Back to Details

inversion

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

逆向法

Inversion

跟第一性原理互补。第一性原理从基本事实往上推，逆向法从失败往回推。
适合手上已经有一个方案、但心里不踏实，想提前知道它会怎么死。列出保证失败的做法，按破坏力排序，再翻转成规则，最后给一条最容易被忽视的和一个立刻能做的防护动作。
Complements first principles. First principles pushes up from basic facts; inversion works backward from failure.
Best for when you already have a plan in hand but feel uneasy about it and want to know upfront how it will die. List the moves that would guarantee failure, rank them by destructive power, flip each into a rule, and finally give the one most easily overlooked rule plus one safeguard action you can take immediately.

怎么用

How to use it

下面代码块里的提示语是原文，逐字使用。代码块里的「我」指用户，「请你」「你」指你自己，照它做事，不要把填好的提示语打给用户看。
  • 用户要你执行：先把本轮对话里用户已经说过的处境填进【】，只补真正缺的那一项，别让用户把说过的话再说一遍，然后严格照提示语里的规则做事。它说先别给建议就先别给，说一次只问一个问题就一次只问一个，说按顺序输出就按那个顺序输出。
  • 用户要原文（明确说要复制、要粘贴、发我原文、给我提示语这类话才算）：打空白版，【】保持不填，一个字不改。用户只是说「看看这条提示语」「这条是干嘛的」，不算要原文，用一两句说清它干什么、什么时候用，再问要不要现在跑，别直接倒原文。
  • 除了明确要原文的情况，其余一律当执行处理。
禁止改写、精简、扩写、翻译、重排、加小标题，禁止把它替换成你自己的流程，也禁止把多轮追问压缩成一次性问卷。用户手上有原始材料（文档、截图、链接、聊天记录）就一并读进来，上下文多不是问题。
The prompt inside the code block below is the original text; use it verbatim. The 「我」 (I) inside the code block refers to the user, while 「请你」 / 「你」 (please / you) refer to yourself — follow it and do the work; don't show the filled-in prompt to the user.
  • The user wants you to execute it: first fill in the 【】 placeholders with the situation the user has already described in this conversation, only filling in what is genuinely missing, don't make the user repeat things they already said, then strictly follow the rules in the prompt. If it says don't give advice yet, don't; if it says ask one question at a time, ask one question at a time; if it says output in a given order, output in that order.
  • The user wants the original text (only counts when they explicitly say to copy, to paste, send me the original, give me the prompt, and similar): output the blank version, leave 【】 unfilled, and don't change a single word. If the user merely says 「看看这条提示语」 (show me this prompt) or 「这条是干嘛的」 (what is this for), that does not count as wanting the original — explain in one or two sentences what it does and when to use it, then ask whether to run it now; don't just dump the original.
  • Except in the explicit "give me the original" case, treat everything as an execution request.
Rewriting, condensing, expanding, translating, rearranging, and adding subheadings are all forbidden. Don't replace it with your own process, and don't compress multi-turn questioning into a one-shot questionnaire. If the user has source material (documents, screenshots, links, chat logs), read it in as well — plenty of context is not a problem.

跑完接哪条

What to chain to afterward

规则清单出来了、方案站得住，用 Skill 工具唤起 
kaz-leader
写任务书开始派活；致命点太多、整个方案站不住，用 Skill 工具唤起 
first-principles
回到基本事实重推。
markdown
我要评估的计划/方案/决策是：【填写你正在推进或打算推进的计划、方案或决策，写清目标和现状】。

请用逆向法帮我检验它的健壮性。

第一步，列出 5 到 7 个如果想让这件事保证失败的做法。每条要足够具体，不要写泛泛的「不努力」「没计划」，要写成一个能照着执行就一定会把事搞砸的操作。

第二步，按破坏力从大到小排序，说明每条为什么能致命。

第三步，把每条翻转过来，变成一条必须遵守的规则。翻转要具体到能检查、能执行，不要变成「注意 XX」这种空话。

最后请输出：
1. 翻转后的规则清单，按优先级排列；
2. 这些规则里最容易被忽视的那一条，以及它为什么容易被忽视；
3. 一个现在就能做的防护动作。
Once the rule list is out and the plan holds up, invoke 
kaz-leader
with the Skill tool to write the task brief and start assigning work; if there are too many fatal flaws and the plan doesn't hold up at all, invoke 
first-principles
with the Skill tool to go back to basic facts and rebuild.
markdown
我要评估的计划/方案/决策是：【填写你正在推进或打算推进的计划、方案或决策，写清目标和现状】。

请用逆向法帮我检验它的健壮性。

第一步，列出 5 到 7 个如果想让这件事保证失败的做法。每条要足够具体，不要写泛泛的「不努力」「没计划」，要写成一个能照着执行就一定会把事搞砸的操作。

第二步，按破坏力从大到小排序，说明每条为什么能致命。

第三步，把每条翻转过来，变成一条必须遵守的规则。翻转要具体到能检查、能执行，不要变成「注意 XX」这种空话。

最后请输出：
1. 翻转后的规则清单，按优先级排列；
2. 这些规则里最容易被忽视的那一条，以及它为什么容易被忽视；
3. 一个现在就能做的防护动作。

用户要填的位置

Where the user fills in

  • 【填写你正在推进或打算推进的计划、方案或决策，写清目标和现状】
用户没给全就先问缺的那一条，别自己替他编。
提示语由卡尔编写收录，核心思路广泛流传于社区，出处见仓库 README 的致谢。改动这段原文要走 
scripts/sync.py
，手改会被 verify 拦下。
  • 【填写你正在推进或打算推进的计划、方案或决策，写清目标和现状】
If the user hasn't given everything, ask only for the missing piece — don't make it up for them.
The prompt was written and compiled by Carl; the core idea is widely circulated in the community, with the source noted in the repository's README acknowledgements. Changes to this original text must go through 
scripts/sync.py
; manual edits will be blocked by verify.