The body is dense and method-specific (pinned 0.2.13 signatures, schemas) that Claude would not reliably know, with no padding of basic concepts; version/date detail is confined to a dedicated "Scope and evidence cutoff" section rather than scattered prose.

Provides copy-paste-ready executable guidance throughout—`uv pip install "neurokit2==0.2.13"`, full script invocations with flags, and complete Python snippets such as `signals, info = nk.ecg_process(ecg, sampling_rate=250)`.

The "Core workflow" sequences a 10-step preprocessing order with explicit validation checkpoints (inspect_signal.py before filtering, validate_multimodal.py before bio_process), satisfying the clear-sequence-with-validation anchor.

SKILL.md is a concise overview with a well-signaled one-level-deep reference table (all 12 listed files exist under references/) plus an organized scripts table; no nested references.

Lists multiple concrete action domains—"build or audit reproducible research workflows for physiological time-series preprocessing, event/interval analysis, multimodal alignment, variability, and complexity"—matching the multi-action anchor.

Explicitly answers both what ("build or audit reproducible research workflows...") and when ("Trigger when code imports neurokit2..."), plus a negative boundary, so it is not capped at 2.

"Trigger when code imports neurokit2" supplies the natural keyword a user would actually say, with good coverage of relevant trigger phrasing.

NeuroKit2 is a clearly bounded biosignal niche with distinct triggers and an explicit "not for diagnosis or device validation" exclusion, making conflict with other skills unlikely.