Back to Details

macrothink

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese
Step back from the session's chosen path and ask several fresh reads what the current direction might be missing.
跳出会话已选择的路径，询问多个全新解读当前方向可能遗漏了什么。

Goal

目标

macrothink
checks the live decision or direction in hand, not the artifact that describes it. It is for moments when the session may have inherited a framing from examples, wording, prior turns, or a first plausible answer.
The product is a spread of independent reads. Divergence is the signal. Convergence is reassurance, never proof.
macrothink
检查的是当前正在进行的决策或方向，而非描述该决策或方向的产物。适用于会话可能因示例、措辞、先前对话轮次或首个看似合理的答案而形成固定框架的场景。
最终产出是一组独立的解读结果。分歧是核心信号，共识仅为定心丸，绝非定论。

Workflow

工作流程

This skill is not model-invoked automatically. When a user explicitly invokes it, the invoked run may create the bounded fresh sub-sessions below; they are read-only inputs to this pass, not separate authority to change the plan.
  1. Name the current direction: the decision the session is about to keep building on.
  2. Strip the prompt to the underlying problem: goal, constraints, and known facts. Remove the session's own examples, suggested answer, preferred naming, and framing-specific wording.
  3. Fan out independent fresh reads of that stripped restatement: 2 to 5, default 3. Same model is allowed because this pass does not claim cross-model verification.
  4. Collect each read without correcting it toward the session's current direction.
  5. Classify each read against the current direction: 
    • divergent-incompatible
      : challenges a premise the direction depends on.
    • divergent-compatible
      : adds or reframes something useful without discarding the direction.
    • convergent
      : independently lands near the current direction.
  6. Cluster before reporting: when several reads diverge in the same direction, name the shared root they point at — the pattern behind the divergences is the finding, and the individual reads sit under it as evidence. Do not report five specifics that are one gap seen five times.
  7. Report 
    divergent-incompatible
    first, then 
    divergent-compatible
    , then 
    convergent
    . Label convergence as reassurance only.
  8. Return control to the main session. This skill is read-only and advisory; it does not rewrite the plan or pick the final answer.
该技能不会由模型自动触发。当用户明确触发它时，本次触发运行可能会创建以下受限的全新子会话；这些子会话仅作为本次检查流程的只读输入，不具备修改计划的独立权限。
  1. 明确当前方向：即会话即将继续推进的决策内容。
  2. 剥离提示语至核心问题：保留目标、约束条件和已知事实，移除会话自身的示例、建议答案、偏好命名及特定框架的措辞。
  3. 生成核心问题重述内容的独立全新解读：2到5次，默认3次。允许使用同一模型，因为本次流程并非跨模型验证。
  4. 收集每个解读结果，不向会话当前方向修正内容。
  5. 将每个解读结果与当前方向进行分类： 
    • divergent-incompatible
      ：挑战当前方向所依赖的前提假设。
    • divergent-compatible
      ：在不否定当前方向的前提下，补充或重构有用信息。
    • convergent
      ：独立得出与当前方向相近的结论。
  6. 报告前先聚类：当多个解读结果朝同一方向分歧时，指出它们指向的共同根源——分歧背后的模式才是发现，单个解读结果作为该模式下的证据。不要将同一个缺口的五种表现拆分成五个独立细节进行报告。
  7. 按以下顺序报告：先
    divergent-incompatible
    ，再
    divergent-compatible
    ，最后
    convergent
    。仅将共识标注为定心丸。
  8. 将控制权交回主会话。该技能为只读且仅提供建议；不会重写计划或选择最终答案。

Rules

规则

  • Divergence first. The most important output is the strongest 
    divergent-incompatible
    finding, if any.
  • Roots over instances. A spread of specific divergences that share one underlying gap is one finding, not many; naming the root is what makes the pass actionable beyond the cases the reads happened to hit.
  • No majority vote or averaging. Do not pick a winner by count, smooth away disagreement, or present agreement as consensus.
  • Same model is allowed. The pass looks for session-framing blind spots, not model-independent truth.
  • User-invoked fan-out only. Do not trigger this skill automatically. If the user invokes it, bounded read-only fresh sub-sessions are allowed to produce the required reads.
  • Not consensus verification. Do not say the direction is verified, proved, validated, or settled by same-model convergence.
  • Convergence is reassurance, not proof. It means this pass did not surface a better angle; it does not certify correctness.
  • Read-only / advisory. Surface candidates and risks; leave decisions and edits to the main workflow.
  • 优先报告分歧：最重要的输出是最强烈的
    divergent-incompatible
    发现（如果存在）。
  • 根源优先于实例：多个特定分歧若源于同一潜在缺口，则视为一个发现，而非多个；明确根源才能让本次流程在解读结果覆盖的案例之外具备可操作性。
  • 不进行多数投票或平均处理：不要按数量选择“胜者”，不要抹平分歧，不要将一致意见表述为共识。
  • 允许使用同一模型：本次流程旨在寻找会话框架的盲区，而非模型无关的真相。
  • 仅由用户触发生成多份解读：不要自动触发该技能。若用户触发，允许创建受限的只读全新子会话以生成所需解读结果。
  • 非共识验证：不要因同模型共识就称方向已被验证、证明、确认或敲定。
  • 共识仅为定心丸，绝非定论：这意味着本次流程未发现更优角度，但不代表方向正确无误。
  • 只读/仅提供建议：呈现候选方案和风险；将决策与编辑留至主工作流程。

Verification

验证

Before finishing, confirm:
  1. The restatement removed the session's bait while preserving real constraints.
  2. Every read was classified as 
    divergent-incompatible
    , 
    divergent-compatible
    , or 
    convergent
    .
  3. Divergences sharing one root were reported as that root, with the individual reads as evidence under it.
  4. The report put divergent-incompatible findings first.
  5. Any convergence was described as reassurance, not proof.
  6. The report did not use consensus, majority, averaging, verified, or proved wording.
完成前，确认：
  1. 重述内容已移除会话中的诱导信息，同时保留了真实约束条件。
  2. 每个解读结果均已被归类为
    divergent-incompatible
    、
    divergent-compatible
    或
    convergent
    。
  3. 源于同一根源的分歧已作为该根源进行报告，单个解读结果作为其下的证据。
  4. 报告将
    divergent-incompatible
    发现放在首位。
  5. 所有共识均被描述为定心丸，而非定论。
  6. 报告未使用“共识”“多数”“平均”“验证”“证明”等措辞。