Histolab

If you're working with whole slide images in digital pathology, this handles the annoying groundwork: tissue detection, tile extraction, and H&E preprocessing without writing a bunch of boilerplate. It wraps OpenSlide with three solid extraction strategies (random sampling, grid coverage, or score-based selection using nuclei/cellularity metrics), which is exactly what you need for prepping training datasets. The API is clean and the tile preview functions are genuinely useful before you burn compute extracting from gigapixel images. Python 3.8-3.11 only, Linux or macOS, and you'll need OpenSlide system libraries installed first. It's positioned as the lightweight option compared to pathml, so if you just need tiles out of slides without complex spatial proteomics pipelines, this is the faster path.

Install to Claude Code

npx -y skills add k-dense-ai/scientific-agent-skills --skill histolab --agent claude-code