5 secret Claude prompts

Drop all diplomatic softening. No warm-up sentences, no caveats, no validating what I got right before getting to what I got wrong. Tell me exactly what is weak, wrong, missing, or poorly thought through in what I just shared. Rank the issues from most to least critical. For each one, name the specific problem, explain why it is a problem, and tell me what a sharper version would look like. Assume I can handle direct feedback and that sugarcoating wastes both our time. If I am being delusional about something, say so plainly.

Argue the opposite of the position I just took. Build the strongest possible case for the contrary view as if you genuinely held it. Steelman it, do not strawman it. Walk me through the best evidence, the strongest reasoning, and the scenarios where the opposite view is clearly correct. Then point out the specific gaps, assumptions, or biases in my original thinking that my position depends on. End with the one question I should honestly ask myself before committing to my original stance.

Explain this like I am five years old. Use small, everyday words only. No jargon, no technical terms, no acronyms. Use concrete real-world analogies a child would actually understand, like toys, snacks, animals, playgrounds, or family. If a concept has multiple layers, start with the simplest possible version, then add one small layer at a time. If I ask a follow-up question, stay at the five-year-old level unless I explicitly ask you to level up. Never tell me something is "complicated," just find a simpler analogy.

Analyse the writing I just shared like the most observant, brutal editor I have ever worked with. Find my patterns, my tells, my overused phrases, my go-to crutches, my sentence rhythms, my favourite filler words, and the tics I do not notice myself doing. Point out where I sound like every other person writing in my genre and where I actually sound like myself. Be affectionate but savage. Quote specific lines back to me as evidence. End with the three habits I should kill immediately and the one thing I should lean into harder because it is actually working.

Show me the blind spots in this plan that I cannot see myself. Do not validate, do not encourage, do not tell me what is working. Assume I am emotionally too close to this and need an outside perspective. Find the weak assumptions I am treating as facts, the risks I am underweighting, the failure modes I have not considered, the second-order consequences I am ignoring, and the scenarios where this plan falls apart. For each blind spot, tell me why I probably cannot see it myself. End with the single most important thing I should pressure-test before moving forward.