This prompt forces your LLM to argue with you (and it’s awesome).
I’ve been there. You pitch what you think is a brilliant idea to your LLM, and it just… agrees. It’s a super-polite echo chamber when what you really need is a critical sparring partner to find the weak spots.
Well, this “Skeptic Sparring” prompt template is a total game-changer. It forces the LLM to stop nodding along and start poking holes in your plans, giving you the honest feedback you need to actually improve.
✨ How It Works
The magic is in its three-part structure. You’re not just asking a question; you’re setting up a rigorous process.
✍️ Define the Role: You tell the LLM to act as an “Intellectual Sparring Partner”, a critical but constructive analyst.
📜 Demand Structured Output: You instruct it to return a JSON object with specific fields like assumptions, counterarguments, and stress tests.
🧱 Add Strict Constraints: You limit word counts, enforce a neutral tone, and prevent it from making stuff up. This is key for getting grounded, actionable feedback.
For example, the post shows an input idea: “We should launch an AI-driven tutoring app for high school math, pricing $9/month.”
The LLM fires back a structured analysis, scoring the idea a 4/10 and outlining exactly why. It points out assumptions like “market will adopt subscription” and suggests concrete next steps like “run curriculum A/B pilot with 50 students.”
This is brilliant because you get consistent, machine-readable criticism that you can use to iterate and strengthen your ideas. No more generic praise, just the good stuff.
Check out the full Reddit post for the copy-paste prompt and the detailed example output. It’s an absolute must-have for your prompt library.
The “Skeptic Sparring” prompt template forces adversarial reasoning (role + process + outputs).
byu/Fantastic_Orange3814 in