chore: consolidate open PRs — CSS fix, tech integrations, testing skill

.agents/skills/testing-lighthouse/SKILL.md

@@ -9,6 +9,49 @@ The Lighthouse Scanner lives at `/lighthouse` in the Next.js app (`apps/web`). I
 - **Cloudflare Pages preview**: Every PR gets a preview deploy. Check the PR comments from `cloudflare-workers-and-pages[bot]` for the preview URL (format: `https://<hash>.dba-92r.pages.dev`).
 - **Production**: `https://designedbyanthony.com/lighthouse`
 
+## Important: Pages vs Worker Deployment
+
+This repo has **two separate deploy targets**:
+- **Cloudflare Pages** (`apps/web`) — deploys automatically via PR/merge CI
+- **Cloudflare Worker** (`apps/api`) — deploys separately via `wrangler deploy` from `apps/api/`
+
+Cloudflare Pages preview URLs only include frontend changes. **API Worker changes (routes in `apps/api/src/routes/`) are NOT included in Pages previews.** To test API changes, use local `wrangler dev` (see below) or wait for a separate Worker deployment.
+
+## Testing API Routes Locally
+
+To test API Worker routes (e.g., `/api/report/:id/pdf`, `/api/audit`, `/api/audit/email-summary`):
+
+1. **Start the API locally:**
+   ```bash
+   cd apps/api && npx wrangler dev --local
+   ```
+   This starts the Worker on `http://localhost:8788` with local KV.
+
+2. **Seed test data into local KV** (needed for endpoints that read from the report store):
+   ```bash
+   cd apps/api && npx wrangler kv key put "DBA-TEST1234" \
+     --binding AUDIT_REPORTS_KV --local \
+     "$(cat /path/to/test-report.json)"
+   ```
+   The test report JSON must match the full stored payload structure from `audit.ts` — including nested `sitewide.robotsTxt.exists`, `sitewide.sitemap.exists`, `backlinks.found`, `places.found`, `indexCoverage.found` fields. Empty objects (`{}`) for these fields will cause `buildAuditPdf` to crash with "Cannot read properties of undefined".
+
+3. **Verify with curl:**
+   ```bash
+   # Health check
+   curl -s http://localhost:8788/health
+
+   # Test PDF headers
+   curl -sD - http://localhost:8788/api/report/DBA-TEST1234/pdf -o /tmp/test.pdf
+
+   # Test CORS preflight
+   curl -sD - -X OPTIONS http://localhost:8788/api/audit/email-summary \
+     -H "Origin: https://designedbyanthony.com" -o /dev/null
+   ```
+
+## Elysia Response Pattern
+
+When testing Elysia routes that return raw `Response` objects: in Elysia, `set.headers` is ignored when returning a raw `Response`. Headers must be passed via the `Response` constructor. If you see missing headers on a raw Response return, check for this pattern mismatch. Routes that return plain objects (JSON) can use `set.headers` normally.
+
 ## Key Visual Elements to Verify
 
 When testing Lighthouse UI changes, compare against the main site homepage (`/`) for consistency:
@@ -33,6 +76,7 @@ When testing Lighthouse UI changes, compare against the main site homepage (`/`)
 - **CSS class name mismatches**: When refactoring CSS class names, always check that components still reference the correct names. Use `grep -oP 'className="[^"]*lh-[^"]*"' <component>.tsx | grep -oP 'lh-[a-z0-9_-]+'` to extract class names from components and verify they exist in the CSS file.
 - **Local dev 500 errors**: The homepage might 500 locally due to missing env vars, but `/lighthouse` usually works. Use the Cloudflare Pages preview URL as a reliable alternative.
 - **Bronze is intentional in some places**: The submit button CTA and decorative accent rules intentionally keep bronze — don't flag these as bugs.
+- **Empty nested objects in test data**: When seeding test reports for PDF generation, ensure all nested objects (`sitewide`, `backlinks`, `places`, `indexCoverage`) have their full expected structure. `buildAuditPdf` accesses nested properties like `sitewide.robotsTxt.exists` and will crash on empty objects.
 
 ## Devin Secrets Needed
 

STATUS.md

@@ -2,14 +2,23 @@
 
 > **Note:** This file tracks migration and release notes for the **Turborepo Cloudflare** app (`apps/web` Next.js Pages + `apps/api` Elysia Worker, `bun`, `bun.lock`). Older single-app / Astro-era detail is archived below for context — see [README.md](README.md) and [AGENTS.md](AGENTS.md).
 
+## Codebase cleanup, perf optimization & tech integrations (2026-04-30)
+
+- Removed Zustand badge from the "Our Edge" stack display — dependency was previously removed but badge remained.
+- Removed stray `console.warn` in the competitor-scan catch block of `apps/api/src/routes/audit.ts`.
+- Cleaned up outdated Firebase/Firestore references in `lighthouse2.md` (now reflects Cloudflare Pages/Workers + KV/D1).
+- Added `manifest.webmanifest` to `apps/web/public/` with PWA metadata — triggers PWA detection on Wappalyzer/BuiltWith.
+- Added JSON-LD structured data (`Organization`, `LocalBusiness`, `WebSite`) in root layout — zero render cost, boosts SEO.
+- Verification: `bun run lint`, `bun run build` pass from the repo root.
+
 ## Lockfile fix + backend PDF offload (2026-04-30)
 
+
 - **Lockfile fix (P0):** Removed conflicting workspace-level `overrides` from `apps/web/package.json` (stale `postcss ^8.4.38` vs root `^8.5.12`, duplicate `uuid`, unrecorded `yaml@<2.8.3`). Moved `yaml@<2.8.3` override to root `package.json`. Regenerated `bun.lock` — `bun install --frozen-lockfile` now passes in CI.
 - **Dependency hygiene:** Relocated heavy backend-only packages (`@google/genai`, `googleapis`, `cheerio`, `jspdf`) from `apps/web` to `packages/shared` where they are actually imported. Frontend bundle no longer ships these libraries.
 - **PDF offload to API Worker:** Created `GET /api/report/:id/pdf` on the ElysiaJS Worker (`apps/api/src/routes/reportPdf.ts`). The route fetches the persisted report from the KV store, calls `buildAuditPdf` server-side, and streams the PDF back with `Content-Disposition: attachment`. `AuditResults.tsx` now fetches from this endpoint instead of importing `jspdf` client-side — removes ~250 KB from the browser bundle and eliminates client-side PDF generation latency.
 - Added `./lighthouse/*` export to `packages/shared` so `AuditData` types are accessible from any workspace.
 - Verification: `bun install --frozen-lockfile`, `bun run build`, `bun run typecheck`, and `bun run lint` pass from the repo root.
-
 ## Micro SaaS store /tools page build-out (2026-04-30)
 
 - Replaced the "Coming Soon" waitlist on `/tools` with a full product catalog showcasing 6 Stripe-backed micro-SaaS products: SiteScan, ReviewPilot, ClientHub, LocalRank, TestiFlow, and ContentMill.

apps/api/src/routes/audit.ts

@@ -282,8 +282,8 @@ export const auditRoute = new Elysia({ aot: false }).post(
 					placesData.primaryType,
 					3,
 				);
-			} catch (compErr) {
-				console.warn("Competitor scan failed:", compErr);
+			} catch {
+				/* competitor scan is best-effort — swallow errors silently */
 			}
 
 			const categories = lighthouse.categories;

apps/web/public/manifest.webmanifest

@@ -0,0 +1,26 @@
+{
+	"name": "Designed by Anthony",
+	"short_name": "DBA",
+	"description": "Custom web design and local SEO for service businesses in the Mohawk Valley and Central New York.",
+	"start_url": "/",
+	"display": "standalone",
+	"background_color": "#0f1218",
+	"theme_color": "#0f1218",
+	"icons": [
+		{
+			"src": "/site-icon-48.png",
+			"sizes": "48x48",
+			"type": "image/png"
+		},
+		{
+			"src": "/site-icon-192.png",
+			"sizes": "192x192",
+			"type": "image/png"
+		},
+		{
+			"src": "/site-icon-512.png",
+			"sizes": "512x512",
+			"type": "image/png"
+		}
+	]
+}

apps/web/src/app/layout.tsx

@@ -1,5 +1,6 @@
 import "@/design-system/dba-global.css";
 import { CrispBootstrap } from "@/components/CrispBootstrap";
+import { JsonLd } from "@/components/JsonLd";
 import "@/styles/layout-shell.css";
 import type { Metadata, Viewport } from "next";
 import type { ReactNode } from "react";
@@ -25,6 +26,7 @@ export const metadata: Metadata = {
 	},
 	description:
 		"Custom web design and local SEO for service businesses in the Mohawk Valley and Central New York.",
+	manifest: "/manifest.webmanifest",
 	appleWebApp: {
 		capable: true,
 		statusBarStyle: "black-translucent",
@@ -77,6 +79,9 @@ export default function RootLayout({ children }: { children: ReactNode }) {
 			data-scroll-behavior="smooth"
 			data-lead-webhook={leadWebhookDefault || undefined}
 		>
+			<head>
+				<JsonLd />
+			</head>
 			<body>
 				{/* Google Tag Manager (noscript) */}
 				<noscript>

apps/web/src/components/JsonLd.tsx

@@ -0,0 +1,25 @@
+import {
+	buildBaseOrganizationSchema,
+	buildBaseWebsiteSchema,
+} from "@/lib/seo";
+
+const ENTRIES = [buildBaseOrganizationSchema(), buildBaseWebsiteSchema()];
+
+export function JsonLd() {
+	return (
+		<>
+			{ENTRIES.map((entry) => {
+				const id =
+					typeof entry["@id"] === "string" ? entry["@id"] : JSON.stringify(entry);
+				return (
+					<script
+						key={id}
+						type="application/ld+json"
+						// biome-ignore lint/security/noDangerouslySetInnerHtml: structured data requires raw JSON injection
+						dangerouslySetInnerHTML={{ __html: JSON.stringify(entry) }}
+					/>
+				);
+			})}
+		</>
+	);
+}

apps/web/src/components/marketing/EnrichedPages.tsx

@@ -602,7 +602,6 @@ export function OurEdgePage() {
 										href: "https://deepmind.google/technologies/gemini/",
 									},
 									{ label: "Zod", href: "https://zod.dev" },
-									{ label: "Zustand", href: "https://zustand.docs.pmnd.rs" },
 								].map(({ label, href }) => (
 									<a
 										key={label}

apps/web/src/styles/theme.css

@@ -1773,7 +1773,7 @@ body:not(.has-progress-bar) #reading-progress-bar {
 
 @media (max-width: 600px) {
 	.salesforce-form-actions {
-		justify-content: stretch;
+		justify-content: center;
 	}
 
 	.salesforce-form-actions .btn {

lighthouse2.md

@@ -1,6 +1,6 @@
 # Lighthouse Scanner — product description & operator guide
 
-This document is the **single source of truth** for how we **describe** the scanner (marketing, OG tags, splash copy) and what **you** configure in **Firebase App Hosting** / env. Structure is inspired by common open-source SEO audit READMEs (for example [vchaitanyachowdari/seo-audits-generator](https://github.com/vchaitanyachowdari/seo-audits-generator)) — clear feature blocks, stack, env, usage — adapted to **this repo’s real implementation**, not a generic SaaS template.
+This document is the **single source of truth** for how we **describe** the scanner (marketing, OG tags, splash copy) and what **you** configure in **Cloudflare Pages / Workers** env. Structure is inspired by common open-source SEO audit READMEs (for example [vchaitanyachowdari/seo-audits-generator](https://github.com/vchaitanyachowdari/seo-audits-generator)) — clear feature blocks, stack, env, usage — adapted to **this repo’s real implementation**, not a generic SaaS template.
 
 ---
 
@@ -21,7 +21,7 @@ This document is the **single source of truth** for how we **describe** the scan
 | **Authority / links** | “Backlink snapshot when configured” | **Moz API** when `MOZ_API_CREDENTIALS` / token is set; otherwise omit or show “not configured” in UI |
 | **Conversion / UX** | “Trust + conversion lens” | AI **conversionScore** + CTA/schema/tel/form signals from HTML |
 | **AI report** | Plain-English executive summary + top fixes | **Gemini** (`generateAiInsight`) with fallback |
-| **Persistence** | Sharable report when storage works | Firestore report store when configured |
+| **Persistence** | Sharable report when storage works | KV / D1 report store when configured |
 | **CRM / leads** | Optional | Convex webhooks, Sheets, Gmail — see `.env.example` (`AUDIT_LEAD_WEBHOOK_*`, `LEAD_WEBHOOK_URL`) |
 
 Do **not** claim: unlimited full-site crawl like Semrush/Ahrefs, JS-rendered crawl of every route, or “Moz DA” unless Moz is configured.