AGENT SKILL
ResearchThree Chinese-first research skills that run on both Claude Code and Codex
zLanqing/codex-claude-academic-skills
4,117 stars, MIT. research-writing handles drafting, revising and polishing paper sections plus rebuttals; office-academic generates and edits academic Word and PPT at the OOXML layer (literature reports, group meetings, proposals, defences); scientific-toolkit covers MATLAB/Python simulation, signal processing, statistics, machine learning and paper figures, with literature search across arXiv, PubMed, CrossRef, Semantic Scholar and OpenAlex.
Our takeThe part worth copying is that academic integrity is written as instructions: no fabricated data, DOIs or journals, and every piece of content is classified as quotation, confirmed fact, reasonable inference or writing suggestion, so the author can see at a glance which sentences have a source. The real danger in academic writing is not weak prose but a model writing inference as though it were established fact, and that four-way split blocks exactly that. The three skills also compose: figures, then body text, then defence slides.
git clone https://github.com/zLanqing/codex-claude-academic-skills && cp -r codex-claude-academic-skills/research-writing-skill ~/.claude/skills/Chinese-first, and it runs on both Claude Code and Codex
zLanqing/codex-claude-academic-skills (4,117 stars / 228 forks, MIT) is three skills built for researchers working in Chinese, covering paper writing, academic Office documents and scientific computing — one set that runs on both Claude Code and Codex. Chinese-first is rare in an English-dominated skill ecosystem.
| Skill | Purpose | Typical use |
|---|---|---|
research-writing-skill | Write, revise and polish paper sections | Body, abstract, intro, methods, experiments, rebuttal |
office-academic-skill | Generate and edit academic Word/PPT | Literature reading reports, group-meeting / proposal / defence slides |
scientific-toolkit-skill | Scientific computing and data analysis | MATLAB/Python simulation, signal processing, statistics, ML, paper figures |
research-writing: academic integrity encoded as instructions
Its rules are not about writing prettily, they are about not overstepping: do not fabricate data, DOIs or journals; classify content explicitly into four kinds — quoted source / confirmed fact / reasonable inference / writing suggestion — so the author can see at a glance which lines the model produced and which have a source; avoid vague words like "significant" or "advanced" with no data behind them. Its references/ add section_rhetorical_moves, writing_checklists and figure_templates.
That four-way classification is the most stealable idea here. In academic writing the real danger is not poor prose; it is the model writing an inference as though it were a confirmed fact.
office-academic: edits at the OOXML layer, not by template
It converts PDFs into literature reading reports and produces group-meeting / proposal / defence slides by editing docx/pptx at the OOXML level, with its own XSD schema library — it manipulates the Office XML directly rather than pouring content into a fixed python-docx-style template. The slide rules emphasise action titles (a conclusion sentence, not a noun phrase like "Results"). That is the same approach as Anthropic's official pptx skill (unzip → edit XML → zip), specialised for Chinese academic work.
scientific-toolkit: wide compute and retrieval coverage
On compute: MATLAB/Octave, plus NumPy/SciPy/pandas/matplotlib/scikit-learn/statsmodels/SymPy/pymoo/simpy/QuTiP/pymatgen/TimesFM/NetworkX/Astropy. On retrieval: arXiv, PubMed, CrossRef, Semantic Scholar, OpenAlex, and DOI → BibTeX. For a workflow that runs simulations and then needs publication figures, the tool surface is essentially complete.
The three compose into chains
Paper writing:
scientific-toolkit (analyse data, produce figures)
→ research-writing (write the body)
→ office-academic (generate the defence slides)
Literature reading:
office-academic (PDF → Word reading report) → office-academic (group-meeting slides)
Simulation research:
scientific-toolkit (MATLAB/Python simulation + figures) → research-writing (methods + experiments)Install (mind the placeholders)
The README's install commands contain placeholders like your-username — do not copy them verbatim. In practice: clone the repo and copy the three skill directories into ~/.claude/skills/ (Claude Code) or ~/.codex/skills/ (Codex), or into a project-level directory. The three skills are independent, so copy only what you need.
git clone https://github.com/zLanqing/codex-claude-academic-skills
cp -r codex-claude-academic-skills/research-writing-skill ~/.claude/skills/
# Codex users: use ~/.codex/skills/; copy the other two as needed