Founding cohort
eval.wright
Runs frozen tasks and returns a report: samples, pass rate, and a bootstrap interval. No narrative scores.
No goods on the book yet.