Skip to content

Latest commit

 

History

4 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Management Science & Engineering Academic Writing Skill

English | 简体中文 | 日本語 | 한국어

A comprehensive OpenCode/Claude Code/Codex skill for writing management science and operations research papers — from topic selection through modeling, derivation, numerical experiments, to a complete first draft.

Covers 13 flagship MS&E journals across two tiers: MS, OR, MSOM, POM (UTD-24) and TS, TRB, DS, OMEGA, TRE, EJOR, IJPE, IJPR, C&IE.

What This Skill Does

This skill guides AI coding agents through the full 0-to-Draft Pipeline for MS&E papers:

Stage What It Produces Key Output
Stage 1: Topic Positioning Gap table, journal recommendation, contribution statement Which journal + what's new
Stage 2: Model Building Notation system, assumption framework, mathematical formulation §3 Model section draft
Stage 3: Derivation & Analysis Lemma→Theorem→Corollary proof chain, comparative statics §4 Analysis section draft
Stage 4: Numerical Experiments Parameter calibration, sensitivity analysis, counterfactuals §5-6 Experiments draft
Stage 5: Writing & Assembly Introduction, Lit Review, Managerial Insights, Abstract, formatting Complete first draft

The skill is domain-specific: it encodes the conventions, expectations, and stylistic norms of MS&E journals that generic writing skills (like scientific-writing) do not cover — such as model-presentation structure, proof hierarchy, managerial insight SAR framework, and journal-specific reviewer expectations.


Installation

1. Clone the Repository

git clone https://github.com/liyuanbo1024/management-science-writing.git

2. Install for Your AI Agent

Choose your platform below.

OpenCode

Copy the skill directory to your OpenCode skills folder:

# Linux/macOS
cp -r management-science-writing ~/.config/opencode/skills/

# Windows (PowerShell)
Copy-Item -Recurse management-science-writing "$env:USERPROFILE\.config\opencode\skills\"

Or register a custom skills path in ~/.config/opencode/opencode.json:

{
  "skills": {
    "paths": [
      "~/.config/opencode/skills",
      "/path/to/your/cloned/management-science-writing"
    ]
  }
}

Then trigger with: /management-science-writing or just describe your task naturally ("I need to write an MSOM paper on...").

Claude Code (Anthropic)

# Linux/macOS
cp -r management-science-writing ~/.claude/skills/

# Windows (PowerShell)
Copy-Item -Recurse management-science-writing "$env:USERPROFILE\.claude\skills\"

Claude Code auto-discovers skills in ~/.claude/skills/. The skill activates when you mention:

  • Writing for MS, OR, MSOM, POM, EJOR, or any covered journal
  • MS&E modeling, proof writing, numerical experiments, managerial insights
  • Any phrase matching the skill description triggers

Codex (OpenAI)

# Linux/macOS
cp -r management-science-writing ~/.agents/skills/

# Windows (PowerShell)
Copy-Item -Recurse management-science-writing "$env:USERPROFILE\.agents\skills\"

Cursor / Windsurf

These editors use the same skill format. Copy to your configured skills directory, typically:

# Cursor
cp -r management-science-writing ~/.cursor/skills/

# Windsurf
cp -r management-science-writing ~/.windsurf/skills/

Manual (Any Agent)

If your agent supports custom markdown-based skills, you can:

  1. Point the agent's skills path to the cloned directory
  2. Or directly reference SKILL.md in your agent's configuration
  3. Or concatenate SKILL.md + relevant reference files into a single prompt

The skill format follows the agentskills.io specification with YAML frontmatter (name + description fields).


How to Use

Quick Start

Once installed, trigger the skill by describing your task naturally. The agent will detect the skill automatically. Examples:

"I'm writing a paper on dynamic pricing with consumer loyalty.
Help me position the contribution and choose a journal."

"My duopoly model is set up. Help me derive the MPE and prove
structural properties."

"I have all my theorems. Design numerical experiments calibrated
to industry data."

"Write the full paper draft for MSOM submission."

Pipeline Mode

For a complete 0-to-draft workflow, say:

"Take me through the full MS&E pipeline. My topic is [describe your topic]."

The agent will:

  1. Load the skill and assess your current stage
  2. Execute Stage 1 (positioning) → get confirmation
  3. Proceed to Stage 2 (modeling) → get confirmation
  4. Continue through Stage 5 (full draft)
  5. Output a complete LaTeX manuscript with INFORMS formatting

Stage-Specific Mode

Jump to any stage:

Trigger Phrase Stage
"I have a research idea about..." Stage 1: Positioning
"Help me design the mathematical model" Stage 2: Modeling
"I need to derive/prove..." Stage 3: Derivation
"Design my numerical experiments" Stage 4: Experiments
"Write the full paper" Stage 5: Assembly

Reference-Only Mode

The skill also works as a passive reference. Just ask:

"What are MS's reviewer expectations for proof completeness?"
"How should I structure the managerial insights for an EJOR submission?"
"What's the standard notation convention for inventory models?"

File Structure

management-science-writing/
├── SKILL.md                              # Main skill file (578 lines)
│   ├── Journal Selection Quick Reference
│   ├── 0-to-Draft Pipeline (5 stages)
│   ├── Rejection Reasons Checklist
│   └── Cross-References to all reference files
│
├── references/
│   ├── journal-characteristics.md        # 13 journals: detailed profiles
│   ├── modeling-conventions.md           # Notation, assumptions, model types
│   ├── proof-derivation-guide.md         # Proof hierarchy, placement, techniques
│   ├── algorithm-solving-guide.md        # Pseudocode, complexity, heuristics
│   ├── numerical-experiments.md          # Experiment design, calibration
│   ├── managerial-insights.md            # SAR framework, MI writing patterns
│   ├── reviewer-expectations.md          # What reviewers look for, rebuttal tips
│   └── writing-patterns.md               # 6 reusable writing patterns
│
├── assets/
│   └── theorem-structure-reference.md    # Generic theorem hierarchy example
│
├── examples/
│   ├── manuscript_template.tex           # INFORMS-style LaTeX template
│   └── numerical_experiments.py          # Newsvendor experiment template
│
├── README.md                             # This file
└── LICENSE                               # MIT License

Covered Journals

Tier 1 (UTD-24)

Journal Abbreviation Focus
Management Science MS Broad management, high rigor, significant contribution
Operations Research OR Methodological depth, optimization focus
M&SOM MSOM Empirical OM, behavioral operations
Production & Oper. Mgmt POM Data-driven, broad OM scope

Tier 2 (Strong Field Journals)

Journal Focus
Transportation Science (TS) Transportation systems, network models
Transportation Research B (TRB) Transportation methodology
Decision Sciences (DS) Decision-making, MCDM
OMEGA Concise, high-impact, managerial
Transportation Research E (TRE) Logistics & transportation
European J. Operational Research (EJOR) Broad methodology, inclusive
Int. J. Production Economics (IJPE) Production economics, supply chain
Int. J. Production Research (IJPR) Production engineering
Computers & Industrial Eng. (C&IE) Computational methods, IE

Customization

Adding a Journal

Edit references/journal-characteristics.md and add a new section following the template:

### Your Journal Name

**Publisher**: ...
**Focus**: ...
**Key characteristics**:
- ...
**Paper structure**: ...
**What it values most**:
1. ...

Adjusting Parameter Defaults

The numerical experiments use Parameters dataclass defaults. Modify in examples/numerical_experiments.py:

@dataclass
class Parameters:
    tau: float = 0.60    # Adjust to your domain
    c: float = 5.0       # Unit cost
    # ... etc

Creating a New Paper Template

  1. Copy examples/manuscript.tex as your starting point
  2. Replace the title, author, and abstract
  3. Fill in your model, theorems, and experiments
  4. Compile with: xelatex manuscript.tex (two passes)

Example Output

The examples/ directory contains two reference files:

  1. manuscript_template.tex: A clean INFORMS-style LaTeX template with placeholder sections following the MS&E paper structure (Model → Analysis → Experiments → Managerial Insights). Start here when writing a new paper.

  2. numerical_experiments.py: A runnable Python template illustrating the numerical experiment conventions from references/numerical-experiments.md—parameter tables, sensitivity analysis, benchmark comparisons. Uses a simple Newsvendor model so you can run it immediately.

The assets/ directory contains theorem-structure-reference.md, a generic guide to the theorem hierarchy (Lemma → Proposition → Theorem → Corollary) with proof-writing conventions.


Requirements

  • AI Agent: OpenCode, Claude Code, Codex, Cursor, Windsurf, or any agent supporting the agentskills.io format
  • For LaTeX compilation (examples): XeLaTeX or pdfLaTeX with amsmath, booktabs, natbib, geometry, setspace, enumitem, hyperref, caption
  • For numerical experiments (examples): Python 3.10+ with numpy

Contributing

Contributions are welcome. Areas where help is especially valuable:

  • Journal profiles: Detailed editorial statements and reviewer expectations for journals not yet covered
  • Writing patterns: Additional reusable patterns distilled from published MS&E papers
  • LaTeX templates: Official style files for specific journals
  • Calibration examples: Numerical experiments calibrated to different industries/markets
  • Multi-language support: Chinese/Korean/Japanese MS&E writing conventions

Please open an issue or pull request on GitHub.


License

MIT License — see LICENSE file.


Acknowledgments

Built using the agentskills.io specification and the skill authoring methodology from OpenCode. The MS&E domain knowledge draws on editorial statements from INFORMS journals (MS, OR, MSOM, TS) and published style guides from the operations research community.

About

No description, website, or topics provided.

Resources

Stars

50 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors