Tooluniverse Image Analysis

This handles the statistical backend for microscopy experiments when you've already got measurement tables from ImageJ or CellProfiler. It's built around a specific workflow: colony morphometry, fluorescence quantification, cell counts, then statistical comparisons using pandas, scipy, and statsmodels. The skill emphasizes checking for pre-computed results before re-running analyses, which is practical for reproducibility but tells you this was written for a lab environment where notebooks get re-executed. Includes Dunnett's test for multi-group comparisons, Cohen's d for effect sizes, and polynomial or spline regression for dose-response curves. One gotcha clearly learned from experience: it defaults "relative proportion" questions to percentages, not decimal ratios, because that's what biology assay ground truths expect.

skills.sh installs: 656
GitHub stars: 1.8K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add mims-harvard/tooluniverse --skill tooluniverse-image-analysis --agent claude-code

Installs into .claude/skills of the current project.

Files

Microscopy Image Analysis and Quantitative Imaging Data

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).