This is a complete pipeline for analyzing Neuropixels neural recordings, from loading raw SpikeGLX or Open Ephys data through spike sorting with Kilosort4 and unit curation using Allen Institute or IBL criteria. It wraps SpikeInterface with sensible defaults for preprocessing (highpass filtering, bad channel detection, common average referencing), motion correction, and quality metrics computation. The AI-assisted curation feature is genuinely useful for borderline units where SNR sits in that awkward 3-8 range. Built for both Neuropixels 1.0 and 2.0 probes. If you're doing extracellular electrophysiology and tired of stitching together preprocessing scripts, this gives you a tested workflow that matches what major neuroscience labs actually use. The documentation is thorough about pitfalls like remembering phase shift correction for 1.0 probes.
npx skills add https://github.com/k-dense-ai/scientific-agent-skills --skill neuropixels-analysis