重新定义人机协作边界：Claude Code 担任 Autopilot（自动驾驶仪），人类担任 Copilot
随着生成式人工智能的快速发展，数学建模竞赛中越来越多的环节可以被 AI 承接——文献检索、数学推导、代码实现、LaTeX 排版，乃至完整的论文写作。这一趋势在推动效率提升的同时，也引发了一个根本性问题：
当 AI 能够独立完成大部分工作时，"人机协作"的边界应该划在哪里？
本项目的出发点，正是跨越这条边界的一次试验。
传统模式中，AI 是工具，人类是决策者。参赛者主导所有环节，AI 仅作为 Copilot（副驾驶）辅助执行。AutoMCM-Pro 反转了这个关系：Claude Code 智能体承担 Autopilot（自动驾驶仪） 的角色，自主驱动整条建模流水线——从解读赛题、搜索文献、构建数学模型、编写并自证求解代码，到最终生成完整的 LaTeX 论文；人类退居 Copilot 位置，在关键节点介入审阅、提供战略性指导，或在需要时完全接管某个环节的决策权。
这种模式并非让人类"袖手旁观"，而是将人类的精力集中在 AI 难以替代的部分：
- 判断建模方向是否符合题意与现实物理
- 发现模型假设中的潜在缺陷
- 提供领域直觉与经验性修正
/auto-mcm、/cumcm-master 和 /mcm-master 三个 Skill 共同构成一条从零到论文的全自动建模流水线，覆盖：
| 阶段 | AI 完成的工作 |
|---|---|
| 题目解读 | 读取 PDF/文本题目，识别问题类型、约束、数据特征 |
| 文献调研 | 联网搜索相关建模方法，提炼文献依据，记录 DOI |
| 数据预处理 | 编写并运行 EDA 脚本，生成数据分布图，完成缺失值/异常值处理 |
| 模型构建 | 选择数学工具（优化/微分方程/统计/图论），编写求解代码，运行至无误 |
| 强制自证 | 为每个求解脚本配写独立验证脚本，覆盖约束满足、物理合理性、数值稳定性 |
| 敏感性分析 | 分析关键参数扰动对结果的影响，验证模型鲁棒性 |
| LaTeX 论文 | 按竞赛格式（国赛/美赛模板）逐章撰写，插入图表，完成引用 |
| PDF 编译 | 调用 xelatex/pdflatex 编译，输出最终 PDF |
整个过程完全在本地运行，所有中间状态通过 state/pipeline.json 持久化，支持中断续做。
AI 完全自主推进，每个阶段完成后立即写入自评报告并自动前进到下一阶段。人类可随时查看 state/review_request.md 中的自评记录，但不需要主动操作。
适用场景：时间紧迫时快速出解，或对 AI 能力充分信任的情况。
AI 在开始核心建模之前会暂停，等待人类在 human_intervention.md 中填写精确的数学规格（模型类型、决策变量、目标函数、约束条件、求解器）。之后 AI 100% 按照该规格实现，不自行添加任何内容。每个阶段完成后同样需要人类填写 [APPROVED] 才能继续。
适用场景：人类已有明确建模思路，需要精确控制模型细节；或希望对每个阶段保持严格的人工审核。
两种模式可在同一次竞赛中组合使用：数据预处理阶段用 AP 快速完成，核心建模阶段切换至 MANUAL 确保数学严谨性。
每个阶段的状态通过 state/pipeline.json 持久化，形成类似 CI/CD 的审查流程：
not_started → in_progress → pending_review → approved
↘ rework → in_progress（重做）
可随时运行 python scripts/pipeline_manager.py status 查看完整状态看板。
所有求解代码（src/models/）都必须配有对应的验证脚本（src/verifications/verify_*.py）。验证内容包括：
- 目标函数值与约束满足性核验
- 边界条件与物理可行性检查
- 数值稳定性与收敛性验证
所有验证项必须全部通过（✓ PASS），结果才能被引用进论文。 任何 ✗ FAIL 都会强制回退到模型构建阶段。
基于 FastAPI + WebSocket 的实时 Web UI（http://localhost:8080），渲染 AI 在 memory/thought_process.md 中记录的推理过程，支持 KaTeX 数学公式、代码高亮和流水线进度看板。
git clone https://github.com/RealSeaberry/AutoMCM-Pro
cd AutoMCM-Pro
bash install.sh安装完成后，/auto-mcm、/cumcm-master、/mcm-master 三个命令在任意目录的 Claude Code 会话中均可使用。
# 拷贝工作目录（保持源码干净，每次竞赛独立拷贝）
cp -r AutoMCM-Pro /path/to/my-contest-2025
cd /path/to/my-contest-2025
# 放入题目文件和附件数据
cp problem.pdf CUMCM_Workspace/data/
cp data.xlsx CUMCM_Workspace/data/
# 初始化流水线（选择模式和竞赛类型）
python scripts/pipeline_manager.py init --mode AP --contest CUMCMclaude # 启动 Claude Code
/auto-mcm # 或 /cumcm-master（国赛）、/mcm-master（美赛）Skill 启动后会自动读取流水线状态，根据 AP/MANUAL 模式执行对应流程。
若使用 MANUAL 模式，AI 在进入 model_build 阶段前会暂停并提示你填写 state/human_intervention.md：
### 问题一
- **模型类型**: 非线性规划
- **决策变量**: θ（投放方向角，范围 0~360°），v（初速度，70~140 m/s）
- **目标函数**: max T_shield(θ, v)（有效遮蔽时长）
- **约束条件**:
- 70 ≤ v ≤ 140
- 无人机等高度匀速直线飞行
- **求解方法**: scipy.optimize.minimize, method='SLSQP'
- **特殊处理**: 网格搜索初始化，避免局部最优填写完毕后在终端输入「继续」即可唤醒 AI 开始编码。
AP 模式下 AI 自动运行，但你可以在任意时刻：
- 查看
state/review_request.md了解当前阶段的自评报告 - 查看
state/human_intervention.md了解历史自评记录 - 直接向 AI 输入修改意见，AI 会立即响应
说明：以下演示中，流水线全程在 AP（Autopilot）模式下自动运行。人类的全部输入只有以下 5 条提示词。我们认为这是一个接近零干预的基线输出——通过更多的战略性干预，完全可以获得质量更高的论文。
流水线进行中（problem_analysis 阶段完成后）：
-
[APPROVED]（确认 AI 对题目的理解正确，同意继续数据预处理） -
[APPROVED]（确认数据预处理方案，同意进入建模阶段）
流水线完成后（所有 11 个阶段均已 AI 自动推进完毕）：
-
请增加更多的图表和可视化，并且确保可视化的中文文字都能正确显示 -
能否编译成pdf -
请在论文中增加描述性的语言，增加文字长度。请美化可视化。修改后再次编译
流水线共运行约 1小时34分钟，完成 11 个流水线阶段，通过 144 项代码验证，全程 AI 自动推进（仅前两条为流水线内必要的人工确认节点）。
具体建模过程、思考链、生成图表和最终论文详见 demo/ 目录。建议直接阅读 demo/CUMCM_Workspace/latex/main.tex（或编译后的 PDF）和 demo/CUMCM_Workspace/memory/thought_process.md，感受 AI 的完整推理过程。
- Claude Code 已安装（
claudeCLI 可用） - Python 3.10+
- 基础依赖：
pip install pdfplumber scipy numpy matplotlib pandas openpyxl - 可选：TeX Live 或 MiKTeX（LaTeX 编译，也可用 Docker 替代）
git clone https://github.com/RealSeaberry/AutoMCM-Pro
cd AutoMCM-Pro
bash install.sh # 符号链接安装（推荐，git pull 自动更新）
bash install.sh --copy # 文件拷贝安装（无 git 环境）
bash install.sh --check # 检查安装状态cd AutoMCM-Pro
git pull # 符号链接安装下，Skill 自动生效，无需重装bash uninstall.shCUMCM-MASTER/
├── .claude/skills/
│ ├── auto-mcm/SKILL.md # 统一入口（AP/MANUAL 双模式）
│ ├── cumcm-master/SKILL.md # 国赛专用
│ └── mcm-master/SKILL.md # 美赛专用
├── AutoMCM_SOP.md # 操作准则（Skill 行为的最终权威）
├── install.sh / uninstall.sh # 安装/卸载
├── init_gitops.sh # 交互式初始化引导
├── docker-compose.yml # 容器化环境（Python + TeX Live）
├── scripts/
│ ├── pipeline_manager.py # GitOps 状态机 CLI
│ ├── compile_pdf.py # 跨平台 LaTeX 编译（Python）
│ ├── setup_workspace.py # 工作区目录初始化
│ └── agent_memory_manager.py # 记忆管理工具
├── templates/
│ ├── latex_template.tex # CUMCM 论文模板（10节中文格式）
│ ├── mcm_template.tex # MCM/ICM 论文模板（mcmthesis）
│ └── mcm_memo_template.tex # MCM 实用性文件模板（单页 Memo）
├── demo/ # 2025国赛A题完整演示产物
└── CUMCM_Workspace/ # 运行时工作目录（每次竞赛独立拷贝）
├── data/ # 赛题与数据文件
├── src/models/ # 求解代码
├── src/verifications/ # 验证脚本（强制配套）
├── latex/ # LaTeX 源文件与图表
├── memory/ # 思考过程与推理链
├── state/ # GitOps 流水线状态
└── output/ # 最终 PDF 与结果表格
# 本地编译（需已安装 TeX Live 或 MiKTeX）
python scripts/compile_pdf.py # CUMCM
python scripts/compile_pdf.py --mode mcm --memo # MCM + Memo
# Docker 编译（无需本地 LaTeX）
docker-compose up -d cumcm-agent
docker exec -it cumcm-agent python scripts/compile_pdf.pyMIT License — 欢迎 Fork、二次开发和 Pull Request。
Redefining Human-AI Collaboration: Claude Code as Autopilot, Humans as Copilot
As generative AI rapidly advances, an increasing share of math modeling competition tasks can be delegated to AI: literature search, mathematical derivation, code implementation, LaTeX typesetting, and even full paper writing. This raises a fundamental question:
When AI can independently handle most of the work, where should the boundary of human-AI collaboration be drawn?
AutoMCM-Pro is an experiment in crossing that boundary.
In the traditional model, AI is a tool and humans are decision-makers. Competitors lead every step, with AI acting as a Copilot. AutoMCM-Pro inverts this relationship: the Claude Code agent takes the Autopilot role, autonomously driving the entire modeling pipeline — from interpreting the problem, searching literature, building mathematical models, writing and self-verifying solver code, to generating the complete LaTeX paper. Humans step into the Copilot seat, intervening at key checkpoints to review, provide strategic guidance, or take full control of specific decisions when needed.
This is not about humans "stepping aside." It's about concentrating human effort where it's irreplaceable:
- Judging whether the modeling direction aligns with the problem's intent and physical reality
- Spotting flaws in model assumptions
- Providing domain intuition and empirical corrections
Three Skills — /auto-mcm, /cumcm-master, and /mcm-master — form a complete zero-to-paper automated pipeline:
| Stage | What the AI Does |
|---|---|
| Problem Reading | Reads PDF/text problem, identifies problem type, constraints, data characteristics |
| Literature Research | Searches the web for relevant modeling methods, extracts references |
| Data Preprocessing | Writes and runs EDA scripts, generates distribution plots, handles missing/outlier values |
| Model Building | Selects mathematical tools (optimization/ODE/statistics/graph theory), writes solver code |
| Self-Verification | Writes an independent verification script for each solver: constraints, physical feasibility, numerical stability |
| Sensitivity Analysis | Analyzes how key parameter perturbations affect results, validates robustness |
| LaTeX Paper | Writes chapter by chapter in competition format (CUMCM or MCM template), inserts figures, completes citations |
| PDF Compilation | Compiles with xelatex/pdflatex, outputs the final PDF |
The entire process runs locally. All intermediate state is persisted via state/pipeline.json, supporting pause and resume.
The AI runs fully autonomously, completing each stage and then immediately writing a self-evaluation report and advancing to the next stage. Humans can review state/review_request.md at any time but are not required to act.
Best for: Time-constrained competitions, rapid prototyping, or when you trust the AI to handle the full run.
Before starting core modeling, the AI pauses and waits for the human to fill in a precise mathematical specification in human_intervention.md (model type, decision variables, objective function, constraints, solver). The AI then implements exactly that specification — no autonomous extensions. Each stage also requires a human [APPROVED] to continue.
Best for: When you have a clear modeling strategy and want precise control; or when you want strict human review at every checkpoint.
The two modes can be combined within a single competition: AP mode for fast data preprocessing, then MANUAL for the critical modeling stages.
Each stage's status is persisted in state/pipeline.json, forming a CI/CD-like review flow:
not_started → in_progress → pending_review → approved
↘ rework → in_progress (redo)
Run python scripts/pipeline_manager.py status at any time to see the full status dashboard.
All solver code (src/models/) must be paired with a corresponding verification script (src/verifications/verify_*.py), covering:
- Objective function value and constraint satisfaction
- Boundary conditions and physical feasibility
- Numerical stability and convergence
All verification checks must pass (✓ PASS) before results can be cited in the paper. Any ✗ FAIL forces a rollback to the model-building stage.
A FastAPI + WebSocket web UI at http://localhost:8080 renders the AI's thought process from memory/thought_process.md in real time, with KaTeX math, code highlighting, and a pipeline progress dashboard.
git clone https://github.com/RealSeaberry/AutoMCM-Pro
cd AutoMCM-Pro
bash install.shAfter installation, /auto-mcm, /cumcm-master, and /mcm-master are available in any Claude Code session from any directory.
cp -r AutoMCM-Pro /path/to/my-contest-2025
cd /path/to/my-contest-2025
cp problem.pdf CUMCM_Workspace/data/
cp data.xlsx CUMCM_Workspace/data/
python scripts/pipeline_manager.py init --mode AP --contest CUMCMclaude
/auto-mcm # or /cumcm-master (CUMCM) / /mcm-master (MCM/ICM)In AP mode the AI runs automatically, but you can intervene at any time by simply typing a message. The AI will respond immediately and adjust its course.
The AI pauses before model_build and prompts you to fill in human_intervention.md:
### Problem 1
- **Model type**: Non-linear programming
- **Decision variables**: θ (launch direction, 0–360°), v (speed, 70–140 m/s)
- **Objective**: max T_shield(θ, v)
- **Constraints**: 70 ≤ v ≤ 140
- **Solver**: scipy.optimize.minimize, method='SLSQP'Type "继续" (continue) in the terminal to wake the AI and start coding.
The pipeline ran in AP (Autopilot) mode throughout. The complete human input for the entire session is just the 5 prompts listed below. This is a near-zero-intervention baseline output. With more strategic interventions, significantly higher quality is achievable.
During the pipeline (after problem_analysis stage):
-
[APPROVED](Confirming the AI's problem understanding; allowing data preprocessing to begin) -
[APPROVED](Confirming the preprocessing plan; allowing modeling to begin)
After the pipeline completed (all 11 stages auto-advanced by AI):
-
请增加更多的图表和可视化，并且确保可视化的中文文字都能正确显示(Add more charts and visualizations; ensure Chinese text renders correctly) -
能否编译成pdf(Can you compile to PDF?) -
请在论文中增加描述性的语言，增加文字长度。请美化可视化。修改后再次编译(Add more descriptive language, increase paper length; polish visualizations; recompile)
The pipeline ran for approximately 1 hour 34 minutes, completing 11 stages and passing 144 verification checks. The first two prompts were necessary approval confirmations within the pipeline; the final three were optional post-pipeline refinements added after the paper was already complete.
For the full thought process, generated figures, and the final paper, see the demo/ directory. We recommend reading demo/CUMCM_Workspace/memory/thought_process.md and the compiled PDF to experience the AI's complete reasoning chain.
- Claude Code installed
- Python 3.10+
pip install pdfplumber scipy numpy matplotlib pandas openpyxl- Optional: TeX Live or MiKTeX (or use Docker)
git clone https://github.com/RealSeaberry/AutoMCM-Pro
cd AutoMCM-Pro
bash install.sh # symlink install (recommended — git pull auto-updates)
bash install.sh --copy # copy install (for non-git environments)
bash install.sh --check # check installation statusgit pull # symlink install: Skills update automaticallybash uninstall.shMIT License — Fork, extend, and contribute pull requests welcome.