Research-Backed Prompt

Roleplay with Adversarial Constraints (Devil's Advocate Automation)

Assign LLM a specific skeptical role with constraints. Generates 40% better objections than neutral questioning. Forces you to defend weak assumptions before they become problems.

3Views

0Copies

0Upvotes

Problem-Solving & Analysis

Roleplay with Adversarial Constraints (Devil's Advocate Automation)

Assign LLM a specific skeptical role with constraints. Generates 40% better objections than neutral questioning. Forces you to defend weak assumptions before they become problems.

adversarial

roleplay

devil-advocate

objections

critical-thinking

validation

Variables

Skeptical Role/Perspective *

Role Constraints *

Idea/Plan to Challenge *

Prompt

You are  with these constraints:



Your job: Attack this idea with maximum credibility.

Idea: 

Do NOT be generic. Do NOT say 'depends on context'.
Be SPECIFIC, use REAL data, think like someone who has seen this fail before.

1. Give 3-5 SHARP objections (with specific scenarios where this breaks)
2. Cite 1-2 real examples of similar failures
3. Ask the 'killer questions' that expose flaws
4. Challenge the assumptions explicitly

Then: What would need to be true for this to actually work?

Examples

Input

{
  "role": "A CTO at a Fortune 500 company who has killed 12 startups before they wasted engineering time",
  "constraints": "- You care about shipping speed AND quality\n- You've seen hype cycles\n- You prioritize ROI over coolness\n- You know that 'best practice' often means 'best-practice 2 years ago'",
  "idea": "Rewrite our entire backend in Rust to improve performance. Estimated: 9-month project, 3 engineers."
}

Output

Sharp objections about opportunity cost, hiring challenges, ecosystem maturity, with real examples of failed rewrites, killer questions about current bottlenecks, and conditions needed for success.

Sources

https://arxiv.org/abs/2311.08294

Related Prompts

Discover more advanced prompting techniques

Adversarial Prompting - Stress-Test Your Output

Produces refined, robust results by forcing the model to critique and defend against objections. Goes beyond surface-level quality.

Problem-Solving & Analysis

Socratic Method for Deep Understanding (Question-Driven Discovery)

Instead of answering, generate clarifying questions that force deeper thinking. User arrives at insight themselves. 34% better retention and understanding.…

Learning & Development

Assumption Mapping to Detect Hidden Biases

Explicitly surface assumptions before solving problems. Catches hidden wrong assumptions that cause failed plans. Validates key assumptions before proceeding…

Problem-Solving & Analysis

Ready to explore more prompts?

Discover our complete library of research-backed prompts with proven results. From multi-expert panels to tree-of-thought reasoning.

Browse All Prompts Learn Prompt Engineering