agentleFS
chaobo8484chaobo8484/MathX-math-model-skills/skills/ask-mathx/SKILL.md

ask-mathx

Skill0 starsChanged 2 months agoMITView on GitHub

不知道 35 个技能该用哪个时先问我:我帮你选 1-2 个并排好顺序。

The file

Files
SKILL.md

Ask-MathX:技能路由器

用户键入 /ask-mathx 是因为不知道哪个技能合适。本技能永远不干活,只指路:告诉用户跑哪 1–2 个技能、什么顺序,然后停下,让用户去调用它们。

How to route

  1. 用户目标一句话:排序、预测、优化、写作、证明、还是画图?
  2. 先查混淆对(下表)——多数误触发都是其中之一。
  3. 最多点名两个技能加顺序。超过两个说明没理解任务;改问一个澄清问题。
  4. 项目第一次用?先路由 /setup-mathx。公共术语在 CONTEXT.md——原文照用。
  5. 用户直接甩文件(CSV/XLSX/TXT/PDF/JPG/MP4)?先路由 ingest-inputs 归一,拿质量报告再按任务路由。

Confusion pairs(先查这里)

用户说其实是两个活判定
评价/排序/赋权ahp 对 evaluation-entropy-topsis准则靠专家经验打分 → ahp;全是指标实测数据、要客观 → evaluation-entropy-topsis
预测gray-prediction 对 time-series-arima只有 4–10 个点、近指数趋势 → gray-prediction;等间隔 30+ 点、有趋势或季节性 → time-series-arima
回归还是分类regression-family 对 clustering-classification 对 bp-neural-network解释 y~X 或变量选择 → regression-family;发现群组或打标签 → clustering-classification;几十到几千样本的复杂非线性 → bp-neural-network
优化optimization-lp-milp 对 genetic-algorithm目标约束可线性表达 → optimization-lp-milp;非凸、非光滑、组合(调度选址路径)→ genetic-algorithm
画图chart-decision 对其他还没定画什么图 → 先 chart-decision;定了图型再去对应技能

Full index

Modeling——方法纪律,agent 在流程中自动调用: ahp(主观多准则排序)、evaluation-entropy-topsis(客观多指标排序)、gray-prediction(小样本预测)、time-series-arima(时间序列)、bp-neural-network(非线性拟合)、regression-family(回归与正则化)、clustering-classification(聚类与分类)、genetic-algorithm(非凸组合优化)、optimization-lp-milp(线性整数规划)、monte-carlo-simulation(随机模拟)、differential-equation(微分方程)、graph-network(图与网络)。

Writing——交付物生产,由人逐个触发: paper-outline(先搭大纲故事线)→ literature-review(缺口表)→ 方法技能 → figure-table-generation(图表)→ latex-typesetting(填模板编译)→ citation-bibliography(引用)→ polish-proofread(润色)→ reproducibility-checklist(交稿门禁)。

Research——研究纪律,可组合: conjecture-formulation(提可证伪命题)→ proof-assistant(拆引理标缺口)/ symbolic-computation(符号推导)/ numerical-verification(数值证据分级)/ counterexample-search(找最小反例);背景不清要地图时先 arxiv-literature-synthesis;单条断言需外部支撑(选型依据、邻居站位、求解器版本)时先 research-evidence。

Visualization——绘图纪律: 没定图型先 chart-decision;静态矢量图 scientific-plotting;带 N/p 值/CI 的统计结论 statistical-plot;要交互 plotly-interactive;定稿拼版 publication-figure;流程架构 diagram-schematic。

Inputs——用户文件先过 ingest-inputs(解析归一 + 质量报告 + 来源记录),再进上面流程。

Common flows(抄送答案)

  • 建模比赛从零开始:paper-outline → 方法技能 → figure-table-generation → polish-proofread → reproducibility-checklist。
  • 只有一小串数据要预测:先看点数,≤10 点 gray-prediction,30+ 点 time-series-arima,之间先问数据。
  • 论文定稿前:figure-table-generation → latex-typesetting(终稿 PDF 用 TEX,协作草稿用 DOCX 分支)→ citation-bibliography → polish-proofread → reproducibility-checklist。
  • 猜想验证:conjecture-formulation → numerical-verification,证伪走 counterexample-search,证明走 proof-assistant。
  • 需外部支撑(选型、站位、版本核验):先 research-evidence 取证,再进方法技能。
  • 用户甩文件:先 ingest-inputs 归一,再按任务路由。

Tone

简短果断。技能点名,顺序一行一个,停。如果任务真跨三个技能,先说清要用户回答哪个问题再路由——不要一次路由三个。

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.