Prompts.
Copy. Run in ChatGPT or Claude with one click. Stolen and refined from issues you should be reading.
The Compute Leak Auditor
An eight-question interview that finds where your compute spend is buying overhead instead of computation, then ranks the fixes by saving divided by effort.
The Parts Audit
Interviews you about your facility's installed base, then produces a prioritized list of undocumented parts worth capturing in CAD - ranked by replacement risk, lead time exposure, and production criticality.
The Assurance Decoder
Separates the AI safety and security claims you can actually verify from the ones you are taking on trust - then names the single unverifiable claim that is carrying the most weight in your decisions.
The Ownership Auditor
Traces every AI tool you use back to its real owner, its model provider, and its infrastructure - then names the single owner whose disappearance would cost you the most.
The Delegation Auditor
Finds the places where an automated system could do something you would have said no to. Interviews you about one setup, then ranks the real gaps by consequence rather than by how easy they are to fix.
The Agent Swarm Architect
Designs a multi-agent structure for a real task you are stuck on — with an explicit failure budget, validators kept separate from producers, and a stop rule.
The Refusal Decoder
An assistant turned you down. Work out whether you hit a policy boundary, a real capability limit, or your own vague request - and what the next move is for each.
The Vendor Claim Tracer
Take a statistic you have seen quoted about an AI product and trace it back to whoever actually asserted it — the vendor, a customer, a third party, or nobody at all. Ends with a verdict on whether you can repeat it.
The Flavor Pairing Interrogator
Plans tonight's cooking experiment around what is actually in your kitchen — one high-overlap pairing, one low-overlap pairing — and states plainly which parts come from published data and which are its own guess.
The Pattern-vs-Meaning Auditor
Take a pattern you have spotted in your own numbers and find out whether you actually know why it happens — or only that it happens.
The Agent Team Designer
Interviews you about your actual weekly work, one task at a time, then returns a delegation map: what an agent should own outright, what it should draft for your review, and what stays yours — plus the specific check you need for each one.
The Provable Claim Auditor
Walks you through one decision you are about to trust AI with, and tells you whether it is the kind of question that can be checked against a rule, or the kind where you are trusting fluency and calling it confidence.
The Quiet Change Pre-Mortem
Pressure-tests a change you are about to ship — a price, a limit, a policy, a feature you are removing — and tells you who will actually see it, what the worst honest headline is, and what to announce first.
The Agent Access Auditor
Decide account by account which ones an AI browsing agent should be allowed to touch — and which to lock away — with a green, yellow, or red verdict for each.
The AI Travel Companion Interview
Turn AI into a trip planner that remembers your preferences for the whole trip — not just for one question.
The Family Document Digitizer
Turn any old family document into a clean, shareable write-up — with a privacy check and a fact-check built in.
The World-Model Cause-and-Effect Trainer
Experience the difference between predicting words and predicting cause-and-effect through guided interview questions.
AI Layoff Exposure Advisor
The Model Tier Routing Audit
Most workflows were built when one model tier was the only sane default. This prompt interviews you about one workflow you already run, step by step, then tells you which tier each step actually needs, where downgrading would be a mistake, and the single change that saves the most for the least risk.
The AI Compute Dependency Evaluator
The Ruthless Line Editor
Interviews you about your own paragraph, then hands back only cuts, flags, and a one-line verdict — never a rewrite in someone else's voice.
The AI Mode Connected-Apps Audit
Interviews you about a repetitive task, then tells you plainly whether AI Mode's Instacart/Canva/YouTube Music integrations actually save you a step — or not.
The 'Is This Actually Done?' Audit
Interviews you about one piece of AI-produced work, then tells you whether it's actually done or just looks done -- built around what the Remote Labor Index found about where AI deliverables actually fail.
The AI-Hype Headline Check
Interviews you about an "AI did X" headline you just read, walks you to the primary source, and checks whether the paper's own methodology language actually supports the claim — or whether the press release oversold it.
