Math Modeling Agent (3-Role Pipeline)
A rigorous math-modeling-contest workflow that splits work into modeler, coder, and paper-writer roles with mandatory independent QA gates at every stage.
EducationAdvanced★ 736⑂ 42AI score 9/10Last updated: Aug 17, 2026
What it does
- Runs math modeling problems through three separable roles: modeler → programmer → paper writer, either end-to-end or one role at a time.
- Locks down deliverables per stage: problem-analysis report + glossary → runnable Python/MATLAB code, result tables (CSV/XLSX), at least 9 figures covering raw data, model process and results for every sub-question, plus a reproduction manifest (random seed, input SHA-256, dependency versions) → a Word paper (LaTeX + compiled PDF on request).
- Enforces independent Subagent QA gates (M1/P1/P2/W1/W2): self-review never counts as passing, and any FAIL routes work back to the responsible role.
- Uses progressive loading — algorithm index, PDF reader, Excel handler, paper search, figure and DOCX/LaTeX tools are read only when needed.
Who it's for
- Students and teams preparing for math modeling competitions (MCM/ICM, CUMCM and similar).
- Grad students and researchers who need reproducible computation plus a publishable-style write-up.
- Data analysts who want strict consistency between code output and the figures/tables in a report.
Example uses
- "Solve contest problem A end to end" → analysis report → M1 gate → minimal runnable result (P1) → full computation and figures → P2 → evidence outline (W1) → final
完整论文.docx(W2). - "Just help me pick models" → modeler role only: candidate model families per sub-question with justification, plus a glossary.
- "I already have the model — write the code and make the plots" → programmer role: Python/MATLAB scripts, raw/process/result figure sets per sub-question, and a reproduction manifest.
· · · Install guide · · ·
Try it now, no install
Paste this into Claude to use the skill without installing anything.
Read the instructions in this file and follow them to help me: https://raw.githubusercontent.com/XiaoMaColtAI/math-modeling-skill/HEAD/dsh-plugin/math-modeling-agent/skills/math-modeling/SKILL.md What I want: (describe your task here)
If Claude can't open the link, open it yourself and paste the contents instead.
↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.
Install in the Claude app (no terminal)
- Download the ZIP with the button below.
- In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
- Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
Install in Claude Code
Let Claude do it — paste this into Claude Code
Install the skill I found on Claude Skill Mart. Copy the dsh-plugin/math-modeling-agent/skills/math-modeling folder from the GitHub repo XiaoMaColtAI/math-modeling-skill into my ~/.claude/skills/xiaomacoltai-math-modeling/. When it's done, tell me in one line what this skill can do.
Install with a command instead
git clone https://github.com/XiaoMaColtAI/math-modeling-skill.git && mkdir -p ~/.claude/skills && cp -r math-modeling-skill/dsh-plugin/math-modeling-agent/skills/* ~/.claude/skills/⚠ This is a third-party skill. Check the source repository before installing.
- Open a terminal and clone the repo:
git clone https://github.com/XiaoMaColtAI/math-modeling-skill.git - Create the skills folder if needed:
mkdir -p ~/.claude/skills - Copy the skill in:
cp -r math-modeling-skill/dsh-plugin/math-modeling-agent/skills/* ~/.claude/skills/ - Verify with
ls ~/.claude/skills/math-modeling/SKILL.md. - Prepare the toolchain: Python (numpy, pandas, matplotlib) or MATLAB, and a TeX distribution (TeX Live/MiKTeX) if you want the LaTeX/PDF paper.
- Restart Claude Code, cd into your project folder (with the problem PDF and data files), and ask: "Use the math-modeling skill to solve this problem."
- Note: the skill docs are in Chinese; copy
使用指南.mdinto your project folder first to understand the usage boundaries of the generated paper.