# The AI Trigger Code Cheat Sheet — CodexLab > URL: https://codexlab.io/resources/claude-trigger-codes > Source: https://codexlab.io/llms/resources-claude-trigger-codes.txt > Last updated: 2026-08-25 20 one-word trigger codes that change how any AI answers. Free resource. Nothing to download — the whole cheat sheet lives on the page. Works in Claude, ChatGPT, Gemini and Copilot. WHAT A TRIGGER CODE IS A trigger code is one word that stands in for a paragraph of instructions. There is no hidden menu or developer mode inside these tools. You define a word at the top of a chat ("when I type STRAIGHT, give me your most honest answer with no softening") and from then on you type the word instead of re-writing the instruction. Because a code is just a plain-English instruction, all 20 work in Claude, ChatGPT, Gemini and Copilot. THIS IS THE SMALLEST SYSTEM YOU'LL EVER BUILD Every real AI system does the same thing at a bigger scale: define how the work should be done once, store it, call it by name when needed. That is what an agent instruction is, and what a workflow is. A trigger code is that shrunk to ten seconds with no tools. If you type the same instruction three times a week, that is a system waiting to be built. SOME OF THESE WORDS ALREADY MEAN SOMETHING REDTEAM, PREMORTEM, STEELMAN, 80/20 and the Feynman technique are established methods with decades behind them, so models recognise them without a definition. The six codes marked "one of ours" are CodexLab labels and mean nothing until defined. Paste the starter line either way. WHAT CHANGES BETWEEN TOOLS - Bluntness varies. STRAIGHT works everywhere, but enterprise Copilot is the politest of the four. - Definitions drift. Every model slides back to its default tone over a long conversation; the starter lines include an instruction to keep going until you say stop. - Persistent settings differ. Each tool stores permanent instructions in a different place (see below). HOW TO SET ONE UP IN TEN SECONDS 1. Open a fresh chat — codes apply inside the conversation you paste them into. 2. Paste the starter line at the top. It holds for the rest of the chat. 3. Type the code at the end of your real message whenever you want that behaviour. THE 20 CODES, GROUPED BY JOB WORK OUT WHAT TO DO NEXT DIRECTOR (one of ours) What it does: Splits a task three ways — what an AI can run unsupervised, what needs you in the loop, and what shouldn't be automated at all. Use it when: you're about to hand something to AI and haven't thought about where the line is. Starter line: unlocks free with name and email on the page. 80/20 What it does: Names the single highest-leverage thing to do first — the 20% of effort behind 80% of the result. Use it when: you're staring at a to-do list and don't know where to start. Starter line (free on the page): When I type 80/20, identify the single highest-leverage thing I should do first, the 20% of effort that will drive 80% of the result, and tell me to start there. Don't give me a long list. MARGIN (one of ours) What it does: Forces the commercial answer first — cost to deliver, what it earns, where the margin sits. Use it when: something sounds exciting and you haven't done the sums yet. Starter line: unlocks free with name and email on the page. FIRSTPRINCIPLES What it does: Ignores how things are normally done and rebuilds the problem from what's actually true. Use it when: the standard approach feels wrong and you can't articulate why. Starter line: unlocks free with name and email on the page. LINDYMODE What it does: Only advice that has held up for decades. Filters out the hype layer. Use it when: you're making a decision you don't want to redo in six months. Starter line: unlocks free with name and email on the page. YCOMBINATOR What it does: Evaluates an idea like a startup investor — customer, problem, why now, market size, moat. Use it when: you want a sharp first gut-check on a new idea. Starter line: unlocks free with name and email on the page. PRESSURE-TEST A DECISION STRAIGHT What it does: No flattery, no hedging, no "great question". Just the answer, including the part you were hoping to avoid. Use it when: you suspect everyone around you is being too polite about your idea, offer or copy. Starter line (free on the page): From now on, when I type STRAIGHT, give me your most honest answer with no softening. No flattery, no caveats, no "great question". Tell me the thing I don't want to hear. REDTEAM What it does: Turns the AI into your sharpest critic and hunts for holes, risks and weak spots. Use it when: you're about to launch, sign or commit, and you want the flaws found now. Starter line: unlocks free with name and email on the page. PREMORTEM What it does: Pretends the plan has already crashed six months from now, then works backwards. Use it when: you've committed to a plan and want the risks on the table. Starter line: unlocks free with name and email on the page. STEELMAN What it does: Builds the strongest possible case for the view you disagree with. Use it when: you need to understand the other side before a negotiation or hard conversation. Starter line: unlocks free with name and email on the page. PRODUCTION (one of ours) What it does: Assumes it works in a demo, then tells you what breaks at volume, on messy inputs, and when the builder leaves. Use it when: something worked beautifully once and you're about to make it a process. Starter line: unlocks free with name and email on the page. CHECK IT BEFORE IT SHIPS SOURCED (one of ours) What it does: Labels every claim as known, inferred or estimated, instead of blending all three in one confident tone. Use it when: you're about to act on numbers, dates or facts an AI gave you. Starter line: unlocks free with name and email on the page. HUMANCHECK (one of ours) What it does: Names which specific parts need a human to verify, and what goes wrong if nobody does. Use it when: anything is about to reach a client, a customer or a public feed. Starter line: unlocks free with name and email on the page. PLAINENGLISH (one of ours) What it does: Rewrites something technical for a board, client or sceptical CFO in one read. Use it when: you understand it and now you have to present it. Starter line: unlocks free with name and email on the page. UNDERSTAND SOMETHING HARD ELI10 What it does: Explains anything complicated like you're ten years old. Plain words, everyday analogies, zero jargon. Use it when: you've read something three times and still don't get it. Starter line (free on the page): When I type ELI10, explain the topic like I'm 10 years old. Simple words, everyday analogies, no jargon. Assume I know nothing about it. FEYNMAN What it does: Makes you explain the concept back, then catches the bits you got wrong or fuzzy. Use it when: you're about to teach, present or defend something. Starter line: unlocks free with name and email on the page. SOCRATES What it does: Asks you questions, one at a time, until you reason your way to the answer yourself. Use it when: you want to think more clearly rather than be handed a conclusion. Starter line: unlocks free with name and email on the page. RUBBERDUCK What it does: Mostly listens while you talk a problem through, reflecting back and asking the odd question. Use it when: you need to hear yourself say it out loud. Starter line: unlocks free with name and email on the page. GET WORDS ON THE PAGE ROUGH What it does: A deliberately unpolished first draft. Something to react to instead of a blank page. Use it when: you've been staring at an empty document for twenty minutes. Starter line: unlocks free with name and email on the page. VOICE What it does: Studies samples of your real writing and rewrites the piece to sound like you. Use it when: you're sick of editing the AI tone out of everything. Starter line: unlocks free with name and email on the page. WHAT UNLOCKS WITH AN EMAIL The full index above is public. Entering a name and email on the page reveals the remaining 17 copy-paste starter lines, four stacking combinations (STRAIGHT + REDTEAM, ELI10 + SOCRATES, 80/20 + MARGIN, PREMORTEM + PRODUCTION), a table of where to store codes permanently in Claude, ChatGPT, Gemini and Copilot, and the method for writing your own codes. WHERE THIS STOPS WORKING Codes change how an AI answers a question you asked. They don't do the work while you're asleep, read your inbox, or chase an invoice. The moment you want that, you've moved from prompting to building: permissions, escalation, accountability when it gets something wrong. Start with codes anyway — set up four and use them for a fortnight, and you'll get a feel for how much of the result comes down to how precisely you define the work. That's the skill the systems run on. WHY CODEXLAB PUBLISHED THIS Trigger codes are day one of AI fluency: cheap, immediate, no setup. The bigger shift happens when teams stop typing one-off prompts and start building systems. CodexLab trains teams on their real work and builds the AI capability that keeps running afterwards.