The Engine
The Last Prompt Engine is the technology behind The Mandate. You write a plan; a neutral AI evaluates the reasoning; the world responds. The same engine runs every world. Three exist today: Colony, Corporate Reckoning and Lockwood.
The engine is domain-agnostic. A skin supplies the environment — its people, its events, the conditions they live under. The evaluation never changes.
A prompt, in its oldest sense, is a nudge — a cue that moves thinking in a new direction without dictating where it lands. In The Mandate, your advisors prompt you. Your question prompts them. The crisis itself prompts the situation.
But the last prompt — the only input that actually moves the world — is yours. The plan you write, in your own words, which a neutral AI then receives and evaluates. It never writes the plan, and never improves it. We named the engine for that moment: the last word before consequences unfold.
This has no connection to AI prompt engineering, LLM tooling, or anything related to writing prompts for AI systems. The word means what it meant before large language models existed.
Six steps. Every cycle.
The same loop runs in every skin. The context changes. The engine does not.
An event arrives, chosen by the state of the world your earlier decisions created. No two runs meet the same sequence.
Someone in the world brings it to you — a colony advisor, a department head, an archetype. They tell the truth as they see it, from their own corner, shaped by their expertise and their fears. Partial by design, because real expertise is partial.
You may ask that advisor one question. It is there so that no one is shut out of a decision by an unfamiliar word or field — what is a CISO? what is a TLS breach? The answer still comes from their corner.
No menus. No options. Four parts in plain English: the goal, the actions, the contingency, the communication. You write it alone.
A neutral AI reads the plan against six criteria and writes what it saw. Not what you chose — how you reasoned.
The result limits how far the conditions can move. Your plan moves them — including any it spent to solve the problem — and the conditions interact: when one falls too low, it drags on others. Then the world chooses the next event, from the world you just made.
The engine keeps two layers apart. Inside the world, the character — Jon Kelly, Jon Roddy, the Traveller — lives only the consequences. They never see the Mandate's evaluation. Outside it, you see both: what happened, and how the reasoning produced it.
Not everything is visible, even from outside. What drives each advisor — their fears, their doubts — shapes every report but never appears on screen. As in life, you read it in what they say.
See it on a real decision: The MirrorThe world doesn't wait while you're busy elsewhere.
Chronosymbiosis is the principle the engine is built around: you live with the consequences of your decisions, even when new information arrives after you made them. Nothing a decision changes is wiped clean at the next event.
Every event is the marble — vivid, urgent, demanding all the attention in the room. The conditions are the glass. While you deal with the marble, the glass keeps shifting: what you attend to steadies; what you don't, grows. A run ends when the glass breaks — when any condition reaches zero — not when a plan scores poorly.
That is also how the world chooses what comes next. The weakest condition draws the next event into its own domain — not to punish a weak spot, but because neglected things grow. Solve a food shortage by spending security, and the next crisis may arrive at the perimeter.
And calm is not the same as safe. When nothing is visibly failing, the next thing to arrive is the thing nobody saw building — the Black Swan. Some events carry consequences no plan can fully prevent; even a strong response leaves damage. That isn't the world punishing stability. It's what was already growing where nobody was looking.
The story behind it → The Glass
You already know how this works.
Colony
Ever tried to get anything done when nobody’s speaking to each other?
When cohesion is low, every gain comes harder.
What happens to the food when the roof leaks and the tools are broken?
When infrastructure is low, food losses deepen.
Corporate Reckoning
Remember the buy-out rumour, when efficiency died overnight?
When team engagement is low, every gain comes harder.
What does it cost every time the systems go down?
When operational infrastructure is low, cash-flow losses deepen.
Lockwood
What becomes of a discovery — iron, electricity, the first computing machines — when the people it reaches are too divided to share it?
When collective coherence is low, every gain comes harder.
How can we embrace an unknown future while we are still struggling to understand our present?
When temporal awareness is low, perceptual losses deepen.
Six criteria. One question: how did you reason?
The Mandate reads every plan against the same six criteria, in every world. After each plan you see the overall result, and what the Mandate noticed under each criterion.
What you won't find here is a scoring guide. The criteria tell you what is being looked at; the rest you discover by reasoning.
How the evaluator is held honest.
An AI evaluator is only worth something if it can't be flattered, padded or argued into a better result. These are the rules it works under.
A criterion is only marked in full when the plan meets all of it, with no meaningful gap. When in doubt, the evaluator is told to mark lower, not higher.
If the evaluator’s reasoning names a risk the plan missed, it cannot mark risk anticipation in full.
The AI returns its marks for each criterion. The engine adds them up and sets the result itself — the AI never gets to declare a plan Strong.
Each observation is written to stand on its own, with no score in it — something noticed, not something marked.
Adding detail to the plan doesn’t move it. Reasoning beyond the plan does: what it costs, what it risks, what happens if it fails.
They answer from their own expertise and fears, in human terms — ‘morale will shatter’, never ‘+2 cohesion’ — and never offer a balanced, all-domain answer.
Engine vs. environment
The engine is the same in every world. Everything a world contains is written as data, and can be replaced without touching the engine.
Identical in every skin.
What a skin is made of.
Same engine, three environments
| COLONY | CORPORATE RECKONING | LOCKWOOD |
|---|---|---|
| Sustenance | Cash Flow | Cognitive |
| Health | Employee Well-Being | Moral |
| Security | Regulatory Compliance | Perceptual |
| Cohesion | Team Engagement | Collective |
| Infrastructure | Operational Infrastructure | Temporal |
Know an environment worth building?
The engine is modular. A new skin is a new environment — its people, its events, the conditions they live under — written as data, with the engine unchanged underneath. If there's an environment your people actually work in that deserves one, we'd like to hear about it.