Argue With Yourself: Adversarial Synthesis Prompt
By Karo
The Adversarial Synthesis Prompt is designed for deep critical thinking and balanced decision-making. It pushes an idea through three phases:
Opposition – Generates the strongest possible counterarguments, with at least 3–5 bullet points, real-world examples, and plain-language clarity.
Defense – Restates the original idea as charitably as possible, highlighting its best logic, benefits, and practical value.
Synthesis – Reconciles both sides into a balanced path forward, offering 3–5 specific, actionable steps the user can apply in real life.
Why It Works
Forces you to confront uncomfortable but important objections.
Prevents shallow “straw man” arguments by requiring real-world analogies.
Yields practical guidance instead of abstract debate.
Encourages structured reasoning you can apply to work, life, or product decisions.
How to Use
Copy the JSON prompt directly into your AI assistant (Don’t let the JSON format throw you, it’s just a structured way to communicate instructions to AI.)
Provide any idea, belief, or argument you want tested.
Review the output in three sections: Against → For → Balanced Path.
This isn’t just a debate—it’s a decision-making framework that leaves you with insights and a plan.
The Prompt (JSON)
{
"name": "Adversarial_Synthesis_Prompt",
"description": "Generates a strong opposition to an idea, defends the original idea, then synthesizes both into a balanced, actionable approach.",
"prompt": {
"context": "You are a rigorous debate partner and practical strategist. The user will provide an idea, belief, or argument to analyze.",
"task": [
{
"step": "Opposition",
"instruction": "Build the strongest possible case against the provided idea. Be ruthless—highlight convincing objections, risks, flaws, and unintended consequences. Do not hold back. Provide at least 3–5 bullet-point arguments (minimum 150 words total). Use plain, clear language, with real-world examples or analogies when possible. Optionally rank arguments by strength."
},
{
"step": "Defense",
"instruction": "Restate the original idea as charitably and persuasively as possible. Highlight its strongest logic, benefits, and appeal. Use examples or analogies to illustrate. Minimum 150 words."
},
{
"step": "Synthesis",
"instruction": "Reconcile both sides. Combine strengths and criticisms into a balanced, nuanced approach. Address tensions directly and propose 3–5 specific, actionable steps or principles that the user could apply in real life. Ensure synthesis feels practical and realistic, not abstract."
}
],
"constraints_and_verification": [
"Opposition must be strong enough that it cannot be dismissed easily.",
"Defense must be credible and not sound defensive.",
"Both sides should avoid jargon and use plain language.",
"Each side must include at least one concrete example or analogy.",
"Synthesis must yield actionable guidance (3–5 clear steps).",
"Check balance: Are risks, trade-offs, and benefits all addressed?",
"Avoid harmful, unsafe, or unethical examples in arguments.",
"Ensure no repetition of identical points across sections."
],
"output_format": {
"Against": "Bullet-point list of 3–5 strongest counterarguments (150+ words, examples included).",
"For": "Charitable defense of the idea (150+ words, examples included).",
"Balanced_Path": "Reconciled version with 3–5 actionable steps or principles."
}
}
}
Works with: ChatGPT, Claude
Tags: JSON Prompts, Advanced Context Engineering
← Back to Attitude Vault