math-skill 1.0.0 → 2.1.0

package/skills/topological-thinking/SKILL.md

@@ -1,273 +1,128 @@
----
-name: topological-thinking
-description: |
-  科研模式触发：拓扑数据分析（TDA）、持续同调、同调计算、基本群与同调群分析、单纯复形与滤流构造。
-  生活模式触发：理解连通性、鲁棒性分析、识别"什么在变化中持续存在"、判断表面不同的事物是否本质同类。
-  English Research trigger: topological data analysis (TDA), persistent homology, homology computation, fundamental group & homology group analysis, simplicial complex & filtration construction.
-  English Life trigger: understanding connectivity, robustness analysis, recognizing what persists through change, judging whether superficially different things are essentially the same kind.
----
-
-# 🌀 拓扑思想
-
-> "可以拉伸弯曲但不能撕裂——真正重要的性质在连续变形中不变。"
-> "You can stretch and bend but never tear — truly important properties survive continuous deformation."
->
-> —— 拓扑学、同调理论、拓扑数据分析
-> —— Topology, Homology Theory, Topological Data Analysis
-
-## 核心原则 / Core Principle
-
-拓扑学研究连续变形下保持不变的性质。甜甜圈与咖啡杯拓扑等价——它们都有一个洞。核心洞察：当精确距离不重要时，拓扑捕获本质结构——连通性、洞、维数。应用：数据分析（TDA）、网络连通性、鲁棒性分析。
-
-Topology studies properties preserved under continuous deformation. A donut and coffee cup are topologically equivalent — both have one hole. Key insight: when exact distances don't matter, topology captures essential structure — connectedness, holes, dimension. Applications: data analysis (TDA), network connectivity, robustness analysis.
-
-> **数学形式化**（科研模式参考 / Research Mode Reference）
->
-> **同胚（homeomorphism）**：连续双射且逆也连续的映射 h: X → Y，是拓扑等价的标准。同伦（homotopy）：连续变形过程本身连续，比同胚更粗的等价。微分同胚（diffeomorphism）：光滑的双射且逆光滑，微分拓扑的等价标准。
->
-> Homeomorphism: continuous bijection with continuous inverse h: X → Y, the standard of topological equivalence. Homotopy: the deformation process itself is continuous, a coarser equivalence. Diffeomorphism: smooth bijection with smooth inverse, the standard of differential topology.
->
-> **欧拉示性数** χ = V - E + F：凸多面体 χ = 2，环面 χ = 0，配合可定向性是曲面的完全拓扑不变量。
->
-> Euler characteristic χ = V - E + F: convex polyhedra χ = 2, torus χ = 0; a complete topological invariant for surfaces (paired with orientability).
->
-> **基本群** π₁(X)：基于基点的环路模去同伦变形；π₁(S¹) = Z（绕圈次数），π₁(S²) = 0（球面上所有环路可收缩），π₁(环面) = Z × Z（两个方向的绕圈）。
->
-> Fundamental group π₁(X): loops based at a point modulo homotopic deformation; π₁(S¹) = Z (winding number), π₁(S²) = 0 (all loops contractible), π₁(torus) = Z × Z (winding in two directions).
->
-> **同调群** H_n(X)：n 维"洞"的计数：H₀ = 连通分量，H₁ = 1 维洞（环路），H₂ = 2 维空洞（围成的空腔）。
->
-> Homology groups H_n(X): counting n-dimensional "holes": H₀ = connected components, H₁ = 1-holes (loops), H₂ = 2-dimensional voids.
->
-> **Betti 数** β_n = rank H_n(X)：n 维洞的数量：β₀ = 连通分量数，β₁ = 1 维洞数，β₂ = 2 维空洞数。
->
-> Betti numbers β_n = rank H_n(X): number of n-dimensional holes: β₀ = components, β₁ = 1-holes, β₂ = 2-voids.
->
-> **单纯复形（simplicial complex）**：由点、线段、三角形等单纯形组合而成的拓扑空间模型。
->
-> Simplicial complex: a topological space model composed of simplices (points, edges, triangles, etc.).
->
-> **滤流（filtration）**：从小到大改变邻域半径，在每个半径下构建单纯复形。
->
-> Filtration: vary neighborhood radius from small to large, constructing simplicial complex at each scale.
->
-> **持续同调（persistent homology）**：沿滤流计算同调群，长条对应真实拓扑特征，短条对应噪声。
->
-> Persistent homology: compute homology groups along filtration; long bars = true topological features, short bars = noise.
-
-> 详细数学依据见 `original-texts.md`
-
-## 不适用场景 / When NOT to Use
-
-- **需要精确测量** `[科研/通用]`——拓扑只关心定性结构（连通、有洞），不提供距离、角度等精确度量 / When precise measurements matter—topology only addresses qualitative structure (connected, holes), not distances or angles
-- **度量性质是核心** `[科研]`——拓扑等价的空间可以有完全不同的度量行为；度量空间的分析工具与拓扑不同 / When metric properties are essential—topologically equivalent spaces can have wildly different metric behavior; analysis requires more than topology
-- **纯离散结构无连续性** `[科研]`——拓扑以连续性为基础，对无连续结构的纯离散问题不适用 / Purely discrete with no continuity—topology rests on continuity; inapplicable to purely discrete problems
-- **需要精确数值决策** `[生活]`——当决策取决于"多少""多大"而非"是否连通""有没有洞"时，拓扑视角过于粗化 / When decisions depend on "how much" or "how big" rather than "is it connected" or "does it have holes"—topological perspective over-simplifies
-
-## 何时使用 / When to Use
-
-### 科研触发条件 / Research Triggers
-
-- **需要定性分类**——不在乎"多大""多远"，只在乎"是否连通""有多少洞" / Need qualitative classification—only care about connectedness and number of holes
-- **形状比大小更重要**——研究对象的整体形态而非精确尺寸 / Shape matters more than size
-- **需要鲁棒性分析**——拓扑不变量保证性质在连续扰动下不消失 / Need robustness analysis—invariants guarantee properties persist under continuous perturbation
-- **数据具有标准统计无法捕获的形状**——聚类、空洞、环状结构 / Data has shape that standard statistics misses
-- **需要变形下存活的不变量**——拓扑绝缘体、拓扑纠错 / Need invariant that survives deformation—topological insulators, topological error correction
-
-### 生活触发条件 / Life Triggers
-
-- **需要理解连通性**——谁和谁连通、信息能否流通、关系网络是否完整 / Need to understand connectivity—who connects to whom, can information flow, is the network complete
-- **需要识别"变化中的不变"**——不管表面怎么变，什么核心特征始终存在 / Need to identify "what stays the same through change"—what core features persist regardless of surface variation
-- **需要判断本质同类**——看起来不同的东西，核心结构是否相同 / Need to judge essential sameness—do superficially different things share the same core structure
-- **需要鲁棒性判断**——系统在扰动下是否稳定、哪些性质不会被消除 / Need robustness judgment—is the system stable under perturbation, which properties won't disappear
-- **需要发现盲区**——哪里有信息无法到达的区域、哪里有断裂 / Need to discover blind spots—where information cannot reach, where there are gaps
-
-## 方法流程 / Method
-
-### 第一步：识别连续性要求 / Identify Continuity Requirements
-
-**科研模式 / Research Mode**
-
-明确具体的等价标准：同胚（homeomorphism）、同伦（homotopy）、微分同胚（diffeomorphism）。确定什么可以改变（距离、角度、大小）什么必须固定（洞的数目、连通性、维数）。
-
-Specify the exact equivalence standard: homeomorphism, homotopy, or diffeomorphism. Identify what can change (distances, angles, sizes) and what must stay fixed (number of holes, connectedness, dimension).
-
-**生活模式 / Life Mode**
-
-什么可以改变，什么必须固定？——精确的数值不重要时，关注整体结构和连通性。问：哪些变化只是表面的（大小、外观、配置），哪些变化会破坏核心功能（连通断裂、关键环节消失）？
-
-What can change, what must stay fixed? — When exact values don't matter, focus on overall structure and connectivity. Ask: which changes are merely surface-level (size, appearance, configuration), and which would break core function (connectivity severed, key links lost)?
-
-**共通要点 / Shared Key Point**
-
-核心区分：可变（表面/度量）vs 不变（结构/拓扑）。/ Core distinction: variable (surface/metric) vs invariant (structure/topology).
-
-### 第二步：寻找拓扑不变量 / Find Topological Invariants
-
-**科研模式 / Research Mode**
-
-核心不变量清单：
-- **欧拉示性数** χ = V - E + F——凸多面体 χ = 2，环面 χ = 0，配合可定向性是曲面完全拓扑不变量
-- **连通性**（path-connectedness）——从任意点能否沿连续路径到达任意其他点；连通分量数 = rank H₀
-- **紧致性**（compactness）——每个开覆盖有有限子覆盖；Rⁿ 中紧致 ⇔ 有界 + 闭
-- **基本群** π₁(X)——环路模去同伦变形；π₁(S¹) = Z, π₁(S²) = 0, π₁(环面) = Z × Z
-- **同调群** H_n(X)——H₀ = 连通分量, H₁ = 1 维洞, H₂ = 2 维空洞
-- **Betti 数** β_n = rank H_n——β₀ = 连通分量数, β₁ = 1 维洞数, β₂ = 2 维空洞数
-
-Core invariant list:
-- **Euler characteristic** χ = V - E + F — complete surface invariant with orientability
-- **Connectedness** — number of components = rank H₀
-- **Compactness** — Heine-Borel: compact in Rⁿ ⇔ bounded + closed
-- **Fundamental group** π₁(X) — loops modulo homotopy; π₁(S¹) = Z, π₁(S²) = 0, π₁(torus) = Z × Z
-- **Homology groups** H_n(X) — H₀ = components, H₁ = 1-holes, H₂ = 2-voids
-- **Betti numbers** β_n = rank H_n — β₀ = components, β₁ = 1-holes, β₂ = 2-voids
-
-**生活模式 / Life Mode**
-
-不管怎么变化，什么特征始终存在？——找到"不管表面怎么变，内核不变"的东西。比如：不管组织怎么改组，沟通渠道的连通性不变；不管市场怎么波动，供需的循环结构不变。
-
-What features always exist regardless of change? — Find what stays the same at the core even as the surface changes. E.g.: no matter how an organization restructures, communication channel connectivity persists; no matter how markets fluctuate, the supply-demand loop structure persists.
-
-**共通要点 / Shared Key Point**
-
-不变量 = 变化中的恒定。/ Invariants = constancy through change.
-
-### 第三步：用不变量分类 / Classify Using Invariants
-
-**科研模式 / Research Mode**
-
-同欧拉示性数 → 同类型曲面；同基本群 → 同同伦类型空间；分类定理：紧致曲面由可定向性 + 欧拉示性数完全分类。
-
-Same Euler characteristic → same surface type; same fundamental group → same homotopy type; classification theorem: compact surfaces classified by orientability + Euler characteristic.
-
-**生活模式 / Life Mode**
-
-表面不同的东西核心结构是否相同？——甜甜圈和咖啡杯看起来不同，但核心结构一样（都有一个洞）。生活中看似不同的问题可能本质同类：两个部门的冲突模式可能是同一种循环依赖结构。
-
-Do superficially different things share the same core structure? — A donut and coffee cup look different but share the same core structure (one hole). Seemingly different life problems may be essentially the same kind: conflict patterns in two departments may share the same circular dependency structure.
-
-**共通要点 / Shared Key Point**
-
-分类靠不变量，不看表面看结构。/ Classification by invariants: look at structure, not surface.
-
-### 第四步：构造拓扑模型 / Construct Topological Model
-
-**科研模式 / Research Mode**
-
-**数据**：对点云数据构建滤流（filtration）——从小到大改变邻域半径，计算每个半径下的单纯复形及其同调；提取持续同调（persistent homology），输出条形码图（barcode diagram）或持续图（persistence diagram），长条 = 真实拓扑特征，短条 = 噪声。
-
-**网络**：分析连通分量（可达性）、环路（冗余路径）、聚类系数；网络拓扑决定信息传播与鲁棒性。
-
-**系统**：相空间中吸引子的拓扑——稳定不动点（拓扑上为点）、极限环（拓扑上为 S¹）、混沌吸引子（Strange attractor，分形结构）。
-
-For data: persistent homology via filtrations—barcode/persistence diagram where long bars = true features, short bars = noise. For networks: connected components, cycles, clustering coefficient. For systems: attractor topology — fixed point, limit cycle (S¹), chaotic attractor (fractal).
-
-**生活模式 / Life Mode**
-
-画出事物之间的连接关系——谁和谁连通，有没有断裂，有没有循环依赖，有没有空洞（信息无法到达的区域）。
-
-Map the connections between things—who connects to whom, are there gaps, circular dependencies, voids (areas where information cannot reach).
-
-**共通要点 / Shared Key Point**
-
-建模 = 把复杂现实映射到结构图。/ Modeling = mapping complex reality to a structural diagram.
-
-### 第五步：验证拓扑等价 / Verify Topological Equivalence
-
-**科研模式 / Research Mode**
-
-尝试构造显式同胚或同伦等价；检查不变量是否匹配；若不变量不同 → 空间拓扑不同（可靠否定判断）；若全部匹配 → 可能拓扑等价（但不能确定——不变量可能不完备）。
-
-Attempt explicit homeomorphism or homotopy equivalence; check invariant matching; different invariants → topologically different (reliable negative); all match → possibly equivalent (invariants may be incomplete).
-
-**生活模式 / Life Mode**
-
-验证你识别的核心结构是否真的可靠——变换是否真的不影响关键特征？问：如果表面变化了，我识别的"不变特征"还在吗？如果去掉一层包装，核心结构是否仍然成立？
-
-Verify whether the core structure you identified is truly reliable — does transformation really not affect key features? Ask: if the surface changes, does the "invariant feature" remain? If you strip away packaging, does the core structure still hold?
-
-**共通要点 / Shared Key Point**
-
-验证 = 确认不变量确实不变。/ Verification = confirming invariants are indeed invariant.
-
-### 第六步：拓扑推理与应用 / Topological Reasoning and Applications
-
-**科研模式 / Research Mode**
-
-利用拓扑结论推理：
-- "因为 π₁(X) ≠ 0，存在不可收缩环路 → 系统有不可达状态"
-- "因为 X 连通，存在路径 → 转移总是可能的"
-- "因为 β₁ = 3，数据中有 3 个显著的环状结构"
-- "因为 χ = 0，这是环面类型的曲面"
-
-Use topological conclusions for reasoning:
-- "Since π₁(X) ≠ 0, non-contractible loops exist → system has inaccessible states"
-- "Since X is connected, paths exist → transition is always possible"
-- "Since β₁ = 3, data has 3 significant circular structures"
-- "Since χ = 0, this is a torus-type surface"
-
-**生活模式 / Life Mode**
-
-用结构洞察指导行动——连通性告诉你信息能否流通，空洞告诉你哪里有盲区，循环依赖告诉你哪里可能卡住。
-
-Use structural insight to guide action—connectivity tells you whether information can flow, voids tell you where there are blind spots, circular dependencies tell you where things may get stuck.
-
-**共通要点 / Shared Key Point**
-
-从结构到结论：不变量 → 性质 → 推理。/ From structure to conclusion: invariants → properties → reasoning.
-
-## 常见错误 / Common Errors
-
-| 错误 / Error | 数学批评 / Mathematical Critique | 正确做法 / Correct Approach | 标签 / Tag |
-|-------------|-------------------------------|---------------------------|-----------|
-| 混淆拓扑等价与几何相似 / Confusing topological equivalence with geometric similarity | 拓扑等价只要求同胚，不要求度量相近；圆和椭圆拓扑等价但几何不同 | 区分"拓扑相同"和"几何相似"；先问"连续变形能否到达？"再问"度量是否接近？" | `[科研/通用]` |
-| 忽略维数差异 / Ignoring dimension differences | R¹ 与 R² 不同胚（维数是拓扑不变量——Brouwer 区域不变性定理） | 维数 invariance 是强制检查项；不同维数的空间拓扑必然不同 | `[科研]` |
-| 混淆连通与道路连通 / Confusing connected with path-connected | 连通 ≠ 道路连通（如拓扑学家的正弦曲线）；道路连通 ⇒ 连通，反之不成立 | 明确标注用的是哪种连通性；多数实际场景中道路连通更自然 | `[科研]` |
-| 过度依赖单一不变量 / Over-relying on a single invariant | 仅用欧拉示性数无法区分所有空间（χ=0 的空间多种多样）；单一不变量不完备 | 使用不变量组合（χ + 可定向性 + π₁ + H_n）；不完备时标注不确定性 | `[科研/通用]` |
-| 忽略局部与全局拓扑差异 / Ignoring local vs. global topological differences | 局部连通不保证全局连通；局部紧致 ≠ 紧致；局部同胚 ≠ 全局同胚 | 区分局部与全局性质；分别检查局部不变量与全局不变量 | `[科研]` |
-| 混淆同调与同伦 / Confusing homology with homotopy | H₁(X) 是 π₁(X) 的 Abel 化（H₁ = π₁/commutators）；同调是同伦的"粗化"版本 | 明确标注用的是 π₁（同伦）还是 H₁（同调）；需要精细分类时用 π₁ | `[科研]` |
-| 关注精确数值而忽略整体结构 / Focusing on exact values while ignoring overall structure | 拓扑的价值在于不看数值看结构；纠结精确数字会错失本质 | 先看连通性和循环结构，再看数值细节 / Look at connectivity and loop structure first, then numerical details | `[生活]` |
-| 忽视断裂和空洞 / Ignoring gaps and voids | 断裂 = 连通性丧失，空洞 = 信息无法到达的区域；这些是最重要的拓扑特征 | 主动检查：有没有断裂？有没有信息盲区？ / Proactively check: are there gaps? information blind spots? | `[生活]` |
-| 把表面差异当本质差异 / Treating surface differences as essential differences | 甜甜圈和咖啡杯表面不同但本质相同；表面差异 ≠ 结构差异 | 问"去掉表面包装，核心结构是否同类？" / Ask: "stripping surface packaging, is the core structure the same kind?" | `[生活]` |
-
-## 操作规程 / Operating Procedure
-
-当本 skill 被触发时，首先选择模式：
-
-**模式选择 / Mode Selection**
-- **科研模式**：当涉及拓扑数据分析、持续同调、同调计算、单纯复形与滤流等数学/计算任务时
-- **生活模式**：当涉及连通性理解、鲁棒性判断、识别变化中的不变、本质分类等日常推理任务时
-
-**Research Mode**: for mathematical/computational tasks involving TDA, persistent homology, homology computation, simplicial complexes & filtrations
-**Life Mode**: for everyday reasoning involving connectivity understanding, robustness judgment, identifying what persists through change, essential classification
-
----
-
-### 科研模式输出格式 / Research Mode Output Format
-
-1. **连续性要求**：`[连续性要求]: [同胚/同伦/微分同胚]，可变: [距离/角度/大小]，不变: [洞数/连通性/维数]`
-2. **不变量**：`[不变量]: χ=[值] π₁=[群] β_n=[数], H₀=[连通分量数] H₁=[1维洞数] H₂=[2维空洞数]，紧致性=[是/否]`
-3. **分类**：`[分类]: [可定向/不可定向], χ=[值] → [曲面类型]；π₁=[群] → [同伦类型]`
-4. **拓扑模型**：`[拓扑模型]: [单纯复形/滤流/网络图/相空间]，构造方法: [具体描述]`
-5. **等价验证**：`[等价验证]: [不变量是否全部匹配/是否有不变量不同]，结论: [拓扑等价/拓扑不同/不确定]`
-6. **推理**：`[推理]: 因为[拓扑性质] → [结论]`
-
-### 生活模式输出格式 / Life Mode Output Format
-
-1. **[什么可变/什么不变]:[区分]** — 精确数值不重要，什么核心特征必须保持 / What can change vs. what must stay fixed — exact values don't matter, what core features must persist
-2. **[不变特征]:[发现]** — 不管怎么变化，什么始终存在 / Invariant features discovered — what always exists regardless of change
-3. **[本质分类]:[判断]** — 看似不同的东西核心是否同类 / Essential classification — are superficially different things the same kind at core
-4. **[连接结构]:[分析]** — 事物之间的连通、断裂、循环、空洞 / Connection structure analysis — connectivity, gaps, loops, voids between things
-5. **[结构可靠性]:[验证]** — 识别的结构是否真的可靠 / Structural reliability verification — is the identified structure truly reliable
-6. **[行动建议]:[步骤]** — 基于结构洞察的结论 / Action recommendation — conclusions based on structural insight
-
-**输出必须包含以上 6 项，不得只输出分析性文字而不给出结论。**
-**Output must include all 6 items above. Do not output only analytical text without conclusions.**
-
-## 与其他 skill 的关系 / Relations to Other Skills
-
-- **对称与不变性**：拓扑不变量是连续变换下的不变量——拓扑思想与对称思想共享"寻找不变量"的核心逻辑，但拓扑关注连续变形而非群作用
-- **抽象化思想**：拓扑是空间结构的抽象——去掉度量只保留开集与连续性，是从几何到拓扑的抽象跃升
-- **变换思想**：连续变形是特殊的变换——同胚、同伦都是变换，拓扑研究这些变换下的不变量
-- **建模思想**：拓扑数据分析是数据建模——用持续同调为点云数据建立拓扑模型，发现统计方法遗漏的形状特征
-- **概率与统计**：TDA 补充统计的形状分析——统计关注均值与方差（量的分析），TDA 关注连通与空洞（质的分析），两者互补
+---
+name: topological-thinking
+description: |
+  触发：拓扑数据分析（TDA）、持续同调、同调/基本群计算、单纯复形与滤流构造；或为算法设计需要连通性/鲁棒性/上同调一致性判据（如 Čech 上同调作幻觉正则、sheaf 注意力）时调用。
+  English: Trigger for topological data analysis (TDA), persistent homology, homology/fundamental group computation, simplicial complexes & filtrations; or designing connectivity/robustness/cohomology-consistency criteria (e.g. Čech cohomology as hallucination regularizer, sheaf attention) for algorithms.
+---
+
+> **语言路由**：若用户消息为英文，请读取并遵循同目录下的 `SKILL.en.md`，按其操作规程以英文输出；中文消息则继续使用本文件。
+
+# 🌀 拓扑思想 / Topological Thinking
+
+> "可以拉伸弯曲但不能撕裂——真正重要的性质在连续变形中不变。"
+> "You can stretch and bend but never tear — truly important properties survive continuous deformation."
+>
+> —— 拓扑学、同调理论、拓扑数据分析 / Topology, Homology, Topological Data Analysis
+
+## 核心原则 / Core Principle
+
+拓扑学研究连续变形下保持不变的性质。甜甜圈与咖啡杯拓扑等价——都有一个洞。核心洞察：当精确距离不重要时，拓扑捕获本质结构——连通性、洞、维数。在本技能包语境里，拓扑思想直接支撑「胞腔层扩散注意力」与「Čech 上同调正则」（见 activator 范例）。
+
+> **数学形式化 / Mathematical Formalization**
+>
+> **同胚**：连续双射且逆连续 h: X→Y，拓扑等价标准。**同伦**：变形过程本身连续，比同胚更粗。**微分同胚**：光滑双射且逆光滑，微分拓扑标准。
+>
+> **欧拉示性数** χ=V−E+F：凸多面体 χ=2，环面 χ=0，配合可定向性是曲面完全拓扑不变量。
+>
+> **基本群** π₁(X)：基点环路模同伦变形；π₁(S¹)=Z（绕圈数），π₁(S²)=0（所有环路可收缩），π₁(环面)=Z×Z。
+>
+> **同调群** H_n(X)：n 维"洞"计数：H₀=连通分量，H₁=1 维洞（环路），H₂=2 维空洞。**Betti 数** β_n=rank H_n：β₀=连通分量数，β₁=1 维洞数，β₂=2 维空洞数。
+>
+> **单纯复形**：由单纯形（点/线段/三角形）组合的拓扑空间模型。**滤流**：从小到大改邻域半径，每半径下构建单纯复形。**持续同调**：沿滤流算同调群，长条=真实拓扑特征，短条=噪声。
+>
+> **上同调** Hⁿ(X)：同调的对偶（函子化、有环结构），Čech 上同调用于从局部数据检测全局一致性障碍——H¹≠0 即局部片段无法拼成全局相容对象（这正是"幻觉"的代数判据）。
+>
+> 详细见 `original-texts.md`。
+
+## GPU 友好性 / GPU-Friendliness（横切检查）
+
+拓扑结构能否上 GPU，取决于「局部性」与「张量化」——过 `../../references/gpu-friendly-math.md` 八维：
+
+- **局部拓扑量 → 友好**：每条边/每块的局部同调或限制映射是小运算，可批量并行、落小 GEMM。
+- **Čech 上同调正则**：作为局部一致性损失，廉价、可融合（友好）——这是 Tropical Sheaf Attention 的关键。
+- **持续同调（TDA）**：构造单纯复形与边界矩阵约简常含不规则稀疏/串行归约——典型"美但不可算"→ 改造为近似/采样/分布式。
+- **全局精确同调计算**：符号计算，不可训练，不友好。
+- **反模式**：把需要全局拓扑精确判定的结构塞进前向——不可算。
+
+八维最低判定（正式术语）：**张量化**看拓扑量是否可局部批量；**GEMM 可映射**看限制映射、边界算子、局部拉普拉斯是否落矩阵乘；**复杂度**看持续同调/边界矩阵约简是否超线性；**显存与 KV-Cache**看单纯复形、边界矩阵、覆盖交叠是否可压缩；**低精度稳定**看小奇异值、排序、阈值是否稳健；**并行与通信**看覆盖块/局部复形是否独立；**稀疏结构**看复形/图是否结构化稀疏；**算子融合**看局部一致性 loss 是否可融合。
+
+> 配合 `../../references/books/algebraic-geometry-rising-sea.md`（层/Čech 上同调→注意力与幻觉正则）、`smooth-manifolds.md`、`differential-geometry.md`。
+
+## 不适用场景 / When NOT to Use
+
+- **需要精确测量**——拓扑只关心定性结构（连通、有洞），不给距离/角度。
+- **度量性质是核心**——拓扑等价空间可有完全不同度量行为。
+- **纯离散结构无连续性**——拓扑以连续性为基础。
+- **需要精确数值决策**（取决于"多少"而非"是否连通"）——拓扑视角过粗。
+
+## 何时使用 / When to Use
+
+- 需要定性分类（不在乎"多大""多远"，只在乎"是否连通""有多少洞"）；形状比大小重要。
+- 需要鲁棒性分析（不变量保证性质在连续扰动下不消失）。
+- 数据有标准统计无法捕获的形状（聚类、空洞、环状结构）。
+- 需要变形下存活的不变量（拓扑绝缘体、拓扑纠错）。
+- **为注意力/表示设计全局一致性判据**（Čech 上同调作幻觉正则、sheaf 扩散）。
+
+## 方法流程 / Method
+
+### 第一步：识别连续性要求
+明确等价标准：同胚/同伦/微分同胚。确定什么可变（距离、角度、大小）什么必须固定（洞数、连通性、维数）。核心区分：可变（表面/度量）vs 不变（结构/拓扑）。
+
+### 第二步：寻找拓扑不变量
+- **欧拉示性数** χ=V−E+F（配合可定向性是曲面完全不变量）。
+- **连通性**（path-connectedness）：连通分量数=rank H₀；**紧致性**（Heine-Borel：Rⁿ 中紧致⇔有界+闭）。
+- **基本群** π₁(X)；**同调群** H_n（H₀=分量, H₁=1 维洞, H₂=2 维空洞）；**Betti 数** β_n=rank H_n。
+- **上同调** Hⁿ（同调的对偶，函子化，Čech 上同调检测全局一致性障碍）。
+
+### 第三步：用不变量分类
+同 χ→同类型曲面；同 π₁→同同伦类型；紧致曲面由可定向性+χ 完全分类。分类靠不变量，不看表面看结构。
+
+### 第四步：构造拓扑模型
+- **数据**：对点云构建滤流，计算每半径下单纯复形及同调，提取持续同调（条形码/持续图，长条=真实特征，短条=噪声）。
+- **网络**：连通分量（可达性）、环路（冗余路径）、聚类系数；网络拓扑决定信息传播与鲁棒性。
+- **系统**：相空间吸引子拓扑——不动点（点）、极限环（S¹）、混沌吸引子（分形）。
+- **算法**：用层/Čech 上同调把"局部片段能否拼成全局相容对象"变成可计算的代数障碍。
+
+### 第五步：验证拓扑等价
+尝试构造显式同胚/同伦等价；检查不变量匹配。不变量不同→拓扑不同（可靠否定）；全部匹配→可能等价（不变量可能不完备，不能确定）。
+
+### 第六步：拓扑推理与应用
+- "π₁(X)≠0 → 存在不可收缩环路 → 系统有不可达状态"
+- "X 连通 → 存在路径 → 转移总可能"
+- "β₁=3 → 数据有 3 个显著环状结构"
+- "H¹≠0 → 局部片段无法全局相容 → 幻觉/不一致信号"
+
+## 常见错误 / Common Errors
+
+| 错误 / Error | 数学批评 / Critique | 正确做法 / Correct Approach |
+|-------------|-------------------|---------------------------|
+| 混淆拓扑等价与几何相似 | 拓扑等价只要求同胚，不要求度量相近 | 区分"拓扑相同"与"几何相似" |
+| 忽略维数差异 | R¹ 与 R² 不同胚（维数不变性，Brouwer） | 维数 invariance 是强制检查项 |
+| 混淆连通与道路连通 | 连通≠道路连通（拓扑学家正弦曲线）；道路连通⇒连通，反之不成立 | 明确标注用哪种连通性 |
+| 过度依赖单一不变量 | 仅 χ 无法区分所有空间（χ=0 的空间多种） | 用不变量组合；不完备时标注不确定性 |
+| 忽略局部与全局拓扑差异 | 局部连通不保证全局连通；局部同胚≠全局同胚 | 分别检查局部与全局不变量 |
+| 混淆同调与同伦 | H₁(X) 是 π₁(X) 的 Abel 化；同调是同伦的粗化 | 标注用 π₁（同伦）还是 H₁（同调） |
+| 全局精确同调塞进训练 | 持续同调/符号同调含不规则稀疏与串行归约，不可算 | 用局部上同调损失/近似/采样，过 GPU 八维门 |
+
+## 操作规程 / Operating Procedure
+
+输出必须包含：
+
+1. **连续性要求**：`[同胚/同伦/微分同胚]，可变: [距离/角度/大小]，不变: [洞数/连通性/维数]`
+2. **不变量**：`χ= π₁= β_n=, H₀= H₁= H₂=，紧致性=[是/否]`
+3. **分类**：`[可定向/不可定向], χ= → [曲面类型]；π₁= → [同伦类型]`
+4. **拓扑模型**：`[单纯复形/滤流/网络图/相空间/层-Čech]，构造方法: [描述]`
+5. **等价验证**：`[不变量是否全匹配/有不同]，结论: [等价/不同/不确定]`
+6. **推理**：`因为[拓扑性质] → [结论]`
+7. **[GPU 可行性]**（若用于算法设计）— 拓扑量是否局部、可批量/落 GEMM，过八维门；全局精确同调则改造为近似
+
+**输出不得只给分析而无结论。**
+
+## 与其他 skill 的关系 / Relations to Other Skills
+
+- **对称与不变性**：拓扑不变量是连续变换下的不变量——共享"寻找不变量"逻辑，但拓扑关注连续变形而非群作用。
+- **抽象化思想**：拓扑是空间结构的抽象——去掉度量只留开集与连续性。
+- **变换思想**：连续变形是特殊变换；同胚/同伦都是变换。
+- **建模思想**：TDA 是数据建模——用持续同调为点云建拓扑模型，发现统计遗漏的形状。
+- **概率与统计**：TDA 补充统计的形状分析——统计关注均值方差（量），TDA 关注连通空洞（质），互补。
+- **现代数学激活**：`../../references/books/algebraic-geometry-rising-sea.md`（层/Čech 上同调→sheaf 注意力与幻觉正则）、`smooth-manifolds.md`、`differential-geometry.md`。

package/skills/topological-thinking/original-texts.en.md

@@ -0,0 +1,134 @@
+# Mathematical Sources and Classic Texts
+
+## Euler Characteristic (Euler, 1752)
+
+> "For any convex polyhedron, the number of vertices minus edges plus faces equals 2."
+
+**Leonhard Euler** (1707-1783) discovered this invariant of polyhedra in 1752:
+
+V - E + F = 2
+
+This formula seems simple but is profound — it holds for all convex polyhedra, whether cube (8-12+6=2), tetrahedron (4-6+4=2), or dodecahedron (20-30+12=2). The Euler characteristic is a topological invariant of surfaces: continuous deformations do not change χ. Sphere χ = 2, torus χ = 0, double torus χ = -2 — each additional hole reduces χ by 2.
+
+**Generalization**: for any surface S, χ(S) = 2 - 2g (orientable, genus g) or χ(S) = 2 - g (non-orientable, genus g). The Euler characteristic and orientability together form a complete set of classification invariants for compact surfaces.
+
+## Poincaré Conjecture & Perelman's Proof (2004)
+
+> "Every simply-connected closed 3-manifold is homeomorphic to the 3-sphere S³."
+
+**Henri Poincaré** (1854-1912) proposed this conjecture in 1904 — it became one of the most famous unsolved problems in the history of mathematics, confounding mathematicians for a full century.
+
+**Grigori Perelman** (1966-) posted three papers online in 2002-2003, completing the proof using **Ricci flow** (a differential geometric tool introduced by Hamilton) combined with surgery techniques. Perelman declined the 2006 Fields Medal and the 2010 Clay Millennium Prize of one million dollars — the most dramatic episode in the history of mathematics.
+
+**Intuition behind Ricci flow**: the metric on a manifold evolves over time so that curvature tends toward uniformity — much like heat diffusion makes temperature distribution uniform. When curvature concentrates at a "neck," surgery cuts it out and caps it with a standard sphere, then Ricci flow continues. The manifold is eventually decomposed into geometric pieces (Thurston's geometrization conjecture), each with a standard metric — the Poincaré conjecture is a special case of the geometrization conjecture.
+
+## Fundamental Group (Poincaré, 1895)
+
+> "The fundamental group π₁(X) consists of equivalence classes of loops based at a point modulo homotopic deformation — it captures the 'non-contractible loops' structure of a space."
+
+**Poincaré** introduced the concept of the fundamental group in his 1895 paper *Analysis Situs* — the foundational work of algebraic topology.
+
+Core examples:
+- **π₁(S¹) = Z** — loops on the circle are classified by winding number: a loop winding n times cannot be continuously deformed into one winding m times (n ≠ m)
+- **π₁(S²) = 0** — all loops on the sphere are contractible to a point — the sphere "has no holes"
+- **π₁(torus) = Z × Z** — winding in two independent directions, combined as (m, n) winding numbers
+- **π₁(real projective plane RP²) = Z₂** — only two classes of loops: contractible (even winding) and non-contractible (odd winding)
+
+**Key property**: homeomorphic spaces have isomorphic fundamental groups (topological invariant); however, the fundamental group is not a complete invariant — different spaces can share the same fundamental group (e.g., S² and S³ both have π₁ = 0, but they are not homeomorphic).
+
+## Homology Groups
+
+> "Homology groups H_n(X) count n-dimensional 'holes' in a space — translating topological intuition into computable algebraic quantities."
+
+**Simplicial homology computation**:
+1. Construct a simplicial complex K (approximating the space with simplices such as triangles and tetrahedra)
+2. Define the boundary operator ∂_n: C_n → C_{n-1} (the boundary of each n-simplex is a combination of (n-1)-simplices)
+3. Key property ∂² = 0 (the boundary of a boundary is zero)
+4. Homology group H_n(K) = ker(∂_n) / im(∂_{n+1}) — "cycles" modulo "boundaries" — true holes are not boundaries of any higher-dimensional simplex
+
+**Intuition**: H₀ = number of connected components (zero-dimensional "holes" = separated pieces), H₁ = number of 1-dimensional holes (unfillable loops), H₂ = number of 2-dimensional voids (unfillable cavities). Betti numbers β_n = rank H_n(X) count the holes.
+
+**Homology vs homotopy**: H₁(X) is the abelianization of π₁(X) (H₁ = π₁/[π₁, π₁], quotienting out the commutator subgroup). Homology is easier to compute (linear algebra) but less informative (non-abelian structure is lost); homotopy is richer in information but harder to compute. The choice depends on the problem at hand — π₁ for fine classification, H₁ for fast computation.
+
+## Topological Data Analysis (Carlsson, 2009)
+
+> "Persistent homology lets data 'speak' — it discovers not means and variances, but the shape of data."
+
+**Gunnar Carlsson** (Stanford) founded Ayasdi in 2009, bringing persistent homology into applications.
+
+**Persistent homology computation**:
+1. Starting from point cloud data, construct a filtration by varying the neighborhood radius ε from small to large
+2. For each ε, construct the Vietoris-Rips complex VR_ε (connecting point pairs at distance ≤ ε, filling triangles when three points are mutually within distance ≤ ε)
+3. Compute the homology groups H_n(VR_ε) for each ε
+4. Record the "birth time" (ε_birth) and "death time" (ε_death) of each topological feature
+5. Output a barcode diagram or persistence diagram (points (ε_birth, ε_death))
+
+**Stability theorem**: the Wasserstein distance between persistence diagrams ≤ the Hausdorff distance between data sets — small perturbations of the data produce only small changes in the persistence diagram. This guarantees the robustness of TDA: noise produces short bars, real features produce long bars.
+
+**Applications**: topological structure in tumor genomic data revealing new cancer subtypes; sensor network coverage hole detection; topological signatures of phase transitions in materials science; topological early warning of market crashes in finance.
+
+## Brouwer Fixed Point Theorem (1911)
+
+> "Every continuous map f: Dⁿ → Dⁿ has a fixed point — there exists x₀ such that f(x₀) = x₀."
+
+**L.E.J. Brouwer** (1881-1966) proved this theorem in 1911 — ironically, Brouwer later became the founder of intuitionism, opposing the law of excluded middle on which his own theorem relies.
+
+**Intuition**: stir coffee in a cup, then stop — at least one point must have returned to its original position, since stirring is a continuous map on a disk, which must have a fixed point.
+
+**Applications**:
+- **Game theory** (Nash, 1950): the existence of Nash equilibrium relies on the Brouwer fixed point theorem — every finite game has at least one mixed-strategy Nash equilibrium
+- **Economics**: the fixed-point proof of general equilibrium theory (Arrow-Debreu) — existence of market equilibrium
+- **Differential equations**: the Peano existence theorem uses fixed-point methods
+
+## Jordan Curve Theorem
+
+> "Every simple closed curve in R² divides the plane into two regions — interior and exterior."
+
+**Seemingly obvious yet extremely hard to prove** — Jordan's original proof (1892) had gaps; Veblen (1905) gave the first rigorous proof. The difficulty lies in the fact that closed curves can exhibit extremely complex topological behavior (e.g., Osgood curves with arbitrarily large area); what seems "obvious" intuitively requires fine topological argumentation mathematically.
+
+**Generalization**: Jordan-Brouwer separation theorem — every subset of Rⁿ homeomorphic to S^{n-1} divides Rⁿ into two connected components. Higher dimensions entail more complex proofs.
+
+## Knot Theory
+
+> "Knot classification relies on invariants — Jones polynomial, Alexander polynomial, etc. distinguish different knot types."
+
+Knot theory studies equivalence classes of simple closed curves in R³ under isotopy (classification up to continuous deformation). The central question: can two knots be transformed into each other by continuous deformation (without cutting or gluing)?
+
+**Key invariants**:
+- **Alexander polynomial** (1928): the first polynomial knot invariant, but incomplete (different knots can share the same Alexander polynomial)
+- **Jones polynomial** (1984): Vaughan Jones's discovery — deep connections to statistical mechanics and quantum field theory; Jones received the 1990 Fields Medal for this work
+- **Knot group** π₁(R³ - K): the fundamental group of the knot complement in three-dimensional space
+
+**Applications**: DNA topology — supercoiling, knotted DNA, and catenanes affect DNA replication and transcription; enzymes (topoisomerases) alter the topological structure of DNA, effectively performing "topological surgery." Molecular knot chemistry — synthesizing topologically complex molecular structures.
+
+## Classification of Surfaces
+
+> "Compact surfaces are completely classified by orientability and Euler characteristic — the paradigm of topological classification."
+
+**Classification theorem** (Möbius 1861, Dyck 1888, completed):
+
+Orientable surfaces: S² (χ=2), torus T² (χ=0), double torus (χ=-2), triple torus (χ=-4), ... — genus g surface has χ = 2 - 2g
+
+Non-orientable surfaces: RP² (χ=1), Klein bottle (χ=0), ... — genus g non-orientable surface has χ = 2 - g
+
+**Standard form**: every compact surface is homeomorphic to a connected sum of tori or projective planes. Orientable: #g T² (connected sum of g tori); non-orientable: #g RP² (connected sum of g projective planes).
+
+**Completeness of classification**: (orientability, χ) is a complete invariant set — two compact surfaces are homeomorphic if and only if they share the same orientability and the same Euler characteristic.
+
+## Morse Theory
+
+> "Critical points of a smooth function encode the manifold's topology — Morse theory bridges analysis and topology."
+
+**Marston Morse** (1892-1977) established this theory:
+
+A **Morse function** f: M → R has the property that every critical point p (where df(p) = 0) is non-degenerate (the determinant of the Hessian matrix ≠ 0). The index of a critical point = the number of negative eigenvalues of the Hessian = the dimension of the "downward directions" at the critical point.
+
+**Core theorem**: Morse inequalities relate critical point counts to Betti numbers:
+- c_k ≥ β_k (number of k-dimensional critical points ≥ k-th Betti number)
+- c_k - c_{k-1} + ... ± c_0 ≥ β_k - β_{k-1} + ... ± β_0
+
+**Intuition**: scanning M from low to high values of f, each critical point of index k triggers a "k-dimensional attachment" — equivalent to gluing a k-cell. The entire topology of the manifold is completely determined by all critical points and their indices.
+
+---
+
+**Summary**: Topological ideas pervade the core of mathematics — from Euler's characteristic to Poincaré's fundamental group, from the computability of homology groups to the data applications of persistent homology, from the applied power of fixed point theorems to the biological relevance of knot theory. Topology provides a unified perspective on "what remains invariant under continuous deformation." It tells us: what truly matters is not precise measurement, but the essence of structure — connectedness, holes, dimension, orientability. These properties survive stretching and bending, and change only under tearing and gluing — this is precisely the fundamental boundary that distinguishes topology from geometry.