Medsci Skills

Parse peer reviewer comments and generate a structured Response to Reviewers document with tracked manuscript changes. Classifies comments as MAJOR/MINOR/REBUTTAL, coordinates new analyses with /analyze-stats and /make-figures, and produces cover letter for editor.

Detect and remove AI writing patterns from academic manuscripts and response-to-reviewers letters. Scans for 27 common AI-generated text patterns and rewrites flagged passages to sound naturally human-written while preserving technical accuracy, bounding how much of the text a rewrite is allowed to touch.

Sync research references from .bib files to Zotero library + Obsidian literature notes. Extract cross-cutting concept notes when enough literature accumulates. Works after /search-lit or standalone.

Meta-analysis topic discovery and feasibility assessment. Professor-first (profile → gap) or Topic-first (question → gap → co-author). Pre-protocol phase from idea to ranked topic list.

Design or audit a model-agnostic evaluation harness for an LLM or multimodal LLM on a clinical task (radiology report generation, visual question answering, clinical text extraction/classification) — the adjudicated reference standard, clinical-efficacy metrics (RadGraph-F1 / CheXbert-F1 beyond BLEU/ROUGE), faithfulness and hallucination, pretraining-contamination of public benchmarks, prompt-sensitivity and determinism, answer-matching, and a reader study — and gate the plan for those axes. Works on a closed API or open weights. Never fabricates outputs or scores, and never reports n-gram overlap as clinical correctness.

Interactive data profiling and cleaning assistant for medical research. Three-stage workflow (profile, flag, code-generate) with user approval gates at each step. Handles missing values, outliers, duplicates, and type mismatches in CSV/Excel clinical data. Does NOT auto-clean — all decisions require researcher confirmation.

Offer your local changes back to the project — a journal profile you added, a checklist item you fixed, a skill you adapted to your department — as a pull request or an issue, without ever typing a git command. Detects what you changed against the installed version, scans it for patient data and identifiers, shows you every line, and sends nothing until you confirm. Also files feedback: a detector that fired wrongly, a step that failed on your file.

De-identify clinical research data before LLM-assisted analysis. Standalone Python CLI detects PHI via regex + heuristics with 10 country locale packs (kr, us, jp, cn, de, uk, fr, ca, au, in). Interactive terminal review. No LLM touches raw data — the script runs locally without any network or AI calls.

Generate the documentation an engineer-built medical-imaging model must carry — a Model Card (Mitchell et al. 2019), a Datasheet for its dataset (Gebru et al. 2021), and a METRIC-informed data-quality pass — filled from user-supplied facts, then verify every required section is present and non-empty before the card ships to a repo, Hugging Face card, or manuscript supplement. Never fabricates numbers, provenance, consent, or licence; unfilled fields stay flagged. Ships a deterministic completeness gate. Model Card and Datasheet are documentation standards vendored here as templates, not counted reporting checklists.

Literature search and citation management for medical research. Searches PubMed, Semantic Scholar, and bioRxiv/medRxiv with verified citations. Anti-hallucination — every reference verified via API before inclusion. Generates BibTeX entries.

Add a new journal to the MedSci Skills profile database. Extracts metadata from author guidelines, generates write-paper (detailed) and find-journal (compact) profiles in canonical format with quality gates.

Cross-cutting reference manager for medical manuscripts. Single entry point for citation-key validation, journal-CSL pandoc rendering, manuscript ↔ DOCX cross-reference QC, marker conversion (``[N]`` ↔ ``[@key]``), and native Zotero CWYW field-code injection. Replaces the inline reference-handling that previously lived in ``/write-paper`` Phase 7.6 and is reused by ``/revise``, ``/peer-review``, ``/sync-submission``, and any skill that produces a journal submission. Audit-only verification stays in ``/verify-refs`` — this skill writes (renders, injects, converts); that skill only reads.

General-purpose research orchestrator. Routes ambiguous or multi-step requests to the right skill(s) from the medsci-skills bundle. Use when the user describes a research goal without naming a specific skill, or when a task spans multiple skills.

Peer review assistant for medical journals. Generates structured review drafts with journal-specific formatting. Constructive developmental tone with systematic manuscript analysis.

Pre-submission self-review for the user's own manuscripts, applying a reviewer perspective. Systematic check across 10 categories with research-type branching. Outputs Anticipated Major/Minor Comments with severity framing and optional R0 numbering for /revise pipeline integration.

Audit-only verification of manuscript references against PubMed and CrossRef. Detects fabricated or mismatched citations and writes qc/reference_audit.json. Does not modify references/ or refs.bib.

Full-pipeline medical/scientific paper writing. 8-phase IMRAD workflow from outline to submission-ready manuscript. Supports original articles, case reports, case series, meta-analyses, AI validation studies, animal studies, and technical notes. Do NOT trigger for self-checking (use self-review instead).

Medical AI paper optimization for AI search engines (Perplexity, ChatGPT web, Elicit, Consensus, SciSpace) and RAG-based literature tools. Applies when drafting or reviewing titles, abstracts, structured summary boxes (Key Points / Research in Context / Plain-Language Summary), manuscripts for high-impact medical AI journals (Lancet Digital Health, Radiology, Radiology-AI, npj Digital Medicine, Nature Medicine), preprints (medRxiv/arXiv), GitHub README + CITATION.cff + Zenodo archives, and Hugging Face model/dataset cards. Integrates TRIPOD+AI, CLAIM 2024, STARD-AI, TRIPOD-LLM, DECIDE-AI reporting requirements with generative engine optimization (GEO) principles. Produces a visible pass/fail checklist.

Generate N analysis scripts from a single methodology template × multiple exposure/outcome combinations. The "80-person team" pattern — same validated method, swap variables only. Produces batch R/Python code + summary matrix.

Study design and validity review for radiology and medical AI research. Identifies analysis unit, cohort logic, leakage risks, comparator design, validation strategy, and reporting guideline fit before drafting or submission.

Journal recommendation engine for medical manuscripts. 2-pass matching against a curated public profile library plus any user-local private profiles, enriched with detailed write-paper profiles for top-5 output. Returns ranked recommendations with scope fit rationale, AI disclosure policy, and homepage links. Impact-factor and APC figures in a profile are point-in-time and may be stale — verify current metrics at the journal site. A pre-ranking acceptance-readiness pre-flight scans the manuscript for design-ceiling, unfixable-defect, and importance-risk signals to add an acceptance-feasibility axis alongside scope fit, and the output includes a reject-fallback cascade plan.

Generate publication-ready figures and visual abstracts for medical research papers. Supports ROC curves, forest plots, CONSORT/STARD/PRISMA flow diagrams, calibration plots, Kaplan-Meier curves, Bland-Altman plots, confusion matrices, pipeline diagrams, and journal-specific visual/graphical abstracts (python-pptx template-based).

Produce or audit a radiomics / tabular clinical-ML study — imaging or clinical features → any classical learner (penalised logistic [LASSO / ridge / elastic-net], SVM, k-NN, naive Bayes, LDA/QDA, decision tree, random forest, gradient boosting [XGBoost / LightGBM / CatBoost], shallow MLP, stacked ensembles) → a clinical outcome — so it clears the rigor bar reviewers expect: nested cross-validation (tuning never on the reported folds), dimensionality control for the features-far-exceed-events regime, feature selection inside the fold, feature-stability (ICC / test-retest) filtering, calibration, and external/temporal validation. The deterministic gate is learner-agnostic (it audits the pipeline, not the algorithm). Emits a pipeline manifest and the gate. The most common solo-doable clinical-ML workflow — no GPU, no engineer. Integrates scikit-learn / xgboost / lightgbm / catboost / pyradiomics; it does not reimplement them.

Scaffold and draft medical/AI literature reviews (narrative, scoping PRISMA-ScR, or systematic). Asks for the spine axis, builds a 7-part skeleton with a required Intro scope/non-overlap block, a summary-table stub, an evaluation-metrics critique subsection, and reporting-guideline wiring. Reuses the self-review RV1-RV9 narrative-review probes for QC. Does not invent citations.

Diagnostic checklist for the MedSci Skills runtime. Verifies Python, R, Node, Claude Code, Git, Zotero, and configured MCP servers, and prints a pass/fail table with links to the right setup doc for any missing component. Read-only — does not install anything.

Statistical analysis for medical research papers. Generates reproducible Python/R code with publication-ready tables and figures. Supports diagnostic accuracy, inter-rater agreement, meta-analysis, survival analysis, survey data, group comparisons, regression, propensity score, and repeated measures.

Fill institutional Word form templates (.doc/.docx) for IRB protocols, ethics applications, grant proposals, and other structured research documents while preserving the original styles, table layouts, fonts, and page geometry. Pairs with write-protocol — write-protocol drafts the scientific content, fill-protocol renders it into the institutional template. Korean-aware (CJK eastAsia font enforcement, table cantSplit) but works for any language template.

Grant and challenge proposal support for radiology and medical AI projects. Structures significance, innovation, approach, milestones, and consortium roles while keeping claims evidence-based and executable.

Systematic review and meta-analysis pipeline for medical research. Covers protocol registration (PROSPERO), search strategy, screening, data extraction, risk of bias assessment (QUADAS-2/ROBINS-I), statistical synthesis (bivariate/HSROC for DTA, random-effects for intervention), and PRISMA-compliant reporting. Supports both DTA and intervention meta-analyses.

Academic presentation preparation — paper-driven (journal club, grand rounds, seminar) and lecture/teaching decks (course material, workshop slides, conference talks). Analyzes source material, finds supporting references, drafts audience-adapted speaker scripts, generates or augments PPTX with speaker notes, and prepares Q&A.

Convert a personal agent skill into a distributable, open-source-ready skill. Runs PII audit, generalization, license compatibility check, cross-platform adapter review, and packaging workflow.

End-to-end cross-national comparison study using KNHANES + NHANES + CHNS (or other parallel surveys). Variable harmonization, parallel weighted analysis, and comparison tables. Supports 2-country (KR+US) and 3-country (KR+US+CN) designs.

Produce or audit the interpretability/explainability analysis of a medical-imaging model — Grad-CAM / Grad-CAM++ / attention-rollout / saliency / integrated-gradients — so it clears the rigor bar a reviewer expects: mandatory Adebayo sanity checks (model- and data-randomisation), a quantitative localisation metric against ground truth (IoU / pointing game / Dice) instead of eyeballed examples, a cohort-level result rather than cherry-picked cases, and attribution framing rather than "proof the model is correct". Emits an explainability-report manifest and a deterministic rigor gate. Integrates captum / pytorch-grad-cam; it does not reimplement them, and never runs a model on real patient data.

Batch-generate per-author ICMJE Conflict of Interest Disclosure Forms (`coi_disclosure.docx`) for manuscript submission. Pre-fills all 13 disclosure items as "☒ None" + final certification ☒ using a synthetic seed template shipped with the skill, then clones the seed per author with Date, Name, and Manuscript Title replaced. Designed for the common case of hospital-based observational research where no author has real financial conflicts; the circulated forms become "reply 'no changes' + sign" for most authors and only flag those who need to amend.

Intake and normalize a new radiology research project. Classifies project type, summarizes current state, identifies missing inputs, recommends next steps, and scaffolds lightweight project memory files.

Research project management for medical manuscripts. Scaffold project structure, track writing progress across phases, maintain project memory files, generate submission checklists and backwards timelines. Commands: init, status, sync-memory, checklist, timeline.

Generate a reproducible, runnable PyTorch training repo for a medical-imaging task — segmentation, classification, detection, image-to-image synthesis, self-supervised pretraining, or fine-tuning a pretrained backbone (transfer learning) — the missing middle link between choosing an architecture and validating a trained model. Emits a patient-level seed-locked split as an auditable artifact, a task-appropriate model, train and evaluate scripts that seed every RNG and infer under eval mode, a config, requirements, a reproducibility record, and a Methods stub with VERIFY placeholders (no fabricated numbers). Fine-tuning mode adds a frozen-then-unfrozen schedule, discriminative learning rates, and a pretrained-weight provenance record. The reproducibility guarantees hold by construction, so the build is leakage-safe before any training runs. Integrates with MONAI, nnU-Net, TorchIO, timm, and torchvision — it does not reimplement them.

Vet the concrete third-party model a study will be built on — this repository, this revision, this checkpoint — not the architecture family. Records a model dossier (source and version pin, licence and the file it was read from, intended use, pretrained-weight provenance, model task vs study task, reported validation, what the model was developed on, your evaluation arms) and gates it deterministically. Catches what a licence check and a citation count cannot: an evaluation arm sitting on the benchmark the model was developed or tuned on, so the arm reads like validation while being closer to a training-set score. Also an evaluation set inside a pretraining corpus, an unstated or use-incompatible licence, an unpinned revision, and a hardware claim never executed. It vets an artifact; it never downloads or runs one.

Render academic Markdown documents (English or Korean) to publication-quality PDF via pandoc + xelatex. Targets non-bibliography artifacts: research proposals, IRB cover letters, briefing handouts, anchor docs (Q&A grids), and reference tables. Auto-infers pipe-table column widths from content (label column shrinks to fit, data columns share remaining width). CJK-aware font fallback for Korean text (Apple SD Gothic Neo on macOS, Noto Sans CJK KR on Linux). NOT for: manuscripts with bibliography (use /manage-refs render_pandoc.sh), Word form filling (/fill-protocol), figures (/make-figures).

IRB/ethics committee research protocol generator. Produces 4 core sections (Background, Study Design, Sample Size, Statistical Plan) with full prose, plus 6 skeleton sections with TODO markers for institution-specific content. Integrates outputs from design-study, calc-sample-size, and search-lit.

PubMed author profile analysis. Author name → PubMed fetch → study-type classification → visualization → strategy report → optional trajectory-archetype classification.

Check manuscript compliance with medical research reporting guidelines. Supports 49 guidelines including STROBE, STROBE-MR, RECORD, REMARK (prognostic tumor-marker studies), TARGET (target trial emulation), GATHER (burden-of-disease / health-estimate modeling), CONSORT, CONSORT-AI, STARD, STARD-AI, TRIPOD, TRIPOD+AI, TRIPOD-LLM, PGS-RS, ARRIVE, PRISMA, PRISMA 2020 for Abstracts, PRISMA-DTA, PRISMA-P, PRISMA-ScR (scoping reviews), CARE, SPIRIT, SPIRIT-AI, CLAIM, DECIDE-AI, MI-CLEAR-LLM, SQUIRE 2.0, CLEAR, MOOSE, GRRAS, SWiM, AMSTAR 2, CHEERS 2022, CROSS (survey studies), SRQR and COREQ (qualitative research), and risk of bias tools (QUADAS-3, QUADAS-2, QUADAS-C, RoB 2, ROBINS-I, ROBINS-E, ROBIS, ROB-ME, PROBAST, PROBAST+AI, NOS, COSMIN, RoB NMA). Generates item-by-item assessment with PRESENT/MISSING/PARTIAL status.

Research gap finder for longitudinal cohort databases. Profiles cohort strengths, matches PI expertise, scans literature saturation, and outputs ranked topic proposals with gap evidence. Works with any cohort: NHIS, UK Biobank, institutional EMR, health checkup registries, or disease-specific registries.

Academic English consistency linting and non-native (ESL) language polish for medical manuscripts. Deterministically flags abbreviation define-once violations, US/UK spelling drift, hyphen-vs-en-dash numeric ranges, P/p case, hyphenation variants, small-number style, and value/unit spacing, then guides a style-only clarity pass that never alters numbers, citations, or scientific meaning. Distinct from humanize (AI-tell removal) and check-reporting (guideline items).

Profile a medical-imaging dataset before any modelling decision is made — the acquisition grid, voxel spacing and orientation spread, the intensity domain, which label values are actually present, how much of the volume the target occupies, and how large the target is in millilitres — then gate that profile against the researcher's declared plan. Catches, at the point where it is still cheap, the dataset facts that otherwise surface after a training run: a "test set" that carries no ground truth, labels whose grid does not match their image, a stray label index, a target occupying a fraction of a percent while accuracy is planned as a metric, and acquisition heterogeneity nobody declared a resampling decision for. Emits a dataset-profile JSON and a deterministic gate that reads it (stdlib-only, so an audit travels with the JSON). It describes the data and audits the plan against it; it does not preprocess, split, or train.

Replicate an existing cohort study's methodology on a different database. Extracts study design from a source paper, maps variables to the target DB via harmonization table, generates analysis code, and produces a replication difference report.

Audit SSOT-to-submission drift and create journal submission manifests from canonical manuscript artifacts.

Dataset version control for research reproducibility. Builds a deterministic content-hash manifest of a dataset (file SHA-256 + tabular schema + per-column value hashes), verifies a later copy against it to detect drift (schema change, row-count change, value changes), and diffs two manifests. Use to prove an analysis ran on the intended data, lock a dataset version, or reproducibility-lock bundled demos.

Choose a model architecture for a medical-imaging research question before scaffolding. Maps the task (classification, segmentation, detection, transfer), modality and dimensionality, labelled-data scale, and class imbalance to a shortlist of architectures, each grounded in its source paper with a when-to-use, a medical-imaging use, a reference implementation, the typical validation setup, and the matching model-scaffold template. Covers the foundational curriculum (ResNet, DenseNet, EfficientNet, ViT, Swin; U-Net, 3-D U-Net, Attention/Residual U-Net, nnU-Net (+ ResEnc/MedNeXt/STU-Net), Mask R-CNN; SAM/MedSAM, nnInteractive/VISTA3D interactive-3D, TotalSegmentator, BiomedCLIP, DINO/MAE/SimCLR; and graph neural nets — GCN/GraphSAGE/GAT/GIN/BrainGNN — for brain connectomes). It teaches archetypes and the task-to-architecture logic (including the "scale the CNN, new≠better" rigour caveat), not a live SOTA leaderboard.

Interactive sample size calculator for medical research. Decision-tree guided test selection, reproducible R/Python code, effect size interpretation, and IRB-ready justification text. Supports diagnostic accuracy, agreement, proportions, continuous outcomes, survival, ANOVA, logistic regression, and non-inferiority/equivalence designs.

Literature-grounded variable operationalization for observational research. Turns a data dictionary + research question into a citation-backed table of exposure/outcome/covariate definitions, cutoffs, and DB variable mappings. Prevents ad-hoc phenotype definitions that invite reviewer rejection. Bridges /search-lit output into /write-protocol Methods.

Compute and report task-correct held-out metrics for a trained medical-imaging model — segmentation (Dice plus a boundary metric such as HD95 or NSD, per structure), classification (AUROC plus AUPRC and sensitivity/specificity with bootstrap CIs at the deployment prevalence), detection (FROC or mAP with a stated IoU criterion), interactive/promptable segmentation (the interaction-count, convergence, and per-case-time axes a static Dice omits), or generative/synthesis image evaluation (similarity plus the downstream-task efficacy similarity alone cannot establish) — plus calibration and subgroup slices. Emits a per-case results table that analyze-stats turns into publication tables, and gates the metric choice against Metrics Reloaded, CLAIM 2024, and Park et al. 2024 (no pixel accuracy for segmentation, no bare accuracy under imbalance, no static Dice for an interactive method, no similarity-only claim for a generative model). Numbers come only from executed code, never hand-typed.

Design or audit the clinical-validation study for an engineer-built medical-imaging model (segmentation, classification, or detection) before the validation report or manuscript is written. Covers patient-level split disjointness and the data-leakage taxonomy, tuning-on-test, internal versus genuine external validation, comparator design, single-run versus multi-seed variance, task-correct metric selection, test-set sizing, and CLAIM 2024 / TRIPOD+AI / STARD-AI reporting fit. Ships a deterministic split-leakage gate that proves patient disjointness by set arithmetic on the emitted split-assignment table. Does not build or train models — it integrates with MONAI / nnU-Net, it does not replace them.

Generate a citable data dictionary / codebook from a tabular dataset (CSV/TSV/Excel/Parquet/Stata/SAS). Profiles every variable — role, type, units placeholder, level frequencies, range/quantiles, missingness — and emits codebook.md + codebook.json. Flags coded variables whose level meanings are unknown as [NEEDS DICTIONARY] rather than guessing them, feeding /define-variables and the dictionary-first workflow.

Batch download open-access PDFs by DOI using legitimate OA APIs (Unpaywall, PMC, OpenAlex, Crossref). Optional PDF→Markdown conversion for token-efficient LLM analysis.

Design or audit the data-preparation stage of a medical-imaging model — DICOM/NIfTI intake, resampling and intensity normalisation, and the augmentation plan — so the pipeline is leakage-safe before model-scaffold builds the training repo. Emits a declarative preprocessing manifest and a deterministic data-stage leakage gate that catches the leaks a split table cannot see: a dataset-level normaliser fit on non-train data, any data-fitted transform run before the split, and the same patient's slices crossing splits. Integrates MONAI / TorchIO transforms; it does not reimplement them, and it never runs preprocessing on real patient data.

Design or audit the uncertainty-quantification, out-of-distribution (OOD) detection, and selective-prediction layer of a medical-imaging model framed for deployment — so a clinical-use claim carries calibrated per-case uncertainty (MC-dropout / deep ensemble / conformal / Bayesian), an OOD guard validated on a held-out OOD set, an abstention rule at a pre-specified operating point, and uncertainty checked under distribution shift. Emits an uncertainty manifest and a deterministic gate that flags a deployment claim built on point predictions, conformal intervals with unmeasured coverage, and an OOD claim with no held-out OOD data. Integrates MAPIE / captum / pretrained OOD scorers; it does not reimplement them and never runs a model on real patient data.

Turn a folder of research PDFs into an Obsidian knowledge vault — consistently formatted literature notes with frontmatter, PDF embed links, and cross-referenced atomic concept notes. Use whenever the user wants PDFs converted to Obsidian notes, a batch of papers summarized into a common template, a research "second brain" built or extended, or concepts extracted across accumulated notes — even if they never say "Obsidian". Pairs with /lit-sync, which owns the same vault folders from the Zotero/BibTeX side.

Design and validity review for studies that benchmark one or more AI systems against a human-expert panel as the reference. Covers the evaluation question and arm definition, decoupled multi-dimensional rubrics with anchors, planted calibration probes, reviewer-panel construction, inter-rater reliability targets, LLM-as-judge versus human-as-judge adjudication, construct-independence guards, and a structured rating-export schema. Use before data collection on an AI-vs-expert evaluation.