fix(seo): strip HTML from meta description + OG/Twitter/Schema.org

Article excerpts sometimes contain raw <p> tags; Next.js HTML-escapes
them into the page <meta>, surfacing as literal '&lt;p&gt;' in view-source
and breaking SEO previews. cleanExcerpt() strips tags + entities for all
plain-text meta surfaces.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/app/articles/[slug]/page.tsx

@@ -1,4 +1,5 @@
 import type { Metadata } from "next";
+import { cleanExcerpt } from "@/lib/articles/excerpt";
 import { redirect } from "next/navigation";
 import ArticleDetailPage from "./ArticleDetailClient";
 import type { Article } from "@/lib/articles/types";
@@ -53,12 +54,12 @@ export async function generateMetadata({ params }: PageProps): Promise<Metadata>
   }
   return {
     title: article.title,
-    description: article.excerpt,
+    description: cleanExcerpt(article.excerpt),
     alternates: { canonical: `/articles/${article.slug}` },
     openGraph: {
       type: "article",
       title: article.title,
-      description: article.excerpt,
+      description: cleanExcerpt(article.excerpt),
       url: `/articles/${article.slug}`,
       images: [{ url: article.image, width: 1200, height: 630, alt: article.alt }],
       authors: [article.author],
@@ -68,7 +69,7 @@ export async function generateMetadata({ params }: PageProps): Promise<Metadata>
     twitter: {
       card: "summary_large_image",
       title: article.title,
-      description: article.excerpt,
+      description: cleanExcerpt(article.excerpt),
       images: [article.image],
       creator: "@AV Beat",
     },
@@ -93,7 +94,7 @@ export default async function ArticlePage({ params }: PageProps) {
     "@context": "https://schema.org",
     "@type": "Article",
     headline: article.title,
-    description: article.excerpt,
+    description: cleanExcerpt(article.excerpt),
     image: [absoluteImage(article.image)],
     author: { "@type": "Person", name: article.author },
     publisher: {

src/app/news/[slug]/page.tsx

@@ -8,6 +8,7 @@ import {
   getRelatedForumThreads, type RelatedForumThread,
 } from "@/lib/articles/legacy-source";
 import { linkifyAndExtractMentions } from "@/lib/company-mentions";
+import { cleanExcerpt } from "@/lib/articles/excerpt";
 import CompanyMentionsHover from "@/components/CompanyMentionsHover";
 import CompaniesInThisStory from "@/components/CompaniesInThisStory";
 import NewsArticleDetailClient from "./NewsArticleDetailClient";
@@ -58,14 +59,19 @@ export async function generateMetadata({ params }: PageProps): Promise<Metadata>
   if (!article) {
     return { title: "Article Not Found", description: "The article you are looking for does not exist." };
   }
+  // Strip HTML from excerpt before stuffing into <meta description> /
+  // openGraph / twitter — those MUST be plain text. Raw excerpts sometimes
+  // contain <p>...</p>, which Next.js then HTML-escapes, surfacing as
+  // literal `&lt;p&gt;` in page source.
+  const cleanDesc = cleanExcerpt(article.excerpt);
   return {
     title: `${article.title} — AV Beat`,
-    description: article.excerpt,
+    description: cleanDesc,
     alternates: { canonical: `/news/${article.slug}` },
     openGraph: {
       type: "article",
       title: article.title,
-      description: article.excerpt,
+      description: cleanDesc,
       url: `/news/${article.slug}`,
       images: [{ url: article.image, width: 1200, height: 630, alt: article.alt }],
       authors: [article.author],
@@ -75,7 +81,7 @@ export async function generateMetadata({ params }: PageProps): Promise<Metadata>
     twitter: {
       card: "summary_large_image",
       title: article.title,
-      description: article.excerpt,
+      description: cleanDesc,
       images: [article.image],
       creator: "@AV Beat",
     },
@@ -106,7 +112,7 @@ export default async function NewsArticlePage({ params }: PageProps) {
     "@context": "https://schema.org",
     "@type": "NewsArticle",
     headline: article.title,
-    description: article.excerpt,
+    description: cleanExcerpt(article.excerpt),
     image: [absoluteImage(article.image)],
     author: { "@type": "Person", name: article.author },
     publisher: {