PluginBench
Skill
Pass
Audit score 90

1start-mathmodel

jihe520/mathmodelagent

Orchestrate mathematical modeling competition workflows from problem analysis through paper submission.

What is 1start-mathmodel?

This is the entry-point skill for mathematical modeling competitions. It initiates the complete workflow by gathering user preferences (typesetting engine, competition type, language, problem count), generates plan.md and todo.md files, and sequentially invokes downstream skills for problem analysis, modeling, coding, visualization, diagramming, paper writing, and verification.

  • Collects user preferences on typesetting engine (Typst/LaTeX), competition type, paper language, and problem structure
  • Generates plan.md with overall workflow strategy, modeling direction, phase sequence, and expected outputs
  • Creates todo.md as a phase-by-phase checklist to track completion status
  • Orchestrates sequential execution of six downstream skills in proper order: analysis-modeling → coding-visual → drawio → writing → verify
  • Maintains project directory structure with organized folders for reports, code, results, figures, and paper sections
  • Records user decisions and prevents downstream skills from duplicating work across phases

How to install 1start-mathmodel

npx skills add https://github.com/jihe520/mathmodelagent --skill 1start-mathmodel
Prerequisites
  • Downstream skills installed: 2analysis-modeling, 3coding-visual, 4drawio, 5writing, 6verity
  • Reference file available at ../_references/math_modeling_norms.md (optional, for domain guidance)
  • Typst or LaTeX installed (depending on user's typesetting preference)
Claude Code
Cursor
Windsurf
Cline

How to use 1start-mathmodel

  1. 1.Run the skill to begin the workflow entry process
  2. 2.Answer the prompted questions about typesetting engine, competition type, paper language, and problem count
  3. 3.Review the generated plan.md to confirm the workflow strategy and user preferences
  4. 4.Check todo.md to see the phase-by-phase checklist
  5. 5.Execute each downstream skill in order as listed in plan.md, updating todo.md status after each phase completes
  6. 6.Verify that each phase produces its expected output files in the correct directories before moving to the next phase

Use cases

Good for
  • Starting a new mathematical modeling competition project (MCM, ICM, national competitions, etc.)
  • Coordinating multi-phase workflows where problem analysis, implementation, visualization, and paper writing must happen in strict sequence
  • Managing document generation across different typesetting systems (Typst vs LaTeX) with consistent templates
  • Tracking progress through a 5-stage competition pipeline with clear handoffs between specialized skills
  • Ensuring all competition deliverables (code, figures, diagrams, paper) are generated in the correct order and stored in standard locations
Who it's for
  • Mathematical modeling competition participants
  • Team leads coordinating multi-person modeling projects
  • Instructors managing student competition workflows
  • Anyone automating the end-to-end competition submission process

1start-mathmodel FAQ

What if I don't know the number of sub-problems yet?

Select 'to be determined' when asked. The 2analysis-modeling skill will analyze the problem statement and determine the actual number of sub-problems during the problem analysis phase.

Can I switch between Typst and LaTeX after starting?

Yes, but you should update plan.md and regenerate the paper templates in 5writing. Both engines have equivalent template coverage (14 Chinese + 3 English templates).

What is the difference between 3coding-visual and 4drawio?

3coding-visual generates data-driven charts and plots from computational results. 4drawio creates conceptual diagrams, algorithm flowcharts, and architecture diagrams. Do not duplicate work between them.

Do I need to manually call each downstream skill?

This skill generates the plan and checklist. You then invoke each downstream skill (2analysis-modeling, 3coding-visual, etc.) in the specified order, updating todo.md as you complete each phase.

What if I want to skip a phase or change the workflow order?

Edit plan.md and todo.md directly to reflect your custom workflow. However, the standard 5-phase sequence is designed to ensure proper dependencies and output consistency.

Full instructions (SKILL.md)

Source of truth, from jihe520/mathmodelagent.


name: 1start-mathmodel description: "数学建模竞赛工作流入口。用于启动完整建模流程:询问用户偏好,生成 plan.md 和 todo.md,并按阶段调用赛题分析、建模、代码与图表、流程图、论文撰写、验证验收等 skills。" allowed-tools: Bash(*), Read, Write, Edit, Grep, Glob, Agent, WebSearch, WebFetch

数学建模工作流

本 skill 是数学建模竞赛项目的总控入口。它不替代后续阶段 skill,而是负责启动流程、询问偏好、记录决策、生成计划,并按顺序调用各阶段 skill。

数学建模规范参考

如需领域判断,读取 ../_references/math_modeling_norms.md。该文件只提供数学建模基本规范和防错知识,不改变本 skill 的阶段顺序和产出约定。

必须产出

在当前工作目录中创建或更新以下文件:

  • plan.md:整体流程方案、建模方向、阶段顺序、预期产物和风险控制。
  • todo.md:具体待办事项列表,记录每个阶段的任务和状态。

工作流

1. 询问用户偏好 AskUserQuestions

在规划前,只询问会实质影响流程的问题。问题要少而关键。

优先询问(按重要性排序):

  1. 排版引擎:Typst 还是 LaTeX?— 决定 5writing 使用哪套模板和编译命令。两套引擎均覆盖全部模板(14 中 + 3 英)。Typst 使用 typst 命令编译;LaTeX 使用 xelatex 命令编译(需跑两遍解决交叉引用)。
  2. 竞赛类型:国赛/华为杯/华中杯/MCM/...— 决定模板选择,见 5writing 的模板族清单。
  3. 论文语言:中文/英文 — MCM/ICM/COMAP 强制英文,其他默认中文。
  4. 子问题数量是否已知:影响章节文件生成数量。若未知,由 2analysis-modeling 阶段根据题面确定。

将用户的选择记录到 plan.md 的"方案"小节中。

2. 制定方案

按以下结构编写 plan.md:

# 方案

要依次调用这些 skill,按照里面要求完成任务。

用户偏好:
- 排版引擎:<Typst / LaTeX>
- 竞赛类型:<国赛 / 华为杯 / MCM / ...>
- 论文语言:<中文 / 英文>
- 子问题数量:<已知 N 个 / 待分析确定>

workflow:
   step      skills
1. 赛题分析与建模设计 - `2analysis-modeling`
2. 编程实现和图表生成 - `3coding-visual`
3. 流程与架构图绘制 - `4drawio`
4. 竞赛论文撰写 - `5writing`
5. 验证和验收 - `6verity`

项目目录结构

各阶段按此骨架创建和填充文件:

.
├── plan.md                      # 1: 本文件
├── todo.md                      # 1: 待办事项
├── reports/                     # 各阶段文档报告
│   ├── ANALYSIS_MODELING_REPORT.md  # 1: 赛题分析-建模报告(2analysis-modeling)
│   ├── RESULTS_REPORT.md            # 2: 结果报告(3coding-visual)
│   ├── DRAWIO_REPORT.md             # 3: 非数据图说明(4drawio)
│   ├── VERIFY_REPORT.md             # 5: 验收报告(6verity)
├── code/                        # 2: 代码(3coding-visual)
│   ├── problem1.py
│   ├── problem2.py
│   ├── problem3.py               # 问题的数量应该更具题目动态调整
│   ├── ... 
│   └── utils.py
├── results/                     # 2: 结果记录(3coding-visual)
├── figures/                     # 2+3: 所有图表(3coding-visual + 4drawio)
│   ├── *.pdf                    #     数据图 + 非数据图 PDF
│   ├── *.drawio                 #     非数据图源文件
├── paper/                       # 4: 论文(5writing)
│   ├── main.typ / main.tex      #     论文主文件(按用户选择的引擎)
│   └── sections/                #     各节文件(.typ 或 .tex)

方案必须明确每个阶段由哪个下游 skill 负责,以及该阶段应产出什么文件。

3. 生成待办

将 todo.md 写成阶段性 checklist,格式如下:

# 待办事项

- [ ] 1. 赛题分析与建模设计 - `2analysis-modeling`
- [ ] 2. 编程实现和图表生成 - `3coding-visual`
- [ ] 3. 流程与架构图绘制 - `4drawio`
- [ ] 4. 竞赛论文撰写 - `5writing`
- [ ] 5. 验证和验收 - `6verity`

每完成一个阶段,都要更新 todo.md 中对应任务的状态。

4. 依次执行阶段

按以下顺序调用下游 skills:

阶段Skill作用主要产物
赛题分析与建模设计2analysis-modeling解析题意、识别变量/约束/数据/评价指标,并建立数学模型、目标函数、约束条件和求解策略。ANALYSIS_MODELING_REPORT.md
编程实现和图表生成3coding-visual实现可复现代码,运行实验,生成结果表和多种多样的图表。code/, results/ , RESULTS_REPORT.md, figures/图表
流程与架构图绘制4drawio在论文确实需要时,绘制方法流程图、架构图和非数据型概念图。figures/*.drawio, figures/*.pdf, DRAWIO_REPORT.md
竞赛论文撰写5writing基于分析、建模、代码结果和图表撰写最终竞赛论文,并按章节直接插入图表。paper/
验证和验收6verity检查可复现性、一致性、产物完整性、格式规范和提交就绪状态。VERIFY_REPORT.md

阶段边界

  • 3coding-visual 负责生成所有依赖计算结果或实验输出的数据图表。
  • 4drawio 只负责概念图、算法流程图、架构图、路线图等非数据型图示。
  • 不要让 4drawio 重复绘制 3coding-visual 已经生成的统计图或数据图。
  • 5writing 负责决定图表在论文中的位置,并按所选引擎写入图表代码:
    • Typst:#figure(image("../../figures/xxx.pdf", width: 85%), caption: [...])
    • LaTeX:\begin{figure}[H]\centering\includegraphics[width=0.85\textwidth]{../../figures/xxx.pdf}\caption{...}\label{fig:xxx}\end{figure}
  • 不要让 5writing 编造数值结论。论文中的数值必须来自 RESULTS_REPORT.md、结果表或已生成图表的数据。