Three copy-paste lines for ChatGPT, Claude, or Gemini. The AI interviews you first, fact-checks itself, and regrades its own work before you see it.
When AI guesses what you meant, you get the average answer it gives everyone. These 3 lines stop the guessing. Each one makes the AI do a round of work before you see anything: ask you what's missing, check real sources, or grade its own draft and redo it.
They work in any AI. Add one line to the prompt you were already going to write.
Use on everything. The questions alone show you what your prompt was missing, and you end up giving it things you never would have thought to type in.
Before you answer, ask me three questions that would make your answer better.
Use when being wrong costs you money or trust: your industry, your health, anything you'll publish. It pulls from real sources and double checks them instead of answering from memory.
Do online research first, and then fact-check what you find.
Use when the first draft matters: a page, an email, a plan. The AI scores its own work and keeps fixing it, so the version you see is the second or third pass, and it comes back sharper.
Grade your output on the five most relevant criteria, give a critique on why anything scored under ten out of ten, and then redo the work until everything scores a ten.
Want your answers to sound like you too? Pair these with the Humanizer skill.