Claude can hold 200,000 tokens of context — roughly 150,000 words — in a single conversation. Most people never use this. This prompt exploits it. Paste everything — your notes, a conversation, a document, a decision you've been wrestling with — and have Claude find the signal buried inside it. The output is always shorter than you expect. That's the point.

Your stated values and your actual values are not the same thing. Your behavior reveals the real hierarchy. This prompt tells Claude to ignore what you say you believe and instead map what you demonstrably believe — based on the choices and patterns you describe. The output is often uncomfortable. It should be.

Before you commit to anything significant — a launch, a hire, an investment, a strategy — run it through this. You're not asking Claude to evaluate it. You're hiring Claude as an adversary. The scored attack vectors are what make this different from generic feedback. Scores force specificity. The final verdict is the one thing you act on first.

Claude is exceptionally good at analyzing language patterns. This prompt turns that against you — in the best way. You describe a situation. Claude reads your word choices, your framing, your emphasis. Then it tells you which cognitive biases are active in your thinking right now, with your own words as evidence. It's like having a mirror that shows you how you're actually thinking, not how you think you're thinking.

Most learning resources are comprehensive by design. That's the problem. Comprehensive optimizes for coverage, not for speed to competence. This prompt does the opposite — it forces Claude to be ruthlessly selective and return a structured table you can act on immediately. The final instruction (what to stop learning) is usually where the most value is.

Charlie Munger's inversion principle — applied as a Claude prompt. Instead of asking how to succeed, you ask Claude to map every way you could guarantee failure. Then you invert the list. The reason this works better than asking for success advice directly: Claude has to get specific about your actual failure modes, not generic best practices. Step 4 is usually the most uncomfortable.

Gary Klein's pre-mortem technique — but done properly. Most people use it as a five-minute brainstorm. This prompt forces Claude to think in narrative, which surfaces different failure modes than a list would. The "real reason" (item 2) is almost always something the person already suspected but hadn't said out loud. The "decisive moment" (item 4) is often surprising.

Behavioral loops persist because they're rewarded. Always. The reward isn't always obvious — sometimes it's identity protection, sometimes it's avoidance of a harder emotion, sometimes it's a dopamine hit from familiar anxiety. This prompt forces Claude to find the specific reward your loop is providing, which is the only way to design a real exit. Without finding the reward, you're just using willpower against a system. That doesn't work.

Most people ask Claude to validate their thinking. This prompt does the opposite — it asks Claude to argue against you as powerfully as possible, then gives you a confidence percentage at the end. The percentage is what makes this actionable. Below 60% means you're committing to something you shouldn't be committing to yet. Above 80% means you have a defensible position. The gap between those numbers is where decisions get made.

Not where you intend to go. Where you're actually going. This prompt asks Claude to model your real trajectory based on your current habits, decisions, and patterns — then show you what closes off at Years 1, 3, and 5. The three-year financial and professional projection is usually where people get uncomfortable. That's the signal that it's working. The pivot question at the end is the only thing that matters after you've read the rest.