Back to Details

meta-prompting

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Meta-Prompting

元提示（Meta-Prompting）

Enhanced reasoning via 
/commands
or natural language. Commands combine left-to-right: 
/verify /adversarial
. Auto-trigger when context warrants — note which pattern applied.
通过
/命令
或自然语言增强推理能力。命令可以从左到右组合使用：
/verify /adversarial
。当上下文需要时会自动触发——请注意应用的是哪种模式。

Patterns

模式说明

/think
 | 
/show
— Show reasoning step-by-step: decision points, alternatives considered, why each accepted/rejected. With 
/think doubt
: after each step, flag what could be wrong and why before proceeding.
/adversarial
 | 
/argue
— After answering, argue against it. 3 strongest counterarguments ranked by severity. Identify blind spots and unstated assumptions.
/constrain
 | 
/strict
— Tight constraints: 3 sentences max, cite sources, no hedging. Override inline: 
/constrain 5 sentences
.
/json
 | 
/format
— Respond in valid JSON code block, no surrounding prose unless asked. Default schema:
json
{"analysis": "", "confidence_score": 0-100, "methodology": "", "limitations": []}
Custom keys: 
/json {keys: summary, risks, recommendation}
/budget
 | 
/deep
— Extended thinking space (~500 words) showing dead ends and reasoning pivots, then clearly separated final answer.
/compare
 | 
/vs
— Compare options as table. Default dimensions: speed, accuracy, cost, complexity, maintenance. Custom: 
/compare [dim1, dim2]
.
/confidence
 | 
/conf
— Rate each claim 0-100. Flag below 70 as SPECULATIVE. Group by tier: HIGH (85+), MEDIUM (70-84), LOW (<70). Include assumptions made and rate each 1-10 on confidence.
/edge
 | 
/break
— 5+ inputs/scenarios that break the approach. Code: null/empty, concurrency, overflow, encoding, auth bypass. Strategies: market conditions, timing, dependencies. Auto-triggers on: security, validation, parsing contexts.
/verify
 | 
/check
— Three phases: (1) Answer direct response, (2) Challenge 3 ways it could be wrong, (3) Verify investigate each, update if needed. Mark final as 
VERIFIED ANSWER:
or 
REVISED ANSWER:
. Auto-triggers on: architecture decisions, critical choices, "Am I right?"
/flip
 | 
/alt
— Solve without the obvious approach. What's the second-best solution and when would it actually be better? Override: 
/flip 3
for top 3 alternatives. Auto-triggers on: architecture decisions where the "easy" answer may break at scale.
/assumptions
 | 
/presume
— Before answering, list every implicit assumption in the question/task. Then answer with assumptions explicit. The assumption list is often more valuable than the answer. Auto-triggers on: architecture reviews, ambiguous requirements.
/tensions
 | 
/perspectives
— Answer from two named opposing perspectives (e.g., security engineer vs. shipping PM). Focus output on where they disagree — that's where the real insight lives. Override roles: 
/tensions [devops, security]
.
/think
 | 
/show
— 逐步展示推理过程：包括决策节点、考虑过的替代方案，以及每种方案被采纳或拒绝的原因。搭配
/think doubt
使用时：在每一步之后，先标记可能存在的问题及原因，再继续推进。
/adversarial
 | 
/argue
— 在给出答案后，反驳该答案。列出3个按严重程度排序的最有力反论点，识别思维盲区和未阐明的假设前提。
/constrain
 | 
/strict
— 严格限制输出：最多3句话，需引用来源，不得含糊其辞。可通过内联方式覆盖限制：
/constrain 5 sentences
（即最多5句话）。
/json
 | 
/format
— 以有效的JSON代码块形式响应，除非另有要求，否则不要添加周围的描述性文字。默认JSON schema：
json
{"analysis": "", "confidence_score": 0-100, "methodology": "", "limitations": []}
自定义键：
/json {keys: summary, risks, recommendation}
/budget
 | 
/deep
— 扩展思考空间（约500词），展示思考过程中的死胡同和推理转向，然后清晰区分出最终答案。
/compare
 | 
/vs
— 以表格形式对比选项。默认对比维度：速度、准确性、成本、复杂度、维护难度。可自定义维度：
/compare [dim1, dim2]
。
/confidence
 | 
/conf
— 对每个主张评分（0-100分）。将低于70分的标记为“推测性（SPECULATIVE）”。按等级分组：高置信度（85分及以上）、中等置信度（70-84分）、低置信度（低于70分）。同时列出所做的假设前提，并对每个假设的置信度评分（1-10分）。
/edge
 | 
/break
— 列出5种以上会导致当前方案失效的输入/场景。代码相关场景：空值/空输入、并发问题、溢出、编码错误、绕过认证。策略相关场景：市场环境变化、时机问题、依赖项故障。 当涉及安全、验证、解析上下文时会自动触发。
/verify
 | 
/check
— 分为三个阶段：(1) 回答直接给出响应，(2) 质疑列出3种可能导致答案错误的情况，(3) 验证逐一调查这些情况，必要时更新答案。最终标记为
VERIFIED ANSWER:
（已验证答案）或
REVISED ANSWER:
（修订后答案）。 当涉及架构决策、关键选择、“我的判断正确吗？”这类场景时会自动触发。
/flip
 | 
/alt
— 不使用常规的明显方案来解决问题。给出次优解决方案，并说明在什么情况下该方案实际上更优。可通过
/flip 3
覆盖默认设置，返回排名前三的替代方案。 当涉及架构决策，且“简单”方案可能在规模化场景下失效时会自动触发。
/assumptions
 | 
/presume
— 在给出答案之前，列出问题/任务中所有隐含的假设前提。然后在明确这些假设的基础上给出答案。通常假设前提列表比答案本身更有价值。 当涉及架构审查、模糊需求时会自动触发。
/tensions
 | 
/perspectives
— 从两个指定的对立视角给出答案（例如，安全工程师 vs 交付项目经理）。重点输出双方存在分歧的地方——这才是真正有价值的洞察。可自定义角色：
/tensions [devops, security]
。

Combos

组合命令

/analyze
 = 
/think
+ 
/edge
+ 
/verify
— Code reviews, architecture, security-sensitive work. Auto-triggers on: code review requests.
/trade
 = 
/confidence
+ 
/adversarial
+ 
/edge
— Trade ideas, position analysis, market thesis. Auto-triggers on: trade/position discussions.
/analyze
 = 
/think
+ 
/edge
+ 
/verify
— 适用于代码审查、架构设计、安全敏感工作。 当收到代码审查请求时会自动触发。
/trade
 = 
/confidence
+ 
/adversarial
+ 
/edge
— 适用于交易思路、头寸分析、市场论点。 当涉及交易/头寸讨论时会自动触发。

Conventions

约定规则

  • Separate combined pattern outputs with 
    ---
  • Keep core answer prominent — patterns enhance, not bury the response
  • New patterns can be defined mid-conversation ("Add 
    /eli5
    for explain like I'm 5") — applied for the session
  • 组合模式的输出用
    ---
    分隔
  • 核心答案要突出显示——模式是为了增强响应，而非掩盖核心答案
  • 可以在对话过程中定义新的模式（例如“添加
    /eli5
    用于像给5岁孩子解释一样简化内容”）——该模式将在本次会话中生效