This guide explains how to publish, install, maintain, and responsibly use this Codex skills package.
This repository contains two Codex skills:
social-science-reviewerbilingual-social-science-writer
The skills are designed for social-science research workflows: manuscript review, research design diagnosis, literature review structuring, bilingual paper drafting, research logs, evidence mapping, and reproducible R/Quarto project scaffolding.
They do not provide data, citations, findings, interviews, ethics approvals, or journal facts. They provide structure, review logic, templates, and workflow discipline.
Recommended public repository layout:
.
├── README.md
├── README.zh-CN.md
├── INSTALL.md
├── OPEN_SOURCE_GUIDE.md
├── OPEN_SOURCE_GUIDE.zh-CN.md
├── SKILLS_EXPLAINED.md
├── SKILLS_EXPLAINED.zh-CN.md
├── LICENSE
└── skills/
├── social-science-reviewer/
│ ├── SKILL.md
│ ├── agents/openai.yaml
│ └── references/review-rubrics.md
└── bilingual-social-science-writer/
├── SKILL.md
├── agents/openai.yaml
├── references/
├── scripts/create_project.py
└── assets/bilingual-social-science-project/
Keep public-facing documentation at the repository root. Keep skill folders focused on files Codex needs to use the skills.
Copy the two skill folders into your Codex skills directory:
Copy-Item -Recurse .\skills\social-science-reviewer "$env:USERPROFILE\.codex\skills\"
Copy-Item -Recurse .\skills\bilingual-social-science-writer "$env:USERPROFILE\.codex\skills\"Then restart Codex.
Use the reviewer skill:
Use social-science-reviewer to review this methods section and identify rejection-level risks.
Use the bilingual writer skill:
Use bilingual-social-science-writer to create a bilingual Chinese/English social-science article project.
Generate a project manually from the script:
python .\skills\bilingual-social-science-writer\scripts\create_project.py --target . --name my-paperPurpose:
- Diagnose research questions, theory, literature positioning, methods, evidence, ethics, and submission readiness.
Best used for:
- manuscript review
- thesis chapter diagnosis
- proposal assessment
- literature review critique
- response-to-reviewers planning
- method/evidence alignment checks
Do not use it as a mere copyediting tool. It is designed to identify substantive academic risks before language polish.
Purpose:
- Create a complete bilingual writing and research-production scaffold.
Best used for:
- Chinese/English manuscript skeletons
- literature review planning
- research logs
- data dictionaries
- evidence-claim mapping
- ethics and open-science checklists
- reproducible R/Quarto workflows
The generated project includes placeholder scripts and templates. Replace illustrative material with real research materials before using any result in a manuscript.
Do not use these skills to:
- fabricate data
- fabricate interviews
- invent citations
- invent DOI or journal metadata
- fabricate statistical results
- hide methodological weaknesses
- claim ethics approval without evidence
- present simulated data as real empirical evidence
Acceptable uses:
- generate templates
- structure arguments
- identify gaps
- downgrade unsupported claims
- create reproducible project scaffolds
- document analysis decisions
- prepare transparent limitations
When editing a skill:
- Edit
SKILL.mdonly for core instructions. - Put detailed criteria in
references/. - Put reusable files in
assets/. - Put deterministic automation in
scripts/. - Validate the skill after changes.
Run validation:
python path\to\quick_validate.py .\skills\social-science-reviewer
python path\to\quick_validate.py .\skills\bilingual-social-science-writerIf your Python environment lacks yaml, install PyYAML or use an environment that already has it.
Recommended versioning:
v0.1.0: initial public release- patch version: wording fixes, template corrections
- minor version: new templates or references
- major version: breaking changes to project structure or skill behavior
Recommended release artifacts:
codex-social-science-skills-bilingual.zipcodex-social-science-skills-en.zipcodex-social-science-skills-zh-CN.zip
Good contributions:
- improve review rubrics
- add reporting-standard checklists
- strengthen bilingual writing templates
- improve project scaffolding scripts
- add discipline-specific references without bloating
SKILL.md
Avoid:
- adding fake examples as if real
- adding large auxiliary documentation inside skill folders
- adding personal data
- adding copyrighted templates without permission
- weakening academic integrity safeguards
This package uses the MIT License. You may use, modify, and redistribute it, subject to the license terms.