nx_andelinwest_crew.nx

nx_andelinwest_crew.nx -- the NISHI TEAM ORGANS execute the andelinwest build (operator: build the builder + engineer + all the team; climb from hardware up toward a WYSIWYG site builder, S-class at each rung, judged by benchmarking/research not self-assessment). Each organ acts in its RACI lane: CONDUCTOR -- dispatches the deterministic L8->L0 frontier batch (nx_layered_backlog) BUILDER -- builds each rung, wielding the web-builder capability (nx_web_builder) on the spec ENGINEER -- gates each rung (UX-compliance verdict); a rung is DONE only when the gate passes CRITIC -- red-teams the result so a pass is not a blind spot EXAMINER -- grades it, and -- crucially -- the REFEREE refuses a self-graded S-class claim: our UX checklist is a PASS/FAIL compliance gate, NOT a DISCRIMINATING benchmark (it passes our v1 AND an award-winning site equally = low TNR per the judging research), so it can prove COMPLIANT but NOT S-class-exceed. S-class needs a discriminating benchmark (Lighthouse perf / a11y / real conversion) vs award-winning legal sites. So the team BUILDS it and the objective judge BLOCKS the overclaim + names what S-class requires. license_tier: ORIGINAL Cites knowledge/research/2026-06-04-objective-testing-judging.md (+ux-andelinwest).