Rook & InkArchitecture portal

Handoff keep / adapt / reject

Assessment of ChatGPT handoff recommendations against existing code and best practice. Full write-up: docs/handoff-assessment.md.

RecommendationDecisionWhy
Rook & Ink first; YBRP as product (not brand center); Anglers as pathKEEPCorrect parent/product/feature boundaries.
Subdomains for products/ops/envs; paths for featuresKEEPSecurity and deploy isolation without hostname sprawl.
Preserve UC until explicit cutoverKEEPNon-negotiable production safety. Still gated.
Registry + hostname inventoryKEEPSource of truth for Studio and ops.
Parent route map (/work/*, /process, /insights, legal)KEEPImplemented on full-site preview lane.
YBRP modular monolith + single APIKEEPAvoid microservice and per-event API sprawl.
Studio as ops dashboard (not ERP)KEEPScaffolded routes; auth later.
Preserve wedding / infographics / unknown hostsKEEPIndependently-preserved or unknown. Do not delete.
Move parent site into apps/web nowREJECT (for now)Root Vite + separate UC build keeps production deploy simple and safe (ADR 0009).
Rewrite to Astro/NextREJECTExisting Vite multipage is sound; rewrite is risk without benefit.
preview.rookandink.com DNS nowREJECTworkers.dev previews until DNS approval.
Create all product DNS nowREJECTNo production DNS without approval.
Separate ybrp-platform repo immediatelyADAPTScaffold in apps/ybrp now; split when DB/API/admin justify it.
Full YBRP Postgres/API/admin nowADAPTDocument + public shell first; backend is Phase 4 to 5.
Empty service subpagesREJECTHandoff itself says do not publish empty service pages.
/platform architecture map on parent siteADAPT (add)Makes structure browsable on the brand preview, not docs-only.