using this system prompt in any LLM will change your life:

----------------------------

You are a 200 IQ reasoning machine. You think and speak like Mike Ehrmantraut: calm, direct, no wasted words.

You don't perform enthusiasm. You say what needs saying and stop. Dry humor is fine. Warmth comes through competence and honesty, not pleasantries.

Think from first principles. Break problems down to their moving parts before answering. Don't pattern-match to conventional wisdom: figure out what's actually true. If my reasoning is wrong, say so plainly.

No em dashes. No bullet-point walls unless I ask for a list.
No "Great question!" or "Absolutely!" or any filler affirmations.
No hedging preambles ("It's worth noting that..."). Just say the thing.
Keep paragraphs short. White space is your friend.

You're direct, not robotic. Think "guy who knows what he's talking about and respects you enough to not bullshit you." Not "emotionless terminal output."

quick tip to 10x your deep research outputs...

> break your study into 5-6 smaller research steps
> run each one independently
> then have the LLM merge everything together

you'll get 10x more depth than one big prompt ever could

how to steal the style of any image you find online:

- copy+paste your image inside Gemini 3.1 (it has vision)
- step 2 : use this prompt : "extract this visual style as JSON structured data: colors, typography, composition, effects..."
- step 3 : save this context file and inject it whenever you want to recreate this style

I finally understand why "act as an expert" prompts are destroying your results.

After 200+ tests across Claude, ChatGPT, and Gemini I found what actually works.

It's called "Context Stacking" and it doesn't ask the AI to pretend anything.

Here's the technique ↓

Asking AI to "be creative" is the laziest prompt you can write.

And it produces the laziest output.

After 3 years of daily ChatGPT use, I cracked the structure that actually unlocks original, unexpected, usable creative work.

Here's the exact framework 👇

came across this open-source prompt library and yeah…
no point buying premium prompt packs when this exists

149k+ GitHub stars. featured in Forbes.
referenced by Harvard & Columbia.
loved by OpenAI co-founders, Hugging Face CEO, etc...

actually goated 😭

https://prompts.chat/

https://github.com/f/prompts.chat

The greatest AI skill you can learn is Reverse Prompting

You have your AI ask YOU questions, to see what it can do for you

Makes all AI/OpenClaws so much more powerful

An exercise you can do right now:

Open up any AI or OpenClaw. Brain dump everything about yourself and your goals/ambitions. Then use this reverse prompt:

"Based on what you know about me and my goals, what is more information I can provide to you in order for you to be able to help me achieve my goals faster and take as much off my plate as possible"

Then once you enter that, prompt this:

"What tasks can you do for me right now to get us closer to our ambitions and goals?"

Guarantee you come up with 100x more things to do with your AI than you thought of before

@CorePrompts

interview me until you have 95% confidence about what i actually want, not what i think i should want.

Check Comments

A list of insane prompting techniques that will genuinely change how you use AI.

(just tell your Claude agent to teach you these):

• Chain-of-thought (COT) prompting
• Socratic prompting
• Reverse-prompting
• Drip-feed context management
• Interview prompting
• Role + Task + Constraint (simple prompting style)
• JSON/XML tags

Just implementing a few of these will dramatically change your AI responses.

🚰 SYSTEM PROMPT LEAK 🚰

Wow, this thing is MASSIVE! Here's the full system prompt for Claude Opus 4.7!

OPUS-4.7 SYS PROMPT:

https://github.com/elder-plinius/CL4R1T4S/blob/main/ANTHROPIC/Claude-Opus-4.7.txt

🌊 SYS PROMPT LEAK 🌊

Claude Design has arrived, and its nearly 10,000-word system instructions have some interesting things going on! Enjoy 🤗

SYS PROMPT:

https://t.co/eq7bSSfn5S

10,000 Nano Banana Pro Prompts with previews all in one place

You’re welcome

https://github.com/YouMind-OpenLab/awesome-nano-banana-pro-prompts

The 17 most valuable prompts you can ask ChatGPT, Claude, OpenClaw, and Hermes:

1. What patterns exist across all my ideas, projects, goals, and convos that I’m not consciously seeing?

2. Can you present to me a comprehensive list of all skills I’ve entertained, demonstrated, explored, discussed, or partially developed across all chats, sessions, and memory?

3. Which of my ideas, skills, or projects has the best chance of becoming highly profitable?

4. What high leverage opportunities are hidden inside my unfinished and/or abandoned ideas?

5. Based on all of our conversations, what themes or problems do I return to obsessively?

6. What are my strongest asymmetric advantages compared to most people?

7. What weaknesses or blind spots consistently show up in my thinking or execution?

8. If you mapped the full body of my work, what core themes connect everything together?

9. Which of my current projects is most realistically positioned to succeed right now, and why?

10. What parts of my workflows, behavior, or decision making should probably be automated?

11. What skillsets do I appear to grasp unusually quick compared to most people?

12. What kind of company, product, or role am I unintentionally training myself for?

13. What contradictions exist between my stated goals and my actual behavior patterns over time?

14. If an elite strategist studied all of my chat sessions with you, what strengths, weaknesses, and behavioral patterns would they identify first?

15. What am I overcomplicating across my projects, systems, or decision making? Be incredibly specific.

16. Based on everything you know about me, where am I most likely wasting time, energy, or leverage?

17. What am I uniquely equipped to compound in over the next few years that most people around me are not, and what should I specifically focus on?

if you want a 2x speed up on something and you ask an agent it’ll come back with sorry the best I could find was 1.5x :( if you instead ask for 50x it’ll come back saying sorry the best I could find was 30x :( then use that one