deepseek-science-writing
Autonomously draft evidence-led LaTeX manuscripts, retain editable vectors and bibliography, compile real TeX, and inspect publication artifacts in DeepSeek Science.
- License 许可
- MIT
- 最近更新
- 2026-10-22
- Stars
- ★ 1
- Forks
- 0
01 前言 Front Matter
- name
- deepseek-science-writing
- license
- MIT
- description
- Autonomously draft evidence-led LaTeX manuscripts, retain editable vectors and bibliography, compile real TeX, and inspect publication artifacts in DeepSeek Science.
02 资源文件 19
- 资产文件references/figure-example/distribution.csv124 B
- 参考文档references/figure-example/figure.json2.3 KB
- 资产文件references/figure-example/heatmap.csv73 B
- 参考文档references/figure-example/hero.json2.0 KB
- 资产文件references/figure-example/line.csv219 B
- 参考文档references/figure-example/README.md2.3 KB
- 资产文件references/figure-example/scatter.csv75 B
- 参考文档references/figures-equations.md5.5 KB
- 参考文档references/nature-writing-standard.md13.6 KB
- 参考文档references/publication.md7.7 KB
- 参考文档references/publication.schema.json1.1 KB
- 参考文档references/templates.md4.7 KB
- 参考文档references/upstream-workflow.md4.6 KB
- 脚本scripts/layout_fixture.tex33.6 KB
- 脚本scripts/plot_figure.py4.9 KB
- 脚本scripts/plot_layout.py5.5 KB
- 脚本scripts/publication.py41.4 KB
- 脚本scripts/scaffold_manuscript.py6.3 KB
- 脚本scripts/scientific_figure.py39.9 KB
03 SKILL.md 正文
Evidence-led LaTeX publication
Apply automatically when the task requires scientific writing or a paper; use the configured shell and file tools without asking the user to name skills. Continue within existing user scope. Deliver illustrated, professionally composed actual LaTeX: full-width title, two-column body, narrow edge floats and wide top floats unless a specific target journal requires another form. Retain verified bibliography, source figures/data and claim evidence. Formatting never supplies findings or establishes scientific validity.
- Inspect the actual question, source reading depth and results. Automatically load and apply the bundled ARIS/Nature/Scholar workflow through the catalog for plan, section-by-section writing, figures, citations and review, using the configured model. Use scripts/scaffold_manuscript.py for evidence quote/artifact validation; it is evidence staging, while the publication CLI owns final typesetting. Preserve real evidence fields in the manuscript claim register. Never invent findings or citations.
- Read template selection and publication CLI. Choose the target venue's requirements; absent a target, proceed with two-column
neutral. Runpython scripts/publication.py init manuscript --title "Actual study" --template neutral --language eninto a new directory. Usezhorbilingualfor Chinese andnature-inspiredor actualieeewhen appropriate.--demo/--layout-democreate invented build fixtures, never study evidence. - Draft
main.tex,references.bibandclaims.csvfrom real evidence. Read figure, equation and claim rules. Use aligned mathematics, units, booktabs and editable vector PDF/SVG where suitable. Narrow floats require[t]/[b]/[tb]; wide floats require[!t]. Inline[h]/[H]is rejected. Record exact independent-unit counts and error definitions. Verify source support and bibliography, then resolve unknown/ARIS markers or report an unfinished draft. - Run the copied project CLI:
python publication.py doctor .,check .,compile .,check . --require-compiled, thenrender .. Missing engines/packages/fonts/renderers areneeds-setup; retain source and diagnostics. Never substitute a fake PDF, silently switch engines or claim success from source presence. Compilation disables shell escape and automatic MiKTeX installation, runs bounded real TeX/BibTeX and retains exact commands/logs. The helper accepts trusted sources; use the configured OS sandbox for untrusted TeX. - Inspect every rendered page and fix clipping, overfull content, unreadable figures, missing glyphs and reference diagnostics; rebuild and render after source changes. Checkpoint actual source, bibliography, claim register, analysis/figure inputs, vectors, PDF, logs and render/compile reports through
science_projectusing its current revision. Advance only after a successful checkpoint and an honest evidence assessment. Report the build status, visual review and scientific limitations separately; setup failure is incomplete delivery, not successful publication.
04 同仓库更多 Skills
bio-free-energy-calculations
Performs alchemical free-energy calculations including relative binding free energy (RBFE / FEP+) and absolute binding free energy (ABFE) via OpenFE, FEP+, GROMACS, AMBER pmemd, and OpenMM with explicit lambda scheduling, soft-core potentials, MBAR/BAR analysis, cycle-closure validation, and protocol-appropriate enhanced sampling. Compares ML alternatives (Boltz-2 affinity, DeepDock). Use when ranking analogs by binding affinity beyond docking accuracy, performing prospective lead optimization, or validating SAR predictions.
bio-conformer-generation
Generates 3D conformer ensembles using RDKit ETKDGv3 with knowledge-enhanced distance geometry, MMFF94/UFF force-field optimization, CREST + GFN2-xTB semi-empirical refinement, and macrocycle-aware torsion preferences. Provides explicit decision rules for single vs ensemble conformer use, RMSD pruning, energy windows, conformer count, and force-field choice. Use when preparing 3D ligands for docking, generating descriptor input for 3D QSAR, or sampling macrocycle/peptide conformational ensembles.
bio-admet-prediction
Predicts ADMET properties using ADMETlab 3.0 (119 platform features, including 77 prediction models with modeled-endpoint uncertainty), ADMET-AI, DeepChem MolNet, and chemprop D-MPNN with explicit handling of OECD QSAR principles, applicability domain assessment, calibration, hERG/CYP/AMES endpoints, and PAINS / Lipinski / Ro5 / Veber / BBB druglikeness filters. Use when filtering compounds for drug-likeness, prioritizing leads by predicted safety, or building an in-house ADMET QSAR model.
writing-systems-papers
Paragraph-level structural blueprint for 10-12 page systems papers targeting OSDI, SOSP, ASPLOS, NSDI, and EuroSys. Provides page allocation, paragraph templates, and writing patterns. Use when user says "写系统论文", "systems paper structure", "OSDI paper", "SOSP paper", or wants fine-grained structural guidance for a systems conference submission.
wiki-enrich
Fill in the per-paper TODO sections of research-wiki/papers/<slug>.md pages that literature-ingest skills leave as bare scaffolds. Use when user says 'enrich wiki', 'fill paper TODOs', 'wiki body 補完', '把 paper 摘要寫進 wiki', 'research-wiki 自動填', or after a batch ingest that left papers/ as TODO scaffolds.
web-debug-search
Search GitHub, Stack Exchange, Chinese technical communities, official documentation, and general developer web sources for software errors, compatibility problems, API usage questions, and real-world workarounds. Use for debugging and discovery only; results are not paper-citation evidence.