The prompt vault I actually use (and that built all of my projects).
This is not another “1,000+ ChatGPT prompts” dump, it's my personal, battle-tested prompt collection.
I refined over hundreds of real-life tests, experiments, and actual shipped products.
Every prompt here has been:
🧪 Tested in real projects — from AI product design to automation flows.
🎯 Optimized for clarity — so you get useful, relevant output without guesswork.
🛠 Built to be remixed — plug them into ChatGPT, Claude, Replit Agents and other tools.
/prompts
├── karo_giga_prd_generator.md # Full PRD builder, 188 lines
├── implementation_plan_for_prd.md # Implementation plan builder, 228 lines
├── prompt_builder.md # Create new prompts from scratch with my system
├── prompt_evaluator.md # Score & refine prompts for clarity, depth, and accuracy
├── replit_anti_loop_regression.md # Diagnostic prompt to prevent looping & regressions
├── replit_founders_checker.md # Founder's-level checklist for production readiness
Browse → Open any .md file to explore prompt examples.
Copy → Paste into your AI tool of choice.
Customize → Swap placeholders for your project context.
Run & Iterate → Try a prompt and make it your own if needed.
Advanced context engineering, testing & evaluation → Each prompt is stress-tested in real scenarios, with structured input/output checks to ensure it performs reliably across tools, topics, and edge cases.
Context-first → Gives the AI enough background to sound smart.
Step-by-step reasoning → Forces depth instead of fluff.
Constraints → Narrow scope = higher quality output.
Tone control → The AI matches your style, not the other way around.