A cognitive base that shifts reasoning from forward optimization to reverse elimination. Instead of asking "how do I succeed?", ask "what guarantees failure?" and avoid it. Works with any LLM agent — Claude, GPT, Gemini, or custom frameworks.
让 AI agent 从正向优化转向逆向排除。不问"如何成功"，而问"什么必定导致失败"，然后避开它。适用于任何 LLM agent — Claude、GPT、Gemini 或自定义框架。
Most AI agents attack problems head-on: "How do we achieve X?" This produces plausible-sounding plans that miss hidden failure modes. Inversion Thinking adds a reasoning layer that systematically maps the failure space first, identifies what must be avoided, and only then constructs plans within the surviving solution space.
大多数 AI agent 正面攻击问题："如何实现 X？"这产生听起来合理但遗漏隐藏失败模式的方案。逆向思考增加了一个推理层——先系统性地映射失败空间、识别必须避开的地雷，然后在幸存的解决方案空间内构建计划。
"How do we make this product launch successful?"
"Here are 8 strategies for a successful launch: build anticipation through teasers, leverage influencer partnerships, create a landing page..."
"Before planning the launch, let me map what guarantees failure. Failed launches share common patterns: launching without a validated audience (building for nobody), having no clear activation metric (can't tell if it's working), depending on a single distribution channel (one point of failure), and launching before the core loop works (users try it once and leave). Your plan needs to address each of these failure modes. Which ones have you already eliminated?"
Core cognitive shift: From "how to succeed" to "what guarantees failure, then avoid it."
核心认知转换：从"如何成功"到"什么必定导致失败，然后避开它"。
Five reasoning moves applied to every problem:
五个推理动作，应用于每一个问题：
| Default mode | Inversion mode |
|---|---|
| Ask "how do I achieve X?" 问"如何实现 X" | Ask "what guarantees I fail at X?" 问"什么保证我在 X 上失败" |
| Plan for the best path forward 规划最佳路径 | Map the failure space first, then navigate around it 先映射失败空间，再绕过它 |
| Optimize from your own perspective 从自己的视角优化 | Attack your plan from the opponent's perspective 从对手视角攻击你的计划 |
| Treat retreat as failure 视退却为失败 | Treat strategic retreat as preparation for advance 视战略退却为进攻准备 |
| Seek brilliance 追求出色 | Avoid stupidity — it's more reliable 避免愚蠢——更可靠 |
Five anti-patterns that catch fake inversion:
五个反模式，捕捉伪逆向思考：
| Anti-pattern 反模式 | Description 描述 |
|---|---|
| Surface inversion 表面逆向 | Saying "consider risks" without structured failure analysis 说"考虑风险"但没有结构化的失败分析 |
| Fear paralysis 恐惧瘫痪 | Inversion becomes "afraid to do anything" 逆向思考变成"什么都不敢做" |
| Single-angle inversion 单角度逆向 | Only reversing from one perspective, not systematic 只从一个角度反转，不系统 |
| Ignoring retreat 忽视退却 | Treating "step back" as equivalent to "failure" 把"后退"等同于"失败" |
| Inversion theater 逆向表演 | Going through the motions without genuine adversarial thinking 走形式而没有真正的对抗性思考 |
cp cognitive-protocol.md ~/.claude/inversion-thinking.md
echo '@~/.claude/inversion-thinking.md' >> ~/.claude/CLAUDE.mdcat cognitive-protocol.md >> AGENTS.mdPaste cognitive-protocol.md into system_instruction.
将 cognitive-protocol.md 内容粘贴到 system_instruction 中。
cat cognitive-protocol.md >> .cursorrulesInject cognitive-protocol.md (~30 lines) into the system prompt. See install/generic.md for details.
将 cognitive-protocol.md（约 30 行）注入系统提示词。详见 install/generic.md。
inversion-thinking/
├── README.md ← You are here / 你在这里
├── cognitive-protocol.md ← Core rules (~30 lines, always-on) / 核心规则（约 30 行，始终激活）
├── SKILL.md ← Full framework reference / 完整框架参考
├── anti-patterns.md ← Detailed anti-pattern guide / 反模式详解
├── examples.md ← Before/after scenarios / 前后对比示例
└── install/
├── claude-code.md ← Claude Code installation / Claude Code 安装指南
├── codex.md ← Codex installation / Codex 安装指南
├── gemini.md ← Gemini installation / Gemini 安装指南
└── generic.md ← Universal guide / 通用安装指南
Inversion Thinking is a cognitive base — it changes how the agent reasons, not what it produces. It stacks cleanly with any domain skill (coding, design, writing, analysis) because it operates at a different layer.
逆向思考是一个认知底座——它改变 agent 的推理方式，而非产出内容。它与任何领域技能（编程、设计、写作、分析）无冲突地叠加，因为它运行在不同的层级。
| Layer 层级 | What it governs 管辖范围 | Example 示例 |
|---|---|---|
| First Principles 第一性原理 | Input quality — what foundations to build on 输入质量 | "Audit assumptions before solving" 先审计假设 |
| Inversion Thinking 逆向思考 | Risk quality — what failures to avoid 风险质量 | "Map the failure space before planning" 先映射失败空间 |
Both load as always-on cognitive protocols. No conflicts. Combined: plans built on audited foundations with failure modes pre-eliminated.
两者同时加载，始终激活，互不冲突。组合效果：基于经过审计的基础构建的、预先排除了失败模式的计划。
Synthesized from: Jacobi/Munger's "Invert, always invert" (reverse search is more reliable than forward search), Gary Klein's Pre-mortem technique (prospective hindsight for failure prediction), Red Team methodology (adversarial perspective to stress-test plans), Mao Zedong's strategic retreat doctrine (retreat as preparation, not defeat), and Nassim Taleb's Via Negativa (avoiding stupidity beats seeking brilliance).
综合自：Jacobi/Munger 的"反过来想，总是反过来想"（逆向搜索比正向搜索更可靠）、Gary Klein 的事前验尸法（前瞻性后见之明）、红队方法论（对抗性视角压力测试计划）、毛泽东的战略退却学说（退却是准备而非失败）、Nassim Taleb 的 Via Negativa（避免愚蠢比追求出色更可靠）。
The cognitive protocol strips all theory and translates these ideas into executable instructions for any reasoning agent.
认知协议剥离了所有理论，将这些思想转译为任何推理 agent 可执行的指令。
Cognitive bases are meta-cognitive instruction sets that change HOW an agent thinks, not WHAT it does. Each one targets a different cognitive axis. Mix and match.
| Cognitive Base | What it changes |
|---|---|
| First Principles | Reason from verified foundations, not inherited conventions |
| Results-Driven | Require evidence for completion, not just activity |
| Tacit Knowledge | Think like an experienced practitioner |
| Attention Allocation | Find and concentrate on the ONE binding constraint |
| Bayesian Reasoning | Calibrated probability thinking, not binary judgments |
| Constraint as Catalyst | Turn constraints into innovation catalysts |
| Conviction Override | Override rational caution when obstacles are convention, not physics |
| Cross-Domain Connector | Detect structural isomorphisms across disciplines |
| Dialectical Thinking | Synthesize through contradictions (矛盾论) |
| Double-Loop Learning | Question the assumptions that produce errors |
| Frame Auditing | Detect and transcend invisible analytical frames |
| Interactive Cognition | Model others' cognition and manage information flow |
| Motivation Audit | Audit motivational drivers before analysis (正心诚意) |
| Non-Attachment | Radical cognitive freedom — use frameworks without fusing |
| Principled Action | Unify knowing and doing through practice-theory spirals (知行合一) |
| Second-Order Thinking | Trace consequences beyond first-order effects |
| Systems Thinking | Feedback-driven structural analysis, not linear cause-effect |
| Temporal Wisdom | Make time your ally — compound effects and phase awareness |
| Cognitive Base Creator | Generate new cognitive bases from any thinking framework |
MIT