platforms

What Upgrades next 15.3.6 → 16.2.10 with react 19.2.7. → , exported function renamed to (middleware is deprecated in 16) Dropped from the dev script, it's the default now / changes are auto-applied by README: fixed remaining middleware references, Node minimum is 20.9+ No app code changes needed ( was already awaited). Targeting 16.2.10 specifically to pick up the May proxy-bypass fixes, which matter here since the proxy is the tenant routing boundary. Validation Verified with and a local smoke test: subdomain root rewrites to , on a subdomain redirects to .

What Captures the owner's goal of publishing a mobile app to the app stores later, at the lowest build and maintenance cost, as a standing constraint on all current feature work. No mobile app is being built now — this is documentation and agent guidance only. The intended low-lift path: a PWA or a thin Capacitor shell that loads the live web app, with the web app staying the single source of truth. A native rewrite (React Native / Expo) is explicitly rejected. Changes — new "Standing constraint: keep the door open to a mobile app" block (build mobile-first; flag any door-closing decision in one sentence before proceeding), a matching "What NOT to Do" guardrail, a docs-table entry, and an update trigger. — the working reference: do this / don't do now / decisions that need a heads-up to the owner. — ADR recording the rationale. — indexes the new reference + ADR. — deferred mobile section (not to start without owner go-ahead). Net behavior change for agents Build mobile-first by default, and pause to surface the tradeoff before any decision that would make a future mobile app harder — desktop-only UX, hover/right-click-only interactions, browser-only auth redirects, rendering/hosting model switches, or webview-incompatible APIs. The owner decides. Risk Docs only — no application code or behavior changes. 🤖 Generated with Claude Code