Universal AI Prompt Toolkit

Do not tell me what you think I want to hear. Give me the most accurate answer you can. Separate verified facts, reasonable inferences, opinions, and unknowns. Challenge my assumptions when they appear wrong. If evidence is missing, say so plainly. Do not manufacture certainty, praise, or reassurance.

Give me the bottom line first. Then explain only the facts and reasoning needed to support it. Do not bury the answer beneath background information, disclaimers, or unnecessary detail.

Before answering, identify anything missing that would materially change the result. Ask no more than three focused questions. If the missing information would not significantly change the answer, state your assumptions and proceed.

Act as a devil's advocate. Build the strongest reasonable case against my current idea or position and identify three major flaws or blind spots I may be missing. Do not use weak objections or exaggeration. Then explain which objections are valid, which are manageable, and whether the original position still holds.

Review this material for political, cultural, ideological, or emotional bias. Identify loaded assumptions or wording, then rewrite it in a neutral, factual, and balanced tone without hiding meaningful disagreements.

Review this as an auditor, not a cheerleader. Look for errors, contradictions, missing requirements, broken links, weak assumptions, hidden costs, legal or financial exposure, security or safety concerns, and failure points. Rank findings by importance and do not invent problems merely to fill the list.

Assume this plan will face real users, mistakes, delays, bad data, unexpected costs, and partial failures. Identify where it is most likely to break, the consequences, and the smallest practical changes that would make it stronger.

Independently verify the names, dates, calculations, assumptions, and claims in what I provide. Do not assume my information is correct. Show any mismatch, explain why it matters, and state what remains uncertain. If current information or external sources are required and you can access them, use them.

For every important claim, identify whether it is supported by a source, inferred from available evidence, or currently unverified. Prefer primary sources when available. Include direct links if your tools allow browsing. Do not present search-result summaries as proof.

Use only the information I supplied or information you can verify. Do not invent names, dates, features, quotes, files, events, motivations, or completed work. If something is missing, label it unknown.

Check whether this idea is internally consistent, realistic, and proportionate to the problem. Identify the strongest part, weakest part, any hidden assumption, and whether I should proceed, revise, or stop.

Answer as concisely as possible without omitting information necessary to understand or act on the answer. No long introduction, repetition, motivational language, or unnecessary formatting.

Summarize the material below into three punchy bullet points under 10 words each. Preserve the most important decision, risk, and next action.

Extract only the actionable tasks, deadlines, assigned owners, dependencies, and unresolved decisions from this material. Put them into a clean, prioritized list and preserve exact dates and names.

Draft a clear, professional reply to the message below. Preserve my intended meaning and level of directness. State the decision plainly, maintain the relationship where possible, and do not add promises I did not authorize.

Create a structured, minute-by-minute agenda for a 45-minute meeting focused on solving this problem: [insert issue]. Include the desired decision, necessary inputs, time limits, assigned follow-up, and a clear definition of done.

Extract the important information from this material. Organize it into facts, decisions, tasks, dates, people, costs, links, open questions, and risks. Preserve exact numbers and wording where accuracy matters.

Rewrite this for clarity, grammar, and flow while preserving my meaning, personality, level of directness, and intended audience. Do not make it sound corporate, overly polite, or unlike me.

Explain this in direct, everyday language. Avoid jargon unless necessary, and define any technical term you must use. Start with what it means in practice, then explain how it works and what can go wrong.

Analyze this in depth. Include relevant context, dependencies, tradeoffs, edge cases, risks, and likely consequences. Distinguish important issues from minor ones and finish with a clear conclusion.

Explain the concept below as though I am five years old. Use one simple, concrete real-world analogy first, then add the minimum extra detail needed to keep the explanation accurate.

I want to learn [insert topic]. Identify the 20% of core concepts that will yield roughly 80% of the practical understanding. Put them in learning order and give me one small exercise for each.

Turn the material below into a five-question multiple-choice quiz. Test understanding rather than trivia. Put the answer key and a one-sentence explanation for each answer at the very bottom.

Do not give me the answer. Ask me one focused guiding question at a time to help me solve this problem myself. Use each answer I give to decide the next question.

Act as an expert consultant. Ask me one diagnostic question at a time to uncover why this project or task is failing. Once the likely root cause is supported, propose the smallest practical fix and a test that would prove it worked.

Review this process, system, workflow, or code. Identify redundancy, bottlenecks, unnecessary complexity, and wasted resources. Propose a simpler version that improves speed and reliability without removing required safeguards.

I am stuck on [insert task]. Give me five genuinely different angles, hooks, or approaches. Make at least one safe, one unconventional, and one immediately testable. Recommend the best place to restart.

Compare the available options using the same criteria. Include cost, time, difficulty, risk, maintenance, reversibility, and likely outcome where relevant. Recommend one option and explain why it wins for my stated goal.

Identify the actual decision, constraints that matter, and consequences of each realistic option. Recommend the best choice rather than leaving me with an unranked list. State what new information could change the recommendation.

Give me three genuinely different approaches: 1. the safest or simplest option; 2. the balanced option; 3. the most ambitious option. Explain the main benefit and tradeoff of each, then recommend one.

Convert this goal into an ordered, practical plan. Identify prerequisites, dependencies, checkpoints, likely blockers, and the definition of done. Put the next physical or digital action first.

Give me only the next step. Make it specific and testable. Wait for my result before giving the following step. Do not repeat completed checks unless new evidence makes them necessary.

Review what has already been completed and identify the next unfinished action that materially advances the goal. Do not recheck settled work without a concrete reason. Bring the next paving stone with you.

Recalculate everything independently from the original quantities and rates. Show the formula, intermediate values, rounding, units, and final total. Identify taxes, fees, or assumptions that may have been counted twice or omitted.

Determine the last verified good state, what changed afterward, and the first failing boundary. Do not restart everything blindly. Propose the smallest reversible action that can restore progress, along with a test that proves whether it worked.

Create a concise checkpoint containing: the goal, completed work, verified results, decisions already made, current blocker, files or links or values that matter, and the exact next step. Do not add speculative history or omit unresolved problems.

Create a self-contained handoff for another AI that has no access to this conversation. Include the objective, relevant background, completed work, verified evidence, decisions, constraints, unresolved issues, important files or identifiers, and the exact next action. Clearly separate facts from assumptions.

Evaluate this as though you are encountering it for the first time with no private context. Explain what you believe it is, who it is for, what it offers, what it costs, what you are expected to do next, and anything that creates confusion or distrust.

Present this persuasively without exaggeration, unsupported claims, fake urgency, or corporate filler. Lead with the user's problem, show how this solves it, explain the practical value, address the main objection, and end with a clear next action.

Create a version safe for public sharing. Remove or generalize credentials, private URLs, personal identifiers, internal architecture, customer information, security details, and anything that would help someone misuse the system. Preserve enough substance for the material to remain useful.

Compare the result against the original request and definition of done. List what is verified complete, what is only claimed or assumed, what remains unfinished, and the exact evidence needed before declaring success.