The Self-Critique Loop — Get a Second Draft Twice as Sharp by Asking Claude to Rewrite Its Own Weakest Parts

One follow-up line that turns any Claude answer into a noticeably sharper second draft

The Self-Critique Loop

By the end of this guide, you'll have a one-line follow-up you can use after any Claude answer that forces a sharper second draft, by making Claude find and rewrite its own three weakest parts.

Who This Is For

  • You get answers from Claude that are fine, but not great, and you don't know how to push for better

  • You accept the first draft because asking for a rewrite feels like more work than fixing it yourself

  • You want a quick, repeatable way to raise output quality without writing a long new prompt

  • You write copy, emails, or plans with AI and want them tighter before you use them

  • You want one habit that makes everything you do with Claude noticeably better

What You'll Build

A single follow-up move you can run after any Claude response:

  • The exact self-critique line that finds and fixes the three weakest parts of any response

  • Targeted variations for persuasion, clarity, or a specific reader when you need a sharper lens

  • A /critique slash command so the whole loop runs in one keystroke

What You'll Need

  • Claude (or any LLM)

  • Any answer you want to improve

  • Nothing else — no tools, no setup, no accounts

Going Further

  • Save this as a permanent command so it's always one keystroke away

  • Have Claude write the prompt before you even start

  • Stack this with four more real Claude techniques for expert-grade output