187 installable AI agent skills for academic research, featuring a first-party Nature-style suite for scientific writing, manuscript polishing, peer review, publication figures, and statistics.
English · 简体中文 · 繁體中文 · 日本語 · 한국어
Qinyan Academic Skills equips Claude Code, Codex, Cursor, Gemini CLI, OpenClaw, and OpenCode with reusable research capabilities. The repository contains 187 skills across 18 domains, including a five-skill workflow for Nature-style scientific writing, academic editing, pre-submission review, scientific visualization, and statistical reporting.
Use the complete collection or install only the skill, category, tool, and scope that your workflow needs.
Move from a defensible scientific claim to a submission-ready manuscript with five focused, composable skills:
| Manuscript stage | Skill | Primary output |
|---|---|---|
| Argument and drafting | qinyan-nature-writing |
Evidence-led narrative, section drafts, title, abstract, and submission package |
| Structural and language revision | qinyan-nature-polishing |
Concise academic prose with meaning-preservation and anti-overclaim checks |
| Pre-submission assessment | qinyan-nature-review |
Prioritized, traceable review findings with stable issue identifiers |
| Figures and visual evidence | qinyan-nature-figures |
Reproducible multi-panel figures with export and preflight validation |
| Study design and statistics | qinyan-nature-statistics |
Estimand-aware analysis plans, reporting audits, and figure-ready results |
Install one skill:
curl -fsSL https://raw.githubusercontent.com/LeonChaoX/qinyan-academic-skills/main/install.sh \
| bash -s -- --skill qinyan-nature-writingOr explore the complete Nature skills workflow, including routing guidance, installation commands, quality gates, and example prompts.
Note
“Nature-style” describes editorial and scientific communication goals. This independent project is not affiliated with, endorsed by, or guaranteed acceptance by Nature Portfolio or Springer Nature.
| Capability | What you get |
|---|---|
| Research coverage | 187 skills spanning literature, writing, life sciences, AI, statistics, databases, and laboratory workflows |
| Agent portability | One installer for Claude Code, Cursor, Codex, Gemini CLI, OpenClaw, and OpenCode |
| Selective installation | Install the full library, one category, or one named skill |
| Project isolation | Choose global installation or keep skills inside a single project |
| Lifecycle management | Search, inspect status, check for updates, and update installed skills |
| Version-controlled source | Every skill is plain text, reviewable, portable, and easy to customize |
The default target is the global Claude Code skills directory.
curl -fsSL https://raw.githubusercontent.com/LeonChaoX/qinyan-academic-skills/main/install.sh | bashNote
The installer requires Bash, Git, and curl. On Windows, run it in WSL or Git Bash. Review install.sh before piping it to Bash if your environment requires source inspection.
INSTALLER="https://raw.githubusercontent.com/LeonChaoX/qinyan-academic-skills/main/install.sh"
# Install one category
curl -fsSL "$INSTALLER" | bash -s -- --category 01
# Install one skill
curl -fsSL "$INSTALLER" | bash -s -- --skill scanpy
# Install into the current project instead of globally
curl -fsSL "$INSTALLER" | bash -s -- --project --skill scientific-writing
# Target another supported agent
curl -fsSL "$INSTALLER" | bash -s -- --tool codexShort forms such as -c, -s, and -t are also supported. Run bash install.sh --help for the complete command reference.
# Browse and search
bash install.sh --list
bash install.sh --list-skills
bash install.sh --search "protein"
# Inspect and update
bash install.sh --status
bash install.sh --check-update
bash install.sh --update
bash install.sh --update --skill scanpygit clone https://github.com/LeonChaoX/qinyan-academic-skills.git
cd qinyan-academic-skills
bash install.sh --help| Agent | --tool value |
Global directory | Project directory |
|---|---|---|---|
| Claude Code | claude (default) |
~/.claude/skills/ |
.claude/skills/ |
| Cursor | cursor |
~/.cursor/skills/ |
.cursor/skills/ |
| Codex | codex |
~/.codex/skills/ |
.codex/skills/ |
| Gemini CLI | gemini |
~/.gemini/skills/ |
.gemini/skills/ |
| OpenClaw | openclaw |
~/.openclaw/skills/ |
.openclaw/skills/ |
| OpenCode | opencode |
~/.config/opencode/skills/ |
.opencode/skills/ |
Each category links directly to its source directory. Use --list-skills for the complete list or --search <term> to discover skills by name and description.
| ID | Category | Skills | Focus |
|---|---|---|---|
| — | Qinyan Academic skills | 10 | Paper search, analysis, citations, and Nature-focused writing, review, figures, and statistics |
| 01 | Literature discovery and reference management | 10 | Scholarly search, reviews, citation workflows, and literature databases |
| 02 | Scientific writing and communication | 6 | Manuscripts, peer review, research proposals, and venue templates |
| 03 | Academic presentation and visualization | 9 | Slides, posters, schematics, infographics, and scientific figures |
| 04 | Research methods and scientific reasoning | 10 | Hypotheses, critical thinking, brainstorming, evaluation, and grants |
| 05 | Bioinformatics and genomics | 21 | Sequences, single-cell analysis, regulatory networks, and variants |
| 06 | Cheminformatics and drug discovery | 12 | Molecular processing, docking, medicinal chemistry, and simulation |
| 07 | Clinical medicine and precision health | 18 | Trials, clinical reports, imaging, variants, and decision support |
| 08 | Protein engineering and structural biology | 7 | Protein language models, structures, domains, and sequence resources |
| 09 | Machine learning and artificial intelligence | 14 | Deep learning, classical ML, forecasting, optimization, and explainability |
| 10 | Materials science and computational physics | 10 | Materials, quantum computing, astronomy, fluids, and simulation |
| 11 | Data analysis and statistical modeling | 11 | Dataframes, statistics, visualization, graphs, and symbolic mathematics |
| 12 | Scientific databases | 22 | Chemistry, genes, proteins, pathways, populations, and patents |
| 13 | Laboratory automation and integrations | 9 | ELNs, cloud labs, robotics, imaging, and workflow platforms |
| 14 | Documents and data tools | 7 | PDF, Word, PowerPoint, spreadsheets, notebooks, and conversion |
| 15 | Finance and economic data | 6 | Markets, filings, fiscal data, economic series, and research reports |
| 16 | Geospatial analysis and remote sensing | 2 | GIS, satellite imagery, spatial analysis, and remote sensing |
| 17 | Platforms and infrastructure | 3 | Resource discovery, cloud compute, and ISO 13485 readiness |
| Skill | Purpose |
|---|---|
qinyan-paper-search |
Search Google Scholar, PubMed, arXiv, and Wanfang through the Qinyan Academic OpenAPI |
qinyan-paper-analysis |
Produce structured, in-depth analysis of individual academic papers |
qinyan-paper-polish |
Improve Chinese and English academic prose while preserving meaning |
qinyan-citation |
Discover sources and format citations in common academic styles |
qinyan-topic-analysis |
Evaluate research topics, evidence landscapes, and promising directions |
qinyan-nature-writing |
Build evidence-led manuscript arguments, sections, and initial-submission materials |
qinyan-nature-polishing |
Restructure, translate, and polish academic prose without changing scientific meaning |
qinyan-nature-review |
Run traceable pre-submission review through conceptual, technical, and evidence lenses |
qinyan-nature-figures |
Design, export, and preflight reproducible publication figures |
qinyan-nature-statistics |
Plan, analyse, audit, and report design-aware manuscript statistics |
| Goal | Recommended starting points |
|---|---|
| Literature search and evidence mapping | qinyan-paper-search, openalex-database, pubmed-database, literature-review |
| Manuscript drafting and revision | qinyan-nature-writing, qinyan-nature-polishing, qinyan-nature-review, venue-templates |
| Grant and proposal development | nsfc-proposal, nssfc-proposal, research-grants, research-proposal |
| Academic slides and posters | paper-slide-deck, scientific-slides, latex-posters, pptx-posters |
| Bioinformatics and omics | biopython, scanpy, pydeseq2, scvi-tools, gget |
| Drug discovery and molecular modeling | rdkit, deepchem, diffdock, molecular-dynamics, rowan |
| Statistical analysis and visualization | statistical-analysis, statsmodels, polars, plotly, seaborn |
| Clinical and precision medicine | clinical-decision-support, clinicaltrials-database, clinvar-database, pydicom |
Tip
Start with one focused skill, inspect its SKILL.md, and combine skills only when the research workflow genuinely needs multiple capabilities.
- Technical guide — installer architecture, search, update detection, and version management
install.sh --help— complete CLI options and examplesskills/— canonical source for every skill
This collection combines first-party Qinyan Academic skills with carefully organized skills derived from the following open-source projects:
- K-Dense-AI/claude-scientific-skills — broad scientific research skill collection
- luwill/research-skills — academic slides, literature reviews, and research planning workflows
- Yuan1z0825/nature-skills — Apache-2.0 architectural inspiration for Nature-focused workflows; the Qinyan suite is independently rewritten
- Qinyan Academic — paper search, paper analysis, academic polishing, citation discovery, and topic analysis
Attribution and license metadata for an individual skill remain in that skill's SKILL.md.
Contributions that improve skills, translations, documentation, and installer compatibility are welcome.
- Fork the repository and create a focused branch from
main - Add or update a skill in the appropriate category, keeping
SKILL.mdas its entry point - Update affected documentation and translations
- Validate links, commands, and formatting
- Open a pull request with a clear description of the change and its impact
When changing shared facts such as counts, supported tools, or CLI flags, update all five README files in the same pull request.
The repository is distributed under the MIT License. Individual skills may declare additional or different terms in their SKILL.md; review those terms before redistribution or commercial use.