</>MCP Agents Market
Skill

MathModelAgent

by jihe5204.1kPythonUpdated 2026-09-02

🤖📐专为数学建模设计的 Agent & skills ,自动完成数学建模,生成一份完整的可以直接提交的论文。 An Agent Designed for Mathematical Modeling ,Automatically complete mathmodel and generate a complete paper ready for submission.

Claude CodeCodex

MathModelAgent is an agent skill that automates mathematical modeling competitions from problem analysis through final paper submission. The skill orchestrates multiple specialized sub-agents to analyze competition problems, select appropriate mathematical models, generate Python code for analysis, create visualizations, and produce publication-ready papers in Typst or LaTeX format. It includes 17 competition templates (MCM, ICM, CUMCM, HuaShu Cup), built-in modeling knowledge bases with decision trees for model selection, and a 9-step automated quality assurance process to catch common errors before submission.

Key Features

End-to-end automation: single /1start-mathmodel command completes problem analysis, modeling, code generation, visualization, and paper formatting
17 built-in Typst templates for major mathematical modeling competitions (MCM/ICM, CUMCM, HuaShu Cup, Huawei Cup) in Chinese and English
Multi-agent architecture with specialized roles: problem analyzer, modeler, coder, paper writer, each configurable with different LLMs
Local Jupyter-based code interpreter plus cloud options (E2B, Daytona) for code execution and debugging
Built-in knowledge base with modeling method decision trees (AHP, TOPSIS, ARIMA, GA), common pitfalls, and MCM/ICM scoring rubrics
9-step automated validation: text leakage detection, numerical consistency checks, Typst compilation, PDF visual inspection
Sister project sci-box with scientific plotting templates (SHAP, ROC, Taylor diagrams, chord graphs) and draw.io flowchart templates
Works with any LLM via litellm provider integration; supports model-per-agent configuration

Use Cases

  • 01Automating mathematical modeling competition submissions for MCM/ICM, CUMCM, and regional contests
  • 02Generating complete research papers with reproducible code notebooks and publication-quality figures
  • 03Rapid prototyping of multiple modeling approaches (optimization, forecasting, machine learning) for problem analysis
  • 04Creating formatted competition papers from raw problem statements in under an hour
  • 05Learning mathematical modeling workflows through automated example generation and knowledge base references
  • 06Batch-testing different LLM combinations across problem analysis, coding, and writing tasks

Related Skills

View more

MathModelAgent — FAQ

What is MathModelAgent?+

MathModelAgent is an agent skill that automates the entire mathematical modeling competition workflow, from reading the problem statement to generating a submission-ready PDF paper with code, analysis, and visualizations. It uses multiple specialized AI agents to handle problem analysis, model selection, coding, and paper writing.

How do I install MathModelAgent?+

Install via npx skills add jihe520/MathModelAgent --all, then run it in Claude Code with /1start-mathmodel or in Codex with $start-mathmodel. Alternatively, download the standalone desktop app from GitHub Releases for macOS (arm64/x64) or Windows (x64).

Which AI clients support MathModelAgent?+

MathModelAgent works with Claude Code (claude --dangerously-skip-permissions) and Codex (codex --yolo). The desktop version bundles Claude Code and all skills for a zero-configuration experience.

Do I need API keys or other prerequisites?+

Yes, you need at least one LLM API key (supports any provider via litellm). Optional: Tavily API key for web search, Redis for local deployment, and Python/Node.js if not using the desktop app or Docker.

Is MathModelAgent free?+

Yes, MathModelAgent is open-source and free for personal use. You only pay for the LLM API calls you make (usage-based costs from your chosen provider). Commercial use requires contacting the author.

What competition templates are included?+

The skill includes 17 Typst templates for major Chinese and international competitions: CUMCM (China Undergraduate), MCM/ICM (COMAP), HuaShu Cup, Huawei Cup, and others. Templates auto-format according to competition requirements.

How do I install MathModelAgent?+

Open the source repository on GitHub and follow its README. MathModelAgent is a skill — MCP Agents Market links you directly to the official repo.

Is MathModelAgent free?+

MathModelAgent is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches