fix: dataset integrity - pruned 67 malformed items

fix: shuffle MCQ candidates, fix length bias, add difficulty tracking

fix: removed all duplicate candidates across the benchmark using rigorous multi-pass review

feat: Artisan Wave 8 completion. 3030 hand-crafted items without templates

feat: Wave 6 complete - 2,164 items, ALL 20 modules have 100+ items

feat: push to 1,439 items (Wave 5), add 5 new modules