, drop it into whatever AI you already use, and give it a job you know well. Then look at what a brain does with eleven of them and a router on top.