Modeling Code and Result Contracts
yushui2022/MathModel-Skill
Generates result-evidence contracts, tables and runnable q1 to q3 modeling code scaffolds for a math modeling paper from a model route, a data plan and cleaned data.
Runs a staged pipeline for mathematical modeling research and contest papers, from problem analysis and computation to paper writing, review rounds and submission checks.
SKILL.md written in Chinese; this summary is our English description.
Install Meta-model-agent Math Modeling Pipeline
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agent --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
Claude Code skills documentation · loads skills from .claude/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .claude/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agent --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .agents/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agent --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .cursor/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agent --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .gemini/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agentInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .github/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install WuXinbo-bo/Math-model-skills meta-model-agent --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "meta-model-agent" agent skill from https://github.com/WuXinbo-bo/Math-model-skills/tree/main into .opencode/skills/meta-model-agent/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "meta-model-agent", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
meta-model-agent- GitHub stars
- 111
- Token cost
- ~2.3k tokens
- SKILL.md length
- 397 words
- Files
- 163 (incl. scripts, references, assets)
- Skills in repo
- 9
- Repo updated
- First seen
- Licence
- MIT
At a glance
Runs a staged pipeline for mathematical modeling research and contest papers, from problem analysis and computation to paper writing, review rounds and submission checks.
- Works in 5 steps: 严守 baseline 基线 → 按阶段渐进加载 → 落实主控与工作角色分层 → …
- Starting or resuming a mathematical modeling contest project
- SKILL.md covers 项目说明, 条件数据准备路线, 快速启动 and 实施准则, plus 2 more sections
- Runs Shell scripts from its folder; calls python
What it does
Meta-model-agent links problem analysis, modeling, program solving, result checks, figures, paper writing and submission checks into one resumable, auditable chain, aimed at contests such as CUMCM, 51MCM and MCM/ICM and similar research tasks. Work runs through seven stages (including DISCOVERY, FORMULATION, COMPUTATION and MANUSCRIPT) with evidence gates, and a project can be started or resumed. If upstream evidence is flawed, the work goes back to the earlier stage instead of polishing the text over the defect.
Three quality tiers exist: baseline keeps stages, deliverables and gates stable; enhancement does controlled rework on weak models, programs, evidence chains or paper sections; championship adds several independent mock reviews and a full-text revision before acceptance. Formulas, numbers, figures and conclusions in the paper must trace to the problem statement, model, program or result files. Data preparation is a conditional sub-flow: when data exist, models read only hashed, preprocessed inputs, and when none exist the skill must say so rather than invent preprocessing.
Model naming has its own contract: a presentable name that pairs a problem-specific mechanism with a standard mathematical model, kept apart from the solving algorithm so that Gurobi or a genetic algorithm is not passed off as the model. The skill states that you must verify the problem reading, data, models, code, citations and final paper yourself. It is written mainly in Chinese, targets Codex first, and ships 166 files including references, shared scripts and TeX demos.
When your agent uses it
- Starting or resuming a mathematical modeling contest project
- Fixing a weak model, program or evidence chain before the paper is finalized
- Running mock review rounds on a modeling paper before submission
- Checking that every number and figure in the paper traces to result files
Example prompts
- “Start a new CUMCM project from the problem statement in ./problem.pdf and set up the stages.”
- “Resume the modeling project and continue from the earliest unfinished stage.”
- “Run the championship review rounds on my paper draft and list what must be fixed before submission.”
Requirements
- An AI tool that can read skills, access the workspace and run local commands
Workflow steps
5 steps, taken from the step headings in SKILL.md.
What it can do on your machine
Read from SKILL.md and the folder at commit ee8a616. It shows what the files ask for, not the result of running them.
Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
Ships 1 file in scripts/ (Shell, from the files we listed), which the agent can run.
Shell commands in SKILL.md call:
python
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Context cost
Meta-model-agent Math Modeling Pipeline loads about 2.3k tokens when it runs, and up to ~22k if it reads all its reference files. Until then it costs about 49 tokens; SKILL.md has 397 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
Safety
Auto-check passed
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
SKILL.md
The full file from WuXinbo-bo/Math-model-skills at commit ee8a616, republished under its MIT licence (© WuXinbo-bo). 397 words, ~2,324 tokens.
Download SKILL.mdSave it as
.claude/skills/meta-model-agent/SKILL.md (or your agent's skills folder). This skill also uses 162 other files; get the full folder from GitHub.- name
- meta-model-agent
- description
- Meta-model-agent 致力于辅助研究者与参赛团队完成高质量数学建模研究和论文产出，提供从问题理解、模型构建、计算实验、证据可视化到论文写作、冠军级多轮审稿与提交验收的完整工程支持。适用于 CUMCM、51MCM、MCM/ICM 及同类数模研究任务，可用于启动或恢复项目、按证据门禁推进、修复薄弱环节并形成可复现、可验证、可提交的高质量数模论文。
Meta-model-agent
Meta-model-agent v1.3 是面向数学建模研究与竞赛论文生产的工程化辅助系统。它将题意研判、数学建模、程序求解、结果验证、图形表达、论文组织和提交检查连接为一条可执行、可恢复、可审计的研究链，帮助使用者把零散的分析过程转化为证据完整、逻辑一致且能够复现的高质量数模论文。
项目说明
本项目的核心目标不是简单生成一篇论文文本，而是辅助完成从研究问题到最终论文的完整质量闭环：先建立可信的问题与模型契约，再通过真实计算获得结果，以图形、表格和结构图组织证据，最后形成符合竞赛规范的论文并接受多轮质量审稿。
项目优先适配 Codex，并兼容能够读取 Skill、访问工作区和运行本地命令的其他 AI 工具。Meta-model-agent 仅提供数学建模辅助；实际使用者必须核验题意、数据、模型、程序、引用、结果与最终论文，并自行承担采用或提交相关产出的风险。
项目按照研究依赖组织工作，并提供三层质量能力：
baseline：保持阶段、交付物和门禁稳定，保证完整流程能够可靠运行；enhancement：针对模型、程序、证据链或论文薄弱项实施受控返工；championship：在最终验收前执行多轮独立模拟审稿和全文修订，以更高标准检验数学正确性、可复现性、逻辑一致性、表达质量与提交合规性。
Meta-model-agent 强调研究证据优先。论文中的公式、数值、图表和结论应能追溯到题面、模型、程序或结果文件；若上游证据存在根本问题，系统应回退相应阶段修复，而不是仅通过文字润色掩盖缺陷。
条件数据准备路线
数据准备是七阶段内的条件子流程，不新增顶层阶段：DISCOVERY 声明 supplied/collected/none 数据模式，FORMULATION 定义预处理合同，COMPUTATION 按“质量审计 → 题目驱动预处理 → 前后质量核验 → 冻结模型输入 → 模型计算”执行。无数据时必须明确豁免且不得伪造预处理产物；有数据时模型只能读取 数据/processed/ 下经过哈希固化的规范输入。预处理方法必须由字段类型、数据质量、时间/空间结构和模型需求决定，禁止对所有题目套用同一个清洗脚本。
模型身份与算法分离合同
- FORMULATION 在现有
建模报告.md中分开登记审计身份与论文表达，不新增文件。建立或扩展模型的子问题写模型定义、模型结构、模型语义和论文表达；模型扩展另写模型增量。仅做比较、验证或应用的子问题只写论文表达，并明确继承的上游模型，禁止强造新模型。 论文表达 QN | 展示名称: ... | 问题角色: new_model/model_extension/comparison/validation/application | 继承模型: none/QN,... | 核心方法: ...。展示名称必须是简洁、可发表的模型名；内部机制、方案代号、运行预算、验证器和工作文件名不得塞入展示名称。- 正式名称使用“题目定制机制 + 标准数学模型”格式，例如“考虑维护约束的机组承诺混合整数线性规划模型”“基于季节项与滞后项的动态回归模型”。
优化模型、决策模型、预测模型、综合模型等泛化名称不能单独通过；标准模型族必须能识别为线性/整数/鲁棒/随机规划、车辆路径、网络流、回归、时间序列、状态空间、微分方程、多指标决策、仿真等明确结构。 - 数学模型、定制机制和求解算法必须分开：模型回答变量如何关联以及目标/约束是什么，定制机制说明题目特有结构，算法只说明如何求解。HiGHS、Gurobi、分支定界、遗传算法、粒子群、匈牙利算法等不得冒充数学模型。
- COMPUTATION 必须在现有
图表/全部结果.json.model_identity中逐问固化名称、模型族、求解算法及模型语义字段，并与 FORMULATION 身份卡逐字一致；不新增结果文件。 - MANUSCRIPT 正文保持模型语义一致，但不复制内部身份卡句式。摘要使用论文表达卡中的展示名称和核心方法；比较、验证或应用类问题应写“基于前述模型进行……”，不得伪造独立数学模型。
统一机理与逐问增量合同
- FORMULATION 必须先识别跨问题复用的机制内核，再按
new_model/model_extension/comparison/validation/application组织后续问题。后续问题不得重复包装前问，也不得仅因更换算法就宣称建立新模型。 new_model/model_extension在现有建模报告.md中增加模型语义卡，明确目标数量、方向、变量类型、关系类型和多目标证据；多资源、高维决策或多个评价指标不自动构成多目标优化。model_extension另登记继承方程、新增变量、新增/修改约束、目标变化、求解变化和验证变化。扩展后的验证必须针对新增机制。- 物理、几何、事件、仿真和混合决策题优先形成“状态/机制内核 -> 判定或事件集合 -> 外层优化/分配 -> 执行方案”的连续主线。具体合同见
references/model-quality-contracts.md。 - COMPUTATION 使用现有
全部结果.json.publication_claims固化发布数值、来源键和派生关系；摘要、正文、图表与结论不得各自手抄同一结果。
摘要内容合同
- 中文竞赛摘要按“总述—逐问—评价”组织：首段说明问题、核心矛盾、主要数学模型及任务；中间每个子问题用自然学术语言写“简洁模型名或继承关系—核心方法—1 至 2 个关键结果—验证结论”；尾段总结模型的可解释性、稳健性、推广性和局限性。
- 摘要不得出现“标准模型族为”“冻结合同”“统一验证器”“搜索预算”“Q1/Q2 求解器”等字段化或内部工作流语言；不得罗列初始化、修复、加速、校验等完整算法链。方案缩写首次出现必须解释，结果数字服务于结论而不是形成流水账。
- 关键词优先使用标准模型族和核心求解算法，通常 3--5 个，例如“混合整数线性规划、动态回归、车辆路径、鲁棒优化”；不得只使用自创长名称，也不得用“研究、分析、结果、优化问题”等空泛词。
- MCM/ICM 的英文 Summary Sheet 使用同一信息链，并额外保留 recommendations 与 limitations；模型名和算法名必须与正文一致。
- 摘要必须最后写，所有模型名、算法名、数值和检验结论均从当前有效正文与验证证据提取，不得凭规划阶段记忆补写。
论文版面与正文密度硬约束
- 图表文件的原始画布尺寸不是论文嵌入尺寸。LaTeX 普通图默认使用
width=0.72\linewidth,height=0.70\textheight,keepaspectratio，再按可读性调整；宽表或宽图最高不得超过\linewidth，禁止依赖 PDF 查看器裁切超出版心的内容。 - DOCX 中按页面可用宽度缩放图表：A4 常规页边距下正文图推荐宽 13.5--15.0 cm，绝不超过实际版心宽；必须锁定纵横比，禁止把像素尺寸或厘米尺寸原样放大到页面之外。
- 图表过大或导致页数超限时，优先裁除空白边缘、合并相关子图、精简重复图、缩短图注、将次要图表移至附录；不得通过把整张图缩到文字不可读来“塞进页面”。图中文字在最终 PDF/DOCX 的 100% 显示比例下必须可读。
MAX_PAGES是上限，不是正文字数目标。页数紧张时必须优先保护问题分析、模型机制、公式推导、求解过程、验证、结果解释和局限性，不能为了满足页数而删成只有结论和图表的空心论文。- 页数合规与正文充分性必须分别验收。即使 PDF 未超页，若任一子问题缺少“机制/推导 -> 结果 -> 验证 -> 解释”，或正文有效字数明显不足，仍不得通过 MANUSCRIPT/ASSURANCE 门禁；禁止用放大字号、拉大行距、堆图或空泛文字补页。
- CUMCM 按 2026 官方规范执行：摘要原则上不超过 1 页，正文不要目录且不超过 30 页，附录单独计数且页数不限。必须通过编译后的
AbstractStart/AbstractEnd与BodyStart/BodyEnd标签实测，禁止用字符估算代替最终门禁。 - 普通图默认使用
0.72\linewidth，0.85\linewidth仅作为普通宽图建议上限；确需更宽时最高不得超过\linewidth，并始终启用keepaspectratio。普通 LaTeX 图表使用[htbp]并在章节边界用\FloatBarrier收束，禁止全篇强制[H]造成空白页和页数膨胀；只有模板或局部版式明确要求时才允许少量[H]。 - LaTeX 核心公式必须使用原生数学环境，执行“机制引入 -> 编号公式 -> 符号/单位/定义域 -> 公式作用”的语义顺序；禁止公式截图、公式无解释堆叠和手写图表公式编号。具体规则见
references/paper-layout/latex-math.md。 - ASSURANCE 必须审计最终 PDF 的页面构成，不以编译成功代替排版通过。重点检查低页面占用率、标题孤悬、正文孤行、图题分离、跨页表头、极小文字、连续大图和密度突变，见
references/paper-layout/page-composition.md。
图表与流程图克制表达硬约束
- 每张图和每个
TABLE_*表必须登记明确论点、数据来源和读者任务；使用图表/figure_manifest.json的figures/tables区分正文、附录和诊断资产，只有publish=true的图表必须嵌入正文。 - 图表同时登记所属问题、
mechanism/result/validation/decision/diagnostic视觉角色和所依赖的结果键。正文证据按“机制 -> 结果 -> 验证/决策”组织，不以图表数量替代论证。 - 图形类别按数据结构和读者任务选择，禁止为了“高级感”或多样性强制使用渐变、阴影、圆角标注框、KDE 背景和无必要多层叠加。
- 所有 DrawIO/TikZ 流程图中的普通矩形、矩形容器和表头必须使用直角矩形；菱形、圆形、六边形、圆柱、平行四边形等其他语义形状保持不变。
- AI Image 默认关闭，禁止用于技术路线图、流程图和模型架构图；仅在用户明确需要且经过内容级核验的物理场景示意中使用。
代码附录硬约束
- COMPUTATION 必须生成
程序/code_manifest.json，登记入口、逐问程序、依赖、代码行数和源码哈希。 - MANUSCRIPT 必须运行
工具/build_code_appendix.py，从当前源码自动生成附录；禁止用一段复现说明或“代码见支撑材料”替代真实核心实现。 - 三种赛制的论文附录均只嵌入主程序和逐问核心实现；数据处理、绘图、校验及公共工具仅进入支撑材料清单，不展开完整源码。附录中的程序显示名统一使用英文，真实路径和源码哈希仅保留在机器审计标记中。源码哈希、附录标记和运行入口不一致时不得通过门禁。
快速启动
- 在目标工作目录初始化运行时工作区：
python scripts/workspace_init.py --workspace . --competition cumcm --output-format pdf--competition 支持 cumcm、51mcm、mcm-icm。竞赛类型会写入状态并控制模板与硬门禁；如需更换竞赛，应重新初始化工作区并重新执行论文及验收阶段。
--output-format 支持 pdf（默认）和 docx。DOCX 路线以 论文/论文正文.md 为写作源，使用 python 工具/docx_export.py --workspace . 生成 论文/数模论文.docx 与图片尺寸报告。
查询当前阶段：
python scripts/stage_executor.py current --workspace .- 启动当前阶段，自动化同步该阶段所需的
工具/参考资料/模板：
python scripts/stage_executor.py begin DISCOVERY --workspace .- 仅加载当前阶段的实施协议：
references/stage_protocols/<skill-name>/SKILL.md- 如该阶段有额外
references/或templates/，只加载当前阶段需的文件。
- 阶段实施完毕后，严格按下面次序收口：
python scripts/stage_executor.py validate DISCOVERY --workspace .
python scripts/stage_executor.py gate_check DISCOVERY --workspace .
python scripts/stage_executor.py complete DISCOVERY --workspace . --artifacts "问题分析.md"- 若该阶段有核验点，继续处理核验点：
python scripts/stage_executor.py checkpoint DISCOVERY --workspace . --action approve --note "checkpoint passed"实施准则
Show full SKILL.md (162 more words)Show less
冠军质量模式
需冲击最高质量时，先设定冠军模式：
python scripts/pipeline_manager.py set-mode championship --workspace .冠军模式保持问题分析、模型构建、计算实验、证据整理、论文写作与提交验收之间的依赖关系不变，并在论文写作完成后强制加入多轮模拟审稿。载入 references/stage_protocols/championship-review/SKILL.md 和 references/championship-review-method.md，依次开展独立审稿、修订规划、全文重写与逐项复核。完成不少于三轮，且终版满足 P0 为 0、P1 不超过 2、综合分不低于 85 后，方可转入提交质量验收。
1. 严守 baseline 基线
- 默认运行模式为
baseline。 - 基线运行严禁打乱上游与下游依赖、缩减交付物目录、改变核验点类别、降低最低文件规模或取消伴随文件条件。
- 只有完整基线证据通过后，才能启动增强处理。
1.1 严格执行竞赛 Profile
- 初始化时将竞赛类型、语言、模板、允许/禁止身份字段、字号和页数限制写入状态。
MANUSCRIPT只能同步当前竞赛模板，不得混用其他竞赛封面或文档类。- MCM/ICM 强制英文、至少 12pt、Summary Sheet、Control Number 且完整 PDF 不超过 25 页。
- 51MCM 强制
51mcmthesis电子提交模式，并移除学校、成员、邮箱和电话字段。 - CUMCM 使用
cumcmthesis，提交前按当届官方通知复核匿名、编号页、页数与附件。 - 规则变化时先更新机器 Profile 和对应竞赛规则文档，再运行工作流。
- 论文章节目录允许项目统一采用
章节/或模板原生采用sections/；主文件引用路径必须与磁盘目录一致。 - 门禁会读取实际竞赛
.cls，检查纸张、页边距、基础字号、页眉或摘要页等关键合同，不能仅靠文档类名称通过。
2. 按阶段渐进加载
- 严禁一次载入全部研究协议。
- 先用
stage_executor.py current确认活动阶段，再查阅相应实施协议。 - 对当前阶段之外的资料，只有在当前协议明确引用时才可加载。
3. 落实主控与工作角色分层
- 主控角色负责阶段编排、运行状态落盘、门禁收束、核验点推进和跨阶段一致性维护。
- 工作角色承担当前阶段内的原子工作项，例如题意拆解、方案质询、程序实现、图形与表格制作以及论文组装。
- 若运行环境不支撑并行工作角色，可按角色次序实施；交付路线、复核独立性和阶段边界仍须保持不变。
4. 保持门禁契约稳定
- 全部阶段都务必核验原产物合同。
computational-realization务必同时核验程序/主程序.py、计算结果.md和伴随结果文件。evidence-visualization/systems-diagramming共用图表/图表引用.tex，避免误删前一步产物。manuscript-synthesis此后再转入delivery-assurance，不可跳。
5. 依据证据恢复运行
- 续跑前先看：
状态/工作流状态.json状态/事件日志.jsonlpython scripts/stage_executor.py status --workspace .
- 若某阶段被标记完成，但产物缺失或尺寸不达标，应重新
begin并补做，而不是盲目向前推进。
项目组成
assets/workflow_manifest.json: 研究工作流机器可读清单assets/competition_profiles.json: 三类竞赛的机器可读模板与合规硬约束references/workflow-map.md: 人类可读的过程总图和阶段映射references/gate-matrix.md: 各个阶段的机器门禁与人工复核要点references/phase-control.md:baseline -> enhancement模式切换准则references/subagent-architecture.md: 主控、工作与复核角色的闭环契约references/enhancement-operations.md: 增强模式的返工与质量提升方法references/cumcm-official-notes.md: 国赛官网准则与时间节点摘录references/guide-*.md: 各项研究工作的主控与执行角色指南references/stage_protocols/: 研究工作协议及配套参考资料与模板references/runtime_reference/:workflow_engine.py与agent_runner.py的运行参考实现scripts/workspace_init.py: 初始化运行时工作区scripts/stage_executor.py: 运行状态机、门禁、核验点、断点续跑入口scripts/gate_contracts.py: 研究工作流的机器门禁引擎scripts/pipeline_manager.py: 总入口、下一步提示、阶段切换scripts/baseline_smoke.py: 完整研究工作流基线冒烟校验scripts/stateful_smoke.py: 核验 checkpoint rerun、rework 传播、resume 行为scripts/enhancement_audit.py: enhancement 阶段的返工与增强推荐入口scripts/championship_review.py: 冠军模式多轮审稿、修订和终版论文回写入口scripts/build_code_appendix.py: 生成源码哈希清单，并从当前程序确定性构建 LaTeX/DOCX 代码附录references/championship-review-method.md: 冠军审稿评分、攻击策略与回退知识库scripts/state_store.py: JSON 运行状态存储assets/shared-scripts/: 原共享脚本assets/templates/: 原论文模板
最终交付说明
完成全部必要阶段后，项目应形成一套能够相互印证的研究成果，包括问题分析、模型报告、可运行程序、结构化结果、论文图形、完整 LaTeX 源稿和最终 PDF。冠军模式启用时，还应保留逐轮审稿报告、修订计划、修改后论文和修复验证记录。
最终论文必须满足以下原则：研究过程可解释，关键结果可复现，论文结论有证据支撑，图表与正文保持一致，格式符合目标竞赛要求，并且不存在已知的 P0 级质量问题。
© WuXinbo-bo, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Files
SKILL.md and 162 other files (scripts, references, assets) in the repository root of WuXinbo-bo/Math-model-skills.
- SKILL.md
- .gitignore
- ENVIRONMENT.md
- LICENSE
- README.md
- agents/openai.yaml
- assets/branding/banner.svg
- assets/branding/logo.svg
- assets/branding/sponsor-chuangshi-xinyuan.jpg
- assets/competition_profiles.json
- assets/shared-scripts/NotoSansSC-Regular.ttf
- assets/shared-scripts/china_provinces.geojson
- assets/shared-scripts/compile_check.sh
- assets/shared-scripts/compile_utils.sh
- assets/shared-scripts/demo_all_templates.tex
- assets/shared-scripts/demo_roadmap_competition.tex
- assets/shared-scripts/demo_roadmap_research_pipeline.tex
- … and 146 more
Open the folder on GitHubat commit ee8a616
Compare with similar skills
Meta-model-agent Math Modeling Pipeline next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Meta-model-agent Math Modeling Pipeline this skillWuXinbo-bo/Math-model-skills | 111 | — | ~2.3k | Automated safety check: Pass | MIT | |
| Modeling Code and Result Contractsyushui2022/MathModel-Skill | 454 | — | ~1.4k | Automated safety check: Pass | MIT | |
| Nature Data AvailabilityYuan1z0825/nature-skills | 47k | — | ~957 | Automated safety check: Pass | Apache-2.0 | |
| Backward Traceabilitylingzhi227/agent-research-skills | 390 | — | ~802 | Automated safety check: Pass | None | |
| Paper Pipeline Assemblylingzhi227/agent-research-skills | 390 | — | ~971 | Automated safety check: Pass | None | |
| Icml Reviewersundial-org/skills | 153 | — | ~2.4k | Automated safety check: Pass | None |
Similar skills
Nature Data Availability
Yuan1z0825/nature-skills
Drafts or audits data and code availability statements, dataset access routes, repository plans and FAIR metadata for Nature-style manuscripts.
Backward Traceability
lingzhi227/agent-research-skills
Makes each number in a LaTeX paper link back to the code line that produced it, using hypertarget and hyperlink tags and compile-time `\num` formulas.
Paper Pipeline Assembly
lingzhi227/agent-research-skills
Orchestrates a research paper from literature review through code, experiments, figures, tables, writing and review, with state passed between phases and checkpoints to resume.
Icml Reviewer
sundial-org/skills
Paper reviewer that evaluates machine learning research projects following official ICML reviewer guidelines.
Scientific Workflow Tools
DrugClaw/DrugClaw
Research-method workflow guide for hypothesis framing, peer-review style critique, reproducibility planning, study-design checks, and scientific-writing structure.
More from WuXinbo-bo/Math-model-skills
All 9 skills in this repoModeling Competition Computation Stage
WuXinbo-bo/Math-model-skills
Pipeline stage that turns a mathematical modeling report into runnable programs per sub-question, frozen numerical results and reviewable evidence files.
Math Model Formulation Stage
WuXinbo-bo/Math-model-skills
Stage protocol for a math-modeling pipeline that turns a problem analysis into a unified mathematical mechanism, formulas, a solution route and a validation plan.
Championship Mode Paper Review
WuXinbo-bo/Math-model-skills
Runs several independent review rounds on a finished math-modeling competition paper and rewrites it until it clears a fixed score and defect bar.
Math Modeling Problem Analysis
WuXinbo-bo/Math-model-skills
Breaks a math modeling competition problem statement into sub-problems, variables, constraints and evidence needs, and writes the result to a problem analysis file.
Delivery Assurance
WuXinbo-bo/Math-model-skills
Meta-model-agent 完成编译、版式、匿名、页数、引用与提交前合规验收。适用于提交质量验收. An agent skill from WuXinbo-bo/Math-model-skills.
Evidence Visualization
WuXinbo-bo/Math-model-skills
Meta-model-agent 将计算成果转换为可发表的数据图形、表格和排版引用。适用于证据图谱构建. An agent skill from WuXinbo-bo/Math-model-skills.
Related
Categories
Questions about Meta-model-agent Math Modeling Pipeline
What does Meta-model-agent Math Modeling Pipeline do?
Runs a staged pipeline for mathematical modeling research and contest papers, from problem analysis and computation to paper writing, review rounds and submission checks. Meta-model-agent links problem analysis, modeling, program solving, result checks, figures, paper writing and submission checks into one resumable, auditable chain, aimed at contests such as CUMCM, 51MCM and MCM/ICM and similar research tasks. Work runs through seven stages (including DISCOVERY, FORMULATION, COMPUTATION and MANUSCRIPT) with evidence gates, and a project can be started or resumed.
When should I use Meta-model-agent Math Modeling Pipeline?
Meta-model-agent Math Modeling Pipeline fits situations like: starting or resuming a mathematical modeling contest project; fixing a weak model, program or evidence chain before the paper is finalized; running mock review rounds on a modeling paper before submission; checking that every number and figure in the paper traces to result files.
How do I install Meta-model-agent Math Modeling Pipeline in Claude Code?
Run `npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a claude-code`. Or copy the skill folder (the WuXinbo-bo/Math-model-skills repository) into .claude/skills/meta-model-agent in your project. Claude Code loads it when a task matches its description.
How do I install Meta-model-agent Math Modeling Pipeline in Codex?
Run `npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a codex`. Or copy the skill folder (the WuXinbo-bo/Math-model-skills repository) into .agents/skills/meta-model-agent in your project. Codex loads it when a task matches its description.
Can I use Meta-model-agent Math Modeling Pipeline in Cursor, Gemini CLI or GitHub Copilot?
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add WuXinbo-bo/Math-model-skills --skill meta-model-agent -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/meta-model-agent, .gemini/skills/meta-model-agent, .github/skills/meta-model-agent and .opencode/skills/meta-model-agent in your project.
What does Meta-model-agent Math Modeling Pipeline need to run?
Going by SKILL.md and its folder, Meta-model-agent Math Modeling Pipeline needs a shell for the scripts in its folder and the command-line tools its instructions call (python). Our summary lists: An AI tool that can read skills, access the workspace and run local commands.
Does Meta-model-agent Math Modeling Pipeline access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Meta-model-agent Math Modeling Pipeline safe to install?
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
What licence does Meta-model-agent Math Modeling Pipeline use?
Meta-model-agent Math Modeling Pipeline is published under the MIT licence (from the LICENSE file in the skill folder). It allows redistribution, so the full SKILL.md is shown on this page.
How many tokens does Meta-model-agent Math Modeling Pipeline use?
About 2.3k tokens (SKILL.md is roughly 9.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 19k tokens, read only when the agent opens those files.
What are the alternatives to Meta-model-agent Math Modeling Pipeline?
Skills that share tags, products or a category with Meta-model-agent Math Modeling Pipeline: Modeling Code and Result Contracts (yushui2022/MathModel-Skill, 454 stars), Nature Data Availability (Yuan1z0825/nature-skills, 47k stars), Backward Traceability (lingzhi227/agent-research-skills, 390 stars) and Paper Pipeline Assembly (lingzhi227/agent-research-skills, 390 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Meta-model-agent Math Modeling Pipeline?
WuXinbo-bo (a GitHub user) maintains it in WuXinbo-bo/Math-model-skills, which has 111 GitHub stars. The repository holds 9 skills in this directory. The repository was last updated on September 21, 2026.
Source: WuXinbo-bo/Math-model-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.