This splits patients into treatment groups based on genomic, clinical, and pharmacogenomic data. It runs a nine-phase pipeline that routes through disease-specific logic (cancer gets TMB/MSI stratification, metabolic diseases get complication risk scoring, rare diseases get genotype-phenotype matching). It fetches from ClinVar, PharmGKB, cBioPortal, OpenTargets, and GWAS databases, then computes a 0-100 precision medicine risk score with evidence tiers for every finding. The workflow is compute-heavy: it expects you to run Python for actual statistical analysis rather than just describe what you'd do. Best for cases where you need a full stratification report with treatment recommendations, not just variant interpretation. The documentation is thorough with scoring matrices and worked examples across six disease types.
npx skills add https://github.com/mims-harvard/tooluniverse --skill tooluniverse-precision-medicine-stratification