Two-stage ensemble for planning meaningful tests. Load when writing tests for new features or reviewing test quality. Counters the tendency to write tests that exercise the stdlib instead of your code. Has full (9-persona) and lightweight (5-persona) modes.
