{"slug": "find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made", "title": "Find what your prompt is missing: the blindspot pass, the debrief, and the permanent fix, made specific", "summary": "A developer introduced three techniques to improve AI interactions by closing knowledge gaps: a blindspot pass prompt to surface unknown unknowns, a debrief prompt to clarify ambiguous requirements before building, and a persistent setup folder (me.md, projects.md, rules.md) to give Claude context at session start. The methods aim to reduce the need for AI to guess user intent. A starter kit is available on GitHub.", "body_md": "Three ways to close the gaps claude fills with guesses. From the video, made specific so you can run them today. Concept credit: a Claude Code team member's field guide on finding your unknowns.\n\nWhen: before you start anything where you don't know the territory, a new part of a codebase, a skill you've never done (design, video, legal docs), a tool you just installed.\n\nThe prompt, fill the brackets:\n\n```\nI'm about to [what you're doing]. My experience with this: [honest level, e.g.\n\"never touched it\" or \"I know the basics\"]. What I already know: [one or two\nthings]. Do a blindspot pass: help me find my unknown unknowns here, the\nquestions I don't know to ask, the mistakes people make, what good actually\nlooks like. Then tell me how to prompt you better for this task.\n```\n\nFilled example:\n\n```\nI'm about to add authentication to my app. My experience with this: I've only\nfollowed tutorials. What I already know: I want email login and Google login.\nDo a blindspot pass: help me find my unknown unknowns here, the questions I\ndon't know to ask, the mistakes people make, what good actually looks like.\nThen tell me how to prompt you better for this task.\n```\n\nWhy the context matters: the pass is only as good as what claude knows about YOUR starting point. Telling it your level changes which blindspots it surfaces.\n\nWhen: right before claude builds anything real, a feature, a document, a design. Run it in the same session, then let it build.\n\nThe prompt, fill the brackets:\n\n```\nBefore you build [the thing], debrief me first. Ask me questions one at a\ntime about anything ambiguous, prioritize the ones where my answer would\nchange the approach. Keep going until you have no important questions left,\nthen summarize everything I told you as the spec, and build from that spec.\n```\n\nFilled example:\n\n```\nBefore you build the money map dashboard, debrief me first. Ask me questions\none at a time about anything ambiguous, prioritize the ones where my answer\nwould change the approach. Keep going until you have no important questions\nleft, then summarize everything I told you as the spec, and build from that\nspec.\n```\n\nThe two upgrades over just chatting: it ends (no endless questions), and the answers become a written spec claude builds from, so nothing you said gets lost.\n\nNot a prompt, a setup. Create a folder claude reads at the start of every session, so it stops guessing at who you are:\n\n- me.md — who you are, what you do, how you like to work, your voice\n- projects.md — what you're building right now, current state, what's next\n- rules.md — your hard preferences: stack, style, things claude should never do\n\nPoint claude at the folder at session start (or use a CLAUDE.md that loads automatically in Claude Code). Every answer you'd give in a debrief that will STILL be true next month belongs in these files.\n\nWant a working version instead of building from scratch? I published a starter\nyou can fork and make yours:\n[https://github.com/kju4q/harness-starter](https://github.com/kju4q/harness-starter)\n\nThe models keep getting better. The gap is you now. Close it once.\n\nBuilt by @kju4q, more at github.com/kju4q", "url": "https://wpnews.pro/news/find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made", "canonical_source": "https://gist.github.com/kju4q/b0d1c2b83e4bc7623e0513ff9e60a466", "published_at": "2026-07-09 20:48:53+00:00", "updated_at": "2026-07-10 07:40:17.718877+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-tools", "developer-tools"], "entities": ["Claude Code", "kju4q", "GitHub"], "alternates": {"html": "https://wpnews.pro/news/find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made", "markdown": "https://wpnews.pro/news/find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made.md", "text": "https://wpnews.pro/news/find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made.txt", "jsonld": "https://wpnews.pro/news/find-what-your-prompt-is-missing-the-blindspot-pass-the-debrief-and-the-fix-made.jsonld"}}