Tooluniverse Statistical Modeling

This is a stats workhorse built for biomedical and clinical trial data, wrapping R and Python libraries into scripts that know the domain-specific gotchas. It handles natural spline regression (via R's `ns()`, not patsy approximations), ordinal logistic models for AE severity with proper SDTM encoding handling, per-gene ANOVA workflows that avoid naive pooling, and power analyses. The scripts enforce workspace isolation so your input data stays clean, and the rule-zero philosophy is smart: check for pre-computed results before re-running anything. If you're doing clinical trial stats or gene expression modeling and tired of debugging why your spline fit doesn't match the paper's R output, this saves the round trips.

skills.sh installs: 407
GitHub stars: 1.8K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add mims-harvard/tooluniverse --skill tooluniverse-statistical-modeling --agent claude-code

Installs into .claude/skills of the current project.

Files

Statistical Modeling for Biomedical Data Analysis

RULE ZERO — Check for pre-computed results FIRST

Before following any instruction below, scan the data folder for:

*_executed.ipynb → read with tu run read_executed_notebook '{"data_folder":"<path>","search":"<keyword>"}' and cite its cell outputs as the authoritative answer
Pre-computed result files (CSV/TSV with names like *results*, *deseq*, *enrich*, *stats*, *_simplified.csv) → read directly and report the requested value
Canonical analysis scripts (analysis.R, run_*.py, find_*.R, *.Rmd) → execute as-is and read the output

Only follow this skill's re-analysis recipe below if none of the above exist. Re-running from raw data produces different numbers than the published answer and is much slower (often 5-10× turn count).

PRIMARY SCRIPTS — use these FIRST

These scripts encode the question-specific gotchas in scripts/ and emit labelled, parseable output. Prefer them over ad-hoc statsmodels / scipy code.

SKILL.md on GitHub