sciagent-skill-creator
Scaffold a new SciAgent-Skills entry. Picks pipeline/toolkit/database/guide template, creates skills/{category}/{name}/SKILL.md with valid frontmatter, appends…
Scientific manuscript writing: IMRAD, citation styles (APA/AMA/Vancouver/IEEE), figures/tables, reporting guidelines (CONSORT/STROBE/PRISMA/ARRIVE), writing principles (clarity/conciseness/accuracy), venue-specific style. For LaTeX see companion assets.
$ npx -y skills add jaechang-hits/SciAgent-Skills --skill scientific-manuscript-writing --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/scientific-manuscript-writingContext preview
The summary Claude sees to decide when to auto-load this skill.
Scientific manuscript writing: IMRAD, citation styles (APA/AMA/Vancouver/IEEE), figures/tables, reporting guidelines (CONSORT/STROBE/PRISMA/ARRIVE), writing principles (clarity/conciseness/accuracy), venue-specific style. For LaTeX see companion assets.
name: "scientific-manuscript-writing" description: "Scientific manuscript writing: IMRAD, citation styles (APA/AMA/Vancouver/IEEE), figures/tables, reporting guidelines (CONSORT/STROBE/PRISMA/ARRIVE), writing principles (clarity/conciseness/accuracy), venue-specific style. For LaTeX see companion assets." license: "CC-BY-4.0"
Scientific manuscript writing is the discipline of communicating research findings with precision, clarity, and reproducibility. This knowhow covers the complete lifecycle of manuscript preparation: from planning and structuring a paper using IMRAD format, through applying correct citation styles and designing effective figures, to ensuring compliance with study-specific reporting guidelines. It applies across biomedical, social science, engineering, and computational fields.
This entry consolidates writing principles, manuscript structure, citation systems, figure/table design, and reporting standards into a unified reference for producing publication-ready scientific documents.
The three pillars of scientific writing govern all manuscript text:
Additional principles include **objectivity** (present results without bias, acknowledge conflicting evidence), **consistency** (same term for same concept, uniform notation), and **logical organization** (clear "red thread" connecting sections). See `references/writing_principles_style.md` for detailed guidance with examples.
IMRAD (Introduction, Methods, Results, And Discussion) is the standard organizational structure for original research articles, adopted across most scientific disciplines since the 1970s. It mirrors the scientific method:
| Section | Question Answered | Primary Tense | Typical Length | |---------|-------------------|---------------|----------------| | **Title** | What is this about? | N/A | 10-15 words | | **Abstract** | Complete summary | Mixed | 100-250 words | | **Introduction** | Why did you study this? | Present/past | 4-5 paragraphs | | **Methods** | How did you do it? | Past | 2-4 pages | | **Results** | What did you find? | Past | 2-4 pages | | **Discussion** | What does it mean? | Past/present | 3-5 pages | | **Conclusion** | Take-home message | Present | 1-2 paragraphs |
The Introduction follows a funnel structure: broad context, narrowing literature review, gap identification, and study objectives. Methods must provide sufficient detail for replication. Results present findings objectively without interpretation. Discussion interprets findings, compares with prior work, acknowledges limitations, and proposes future directions.
**Venue variations**: Nature/Science use modified IMRAD with methods in supplement. ML conferences (NeurIPS/ICML) use Introduction-Method-Experiments-Conclusion with numbered contributions and ablation studies. See `references/manuscript_structure.md` for section-by-section guidance.
Five major citation styles serve different disciplines:
| Style | Format | Primary Disciplines | |-------|--------|---------------------| | **AMA** | Superscript numbers | Medicine, health sciences | | **Vancouver** | Numbers in brackets [1] | Biomedical sciences | | **APA** | Author-date (Smith, 2023) | Psychology, social sciences | | **Chicago** | Notes-bibliography or author-date | Humanities, some sciences | | **IEEE** | Numbers in brackets [1] | Engineering, computer science |
**Best practices across all styles**: Cite primary sources when possible. Include recent literature (within 5-10 years for active fields, 2-3 years for ML). Balance citation distribution across Introduction and Discussion. Verify all citations against original sources. Keep self-citations below 20%. Always check the target journal's author guidelines for the required style.
See `references/citation_guide.md` for complete format specifications, journal-specific requirements, and DOI formatting rules.
Figures and tables are the backbone of a scientific paper; many readers examine them before reading the text.
**Decision rule**: Can the information be conveyed in 1-2 sentences? If yes, use text only. If no and precise values are needed, use a table. If no and patterns/trends matter most, use a figure.
**Core design principles**: 1. **Self-explanatory**: Each display item must stand alone with a complete caption defining all abbreviations, units, sample sizes, and statistical annotations 2. **No redundancy**: Text highlights key findings from displays; it does not repeat all data 3. **Consistency**: Uniform formatting, color schemes, and terminology across all display items 4. **Optimal quantity**: Approximately one display item per 1,000 words of manuscript 5. **Accessibility**: Color-blind safe palettes (blue-orange, viridis); designs that work in grayscale
**Error bar rule**: Always state which measure is shown (SD for data spread, SEM for measurement precision, 95% CI for significance assessment). 95% CI is generally preferred because non-overlapping CIs indicate significant differences.
See
Turn your AI coding agent into a life sciences expert — 199 bioinformatics skills for Claude Code covering RNA-seq, single-cell analysis, genomics, proteomics, drug discovery, and more. Boosted BixBench from 65% to 92%. Open source.
Scaffold a new SciAgent-Skills entry. Picks pipeline/toolkit/database/guide template, creates skills/{category}/{name}/SKILL.md with valid frontmatter, appends…
Bayesian modeling with PyMC 5: priors, likelihood, NUTS/ADVI sampling, diagnostics (R-hat, ESS), LOO/WAIC comparison, prediction. Hierarchical, logistic, GP…
Time-to-event modeling with scikit-survival: Cox PH (elastic net), Random Survival Forests, Boosting, SVMs for censored data. C-index, Brier, time-dependent…
Guided statistical analysis: test choice, assumption checks, effect sizes, power, APA reporting. Pick tests, verify assumptions, or format results for…
Python statistical modeling: regression (OLS, WLS, GLM), discrete (Logit, Poisson, NegBin), time series (ARIMA, SARIMAX, VAR), with rigorous inference,…
DL cell/nucleus segmentation for fluorescence and brightfield microscopy. Pre-trained models (cyto3, nuclei, tissuenet) and a generalist flow-based algorithm…