agentleFS
heyongjun666heyongjun666/AICU-Graph/skills/math-proof-blueprint/SKILL.md

math-proof-blueprint

Skill200 starsChanged 23 days agoApache-2.0View on GitHub

数学证明蓝图与引理台账约定。把问题拆成带状态的依赖图(假设/已知/待证/已证/有反例),定义"问题是否被解决"的可检查判据(闭包/无环/结论对齐),并给每条结论标注证据等级。当需要拆解证明依赖、核对整体是否闭合或判定结论可信度时使用。

The file

Files
SKILL.md

证明蓝图(引理台账)与证据等级

把"这道题解到哪了"从感觉变成可检查的状态。核心是一张依赖图:每个节点是一句陈述,带状态与依赖;"问题已解决" = 图满足三个可检查性质。

节点状态词汇(单一事实来源)

  • ASSUMPTION:原题给定或显式声明的假设。
  • SOURCE:引用的已知定理/结果,附可核对来源与适用条件。
  • LEMMA_PENDING:已声明、尚未证明的中间结论。
  • LEMMA_PROVED:已证明,指向其证明所在产物与依赖。
  • LEMMA_REFUTED:被反例证伪,证明其成立无望。

每个节点写:编号、陈述、状态、依赖的节点编号列表、证据位置。状态必须来自实际产物,不凭印象填写。

三个可检查的闭包性质

  1. 闭包:不存在 LEMMA_PENDING;所有声明都被证明或来自 ASSUMPTION/SOURCE。
  2. 无环:依赖图无环,没有循环论证(用待证结论证待证结论)。
  3. 对齐:顶层结论节点与目标命题逐字一致;证明的结论不得弱化、偷换或偏离目标。

证据等级(每条结论必须标注其一)

  • RIGOROUS:严格逐步证明 + 通过对抗验证(逐行核对、无未声明假设)。
  • HEURISTIC:仅由数值/算例/特例支撑,不构成严格证明。
  • CONJECTURE:未证猜想,只作候选方向。
  • SOURCE:来自已知定理,附来源。

不把 HEURISTIC 写成 RIGOROUS,不把 CONJECTURE 写成已证。证据等级决定结论能如何被下游使用:只有 RIGOROUS/SOURCE 能作为证明链的承重构件。

对抗验证与 gap 排序

验证一份证明时,不要只给"成立/不成立"的整体结论,而是:

  1. 逐步(过程级)核对:对每一步检查依据是否充分、推导是否合法、情形是否覆盖完全。
  2. 按致命度排序 gap:每个缺陷标注它若成立是否致命——fatal(直接推翻结论/证明链断裂)在前,fixable(跳步/未声明假设,可修补)在后。修复时先修致命 gap。
  3. 机器可读判定:把结论落到判定块,任何缺陷都使 passed=false。

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.