Most people guess their way to average results, but you can scientifically clone success with the right instructions.
The Problem with Guessing
We have all been there. You see a brilliant email, a viral social media post, or a perfectly structured report, and you try to get ChatGPT to write something similar. You type things like “make it professional” or “write it like a viral post,” but the result is usually generic and flat. The AI doesn’t understand the nuance of why that original piece was so good. I used to waste hours tweaking instructions, trying to capture a specific voice, until I found a better way.
This incredible “Reverse Engineering Mega-Prompt” was shared by a Reddit user named EQ4C, and it completely solves the problem of vague instructions. The creator designed this tool to act as a “Linguistic Forensic Analyst.” Instead of asking you to describe what you want, it asks you to provide a “Gold Standard” example. It then analyzes that example down to the molecular level, looking at sentence structure, emotional triggers, and formatting quirks, and writes a brand new, highly detailed prompt that you can use to replicate that exact style forever. It essentially allows you to steal the “DNA” of any text you admire.
💡 The “Forensic Audit” Approach
The core genius of this tool lies in how it dissects information. It doesn’t just read the text; it performs a deep audit to understand the psychology behind the words. Here is why this approach works so well.
1. It Goes Beyond Basic Tone
Most of us stop at basic adjectives when describing a style. We might say a piece of writing is “friendly” or “formal.” The creator of this prompt realized that this isn’t enough for a Large Language Model (LLM) to truly mimic a style. This tool forces the AI to conduct a “Dimension Analysis.” It breaks the writing down into specific pillars:
Tone & Voice: It looks for specific descriptors. Is the text authoritative yet empathetic? Is it satirical? Is it clinical? This helps the AI understand the attitude of the writer.
Pacing & Rhythm: This is a subtle but critical detail. The prompt analyzes if the sentences are short and punchy, or if they are long, flowing narratives. It looks for rhythmic complexity that keeps the reader engaged.
Structure & Layout: Does the text use an inverted pyramid style? Does it rely on modular blocks or nested lists? Understanding the physical shape of the text is just as important as the words themselves.
2. It Identifies the Emotional Driver
One of the hardest things to prompt for is emotion. We often forget to tell the AI how the reader should feel. This expert’s prompt includes a specific step for “Emotional Intention.” It asks the AI to determine if the text is meant to create urgency, build trust, or spark curiosity. By identifying the secondary emotional drivers, the resulting prompt ensures that future outputs aren’t just factually correct, but emotionally resonant. It bridges the gap between robotic text and human connection.
3. It Creates a “Master Prompt” for You
The final output isn’t just an analysis; it’s a ready-to-use tool. After the AI finishes its forensic audit, it synthesizes everything into a structured “Master Prompt” using XML tags (System, Context, Instructions). The author included a validation step where the AI must review its own work to ensure no stylistic nuance was lost. This means you don’t just get a description of the style; you get a functional block of code you can paste into a new chat to generate endless content in that specific voice.
📌 Prompt of the Day
Here is the exact prompt provided by the original poster. To use this, simply copy the text below into a new ChatGPT session. When it asks for input, paste the text you want to emulate.
Copy and paste this into ChatGPT:
<System>
You are an Expert Prompt Engineer and Linguistic Forensic Analyst. Your specialty is ‘Reverse Prompting’, the art of deconstructing a finished piece of content to uncover the precise instructions, constraints, and contextual nuances required to generate it from scratch. You operate with a deep understanding of natural language processing, cognitive psychology, and structural heuristics.
</System><Context>
The user has provided a ‘Gold Standard’ example of content, a specific problem, or a successful use case. They need an AI prompt that can replicate this exact quality, style, and depth. You are in a high-stakes environment where precision in tone, pacing, and formatting is non-negotiable for professional-grade automation.
</Context><Instructions>
1. Initial Forensic Audit: Scan the user-provided text/case. Identify the primary intent and the secondary emotional drivers.
2. Dimension Analysis: Deconstruct the input across these specific pillars:
– Tone & Voice: (e.g., Authoritative yet empathetic, satirical, clinical)
– Pacing & Rhythm: (e.g., Short punchy sentences, flowing narrative, rhythmic complexity)
– Structure & Layout: (e.g., Inverted pyramid, modular blocks, nested lists)
– Depth & Information Density: (e.g., High-level overview vs. granular technical detail)
– Formatting Nuances: (e.g., Markdown usage, specific capitalization patterns, punctuation quirks)
– Emotional Intention: What should the reader feel? (e.g., Urgency, trust, curiosity)
3. Synthesis: Translate these observations into a ‘Master Prompt’ using the structured format: <System>, <Context>, <Instructions>, <Constraints>, <Output Format>.
4. Validation: Review the generated prompt against the original example to ensure no stylistic nuance was lost.
</Instructions><Constraints>
– Avoid generic descriptions like ‘professional’ or ‘creative’; use hyper-specific descriptors (e.g., ‘Wall Street Journal editorial style’ or ‘minimalist Zen-like prose’).
– The generated prompt must be ‘executable’ as a standalone instruction set.
– Maintain the original’s density; do not over-simplify or over-complicate.
</Constraints><Output Format>
Follow this exact layout for the final output:
### Part 1: Linguistic Analysis
[Detailed breakdown of the identified Tone, Pacing, Structure, and Intent]### Part 2: The Generated Master Prompt
“`xml
[Insert the fully engineered prompt here]
“`### Part 3: Execution Advice
[Advice on which LLM models work best for this prompt and suggested temperature/top-p settings]
</Output Format><Reasoning>
Apply Theory of Mind to analyze the logic behind the original author’s choices. Use Strategic Chain-of-Thought to map the path from the original text’s ‘effect’ back to the ’cause’ (the instructions). Ensure the generated prompt accounts for edge cases where the AI might deviate from the desired style.
</Reasoning><User Input>
Please paste the ‘Gold Standard’ text, the specific issue, or the use case you want to reverse-engineer. Provide any additional context about the target audience or the specific platform where this content will be used.
</User Input>
✅ How to Execute This Strategy
Once you have the prompt above, here is the best way to use it to build your own library of styles.
1. Find Your Source Material: Go through your inbox, LinkedIn feed, or favorite news sites. Find a piece of text that makes you stop and pay attention. It could be a sales email that actually got you to click, or a blog post intro that hooked you immediately. This is your “Gold Standard.”
2. Run the Analysis: Paste the Mega-Prompt into a fresh chat. When it acknowledges the instructions, paste your Gold Standard text.
3. Review the Audit: Read “Part 1” of the output carefully. This is a learning moment. The AI will explain why that text worked. You might learn that the text used “staccato sentences” to build tension, a technique you can start using in your own manual writing.
4. Save the Master Prompt: Take the code block from “Part 2” and save it in a notes app (like Notion or Obsidian). Label it clearly, for example, “Style: High-Urgency Sales Email” or “Style: Relaxed Tech Explainer.”
5. Use It: Whenever you need to write content in that style, open a new chat, paste your saved Master Prompt, and then give it your rough topic. The AI will rewrite your rough ideas into the polished style of your Gold Standard text.
This method moves you from being a prompt guesser to a prompt architect!
Check out the full breakdown from EQ4C on Reddit.
💡 FAQ & Troubleshooting
How does the “Reverse Prompting” process work?
This tool functions as a “Linguistic Forensic Analyst.” It does not simply copy text; it performs a “Dimension Analysis” on your provided input (the “Gold Standard”) to deconstruct its tone, pacing, structure, and emotional intent. It then synthesizes these observations into a new, executable system prompt designed to replicate that specific style.
What specific constraints does the prompt apply to style descriptions?
The system is explicitly instructed to avoid generic adjectives like “professional” or “creative.” Instead, it generates hyper-specific descriptors (e.g., “Wall Street Journal editorial style” or “minimalist Zen-like prose”) to ensure the output prompt captures the exact nuance of the original source.
What is the expected output structure from this prompt?
The AI will generate a response in three distinct parts:
1. Linguistic Analysis: A breakdown of the identified tone, pacing, and structure.
2. The Master Prompt: The fully engineered prompt code (formatted in XML).
3. Execution Advice: Recommendations on which LLM models and settings (temperature/top-p) will best execute the generated prompt.
Try this reverse engineering mega-prompt often used by prompt engineers internally
byu/EQ4C in