endurance-coach-validation-and-qalisted

Load this when the question is "is this good enough to ship?" for the Endurance Coach repo — what counts as evidence, how to test a change, whether the suite is green, and where coverage is thin. Triggers: "how do I test this", "add a test", "write a unit test", "is the suite green", "npm test failing", "what counts as done / evidence", "acceptance threshold", "is this finding confirmed or exploratory", "should this insight ship", "coverage is thin here", "what's the CI contract", "does this need a test before I commit", "golden / invariant / certified set", "node:test", "tsx --test", "fixture", "faked client", "test inversion", "profileQuestions test", "dashboard script-parse test", "writegate propose/confirm test", "statvalidity", "monitoring validated vs exploratory". Also load before merging any behavioural change to decide if it carries the test its Definition-of-Done requires. DON'T load this for: the derivations of the statistics themselves (use endurance-coach-proof-and-analysis-toolkit), the go/no-go
maxeskell/personal-training-app · ★ 0 · Testing & QA · score 70

Install: claude install-skill maxeskell/personal-training-app

# Endurance Coach — validation & QA **Use this when** you need to know what counts as *evidence* in this repo, how to write or run a test, whether the suite is green, or whether a statistical finding is allowed to ship as "confirmed". **Don't use this when** you need the *math* behind a statistic (→ `endurance-coach-proof-and-analysis-toolkit`), the *step-by-step campaign* to validate one detector on this athlete's data (→ `endurance-coach-n1-validation-campaign`), the *epistemics* of accepting a hunch (→ `endurance-coach-research-methodology`), the *change gate / ship flow* (→ `endurance-coach-change-control` and `endurance-coach-run-and-operate`), or the *diagnostic tools* like `doctor`/`cost`/`verify:reads` (→ `endurance-coach-diagnostics-and-tooling`). Jargon, defined once: **hermetic** = a test that touches no network, clock-dependent state, or real credentials — it runs identically anywhere. **Fixture** = hard-coded sample input built inside the test (or a small committed file), used instead of a live fetch. **Faked client** = a stand-in for an external client (AI Endurance, Garmin) that records what it was asked to do in a `calls` array so the test can assert on it. **Provenanced field** = every `AthleteState` value is `{ value, source, note? }` so a missing datum degrades to `null`, not a crash. **FDR** = false-discovery-rate control (Benjamini–Hochberg). --- ## 1. The definition of evidence (what "done" means here) Two gates, both mandatory, both non-negotiable