CodeWithMith/ultimate-prompt-engineering-pack

🚀 The Ultimate Prompt Engineering Pack (v2.0)

[image]

 

Hey everyone! 👋 Welcome to the Ultimate Prompt Engineering Pack.

I built this because I honestly got tired of seeing the same basic "write me a blog post" prompts being shared everywhere. I wanted to put together a single, open-source collection of truly high-performance prompts—the kind that use actual prompt engineering techniques (like Chain of Thought, ReAct, and Few-Shot) to get the absolute best results out of tools like ChatGPT, Claude, and Midjourney.

Whether you're a developer trying to refactor messy code, a founder writing pitch decks, or just someone having fun with AI art, you'll find something useful here.

📑 Table of Contents

1. 1.What's Inside?
2. 2.Folder Layout
3. 3.How to Use This
4. 4.For the Devs: Training-Ready Datasets
5. 5.Supported Models
6. 6.What's Next?
7. 7.Contributing

🌟 What's Inside?

I've broken the prompts down into a bunch of practical categories:

• —💻 Coding & Dev: Code refactoring, bug hunting, generating unit tests.
• —🧠 Prompt Engineering: Advanced stuff like CoT, ToT, Self-Critique, and ReAct agent loops.
• —📚 Learning: The Feynman Technique and deep-dive conceptual learning.
• —✍️ Content Writing: Viral threads and SEO-optimized blogs.
• —🎨 AI Art: Crazy detailed Midjourney, DALL-E, and Flux templates.
• —🚀 Business & Marketing: SWOT analysis, PAS frameworks, customer personas.
• —🤖 Real-World Use Cases: SaaS support agents and automation logic.
• —💼 Career & Health: Resume optimization and customized macro plans.

Every prompt comes with difficulty levels, the best model to use, and easy copy-paste templates!

📁 Folder Layout

If you're cloning this repo, here's how I set it up:

/prompts
   /coding
   /business
   /marketing
   /ai-art
   /career
   ... (and more)
/datasets
   dataset.json      <-- Grab this for LoRA/Fine-tuning!
   prompts.jsonl     <-- Best for automated pipelines.
/images
   preview.png
README.md
CONTRIBUTING.md
LICENSE

🛠️ How to Use This

Just want to copy/paste?

1. 1.Jump into the prompts/ folder and pick whatever category you're working on.
2. 2.Open the README.md inside that folder.
3. 3.Grab the Prompt text.
4. 4.Swap out the [BRACKETED_PLACEHOLDERS] with your actual details.
5. 5.Drop it into your favorite AI and watch it work.

Example: Prompt: Explain Quantum Computing as if I am a 10-year-old... AI Output: Imagine you have a magical coin. A normal coin is either heads or tails. But this magic coin, while it's spinning in the air, is kind of heads AND tails at the same time...

📊 For the Devs: Training-Ready Datasets

If you're looking to train your own models, I've got you covered. Check out the datasets/ folder. I've formatted all of these prompts and high-quality mock outputs into dataset.json (array format) and prompts.jsonl (JSON Lines format).

They are clean, structured, and completely ready to be dropped into a LoRA fine-tuning pipeline.

🤖 Supported Models

I've tested these prompts pretty heavily across the big players:

• —Text/Code: ChatGPT (GPT-4o), Claude 3.5 Sonnet / Opus, Gemini 1.5 Pro
• —Images: Midjourney v6, DALL·E 3, Stable Diffusion (SDXL), Flux

🔮 What's Next?

I have some big plans for this pack:

• —Scaling up to 500+ premium prompts.
• —Building a custom fine-tuned model specifically for prompt generation.
• —Releasing a Hugging Face Space App to generate these on the fly.
• —Creating an API service.

🤝 Contributing

This is a community effort! If you've got a killer prompt that always works for you, I'd love to see it. Check out CONTRIBUTING.md to see how you can add it to the pack.

📜 License

MIT License - go crazy with it! See the LICENSE file for details.

Created with ❤️ by Mithilesh Kumar | v2.0