Add a View on Goodreads button on Amazon book pages.

An Amazon listing already has the ISBN; jumping to the matching Goodreads
page should not require leaving the product page to search.

Co-authored-by: Cursor <cursoragent@cursor.com>

Author: ion05

AGENTS.md

@@ -57,8 +57,10 @@ src/lib/dnr/fit.ts      Chrome's RE2 check, resplitting a refused shard, coverag
 src/lib/draft.ts        What the edit form edits, and the pure parsing around it
 src/lib/text.ts         String helpers every surface shares
 src/lib/url.ts          Small URL helpers
+src/lib/amazon-book.ts  Amazon product HTML → ISBN → Goodreads URL. Pure.
 src/lib/install.ts      The onInstalled branch: starter pick, rule sync, welcome tab
 src/background.ts       MV3 service worker: listener registration, rule sync, omnibox
+src/content/            Isolated-world content scripts. `amazon-goodreads.ts` is IIFE-bundled.
 src/go/go.ts            Dispatch page: resolves and navigates
 src/ui/dom.ts           `el` / `mark` / `nextId`: the element builders every surface shares
 src/options/            Shortcut manager UI (below)

CHANGELOG.md

@@ -7,6 +7,13 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+
+- On Amazon.com book pages, a **View on Goodreads** button next to the title.
+  BunnyLol reads the ISBN off the product details (locally) and opens
+  `goodreads.com/book/isbn/…` for that same book. Pages without an ISBN are
+  left alone.
+
 ## [1.1.0] - 2026-09-02
 
 ### Added

PRIVACY.md

@@ -1,6 +1,6 @@
 # Privacy Policy
 
-Last updated: 2026-09-03
+Last updated: 2026-09-07
 
 ## Summary
 
@@ -37,15 +37,21 @@ one of your keywords, Chrome rewrites the URL to the extension's own
 then reads the query string, and it never goes anywhere else. Searches that
 do not match are left untouched and go to the search engine as normal.
 
-## What the extension cannot see
-
-BunnyLol has no content scripts, reads no page content, and has no access to
-your browsing history. It does not request the `tabs` permission. Three places
-open a tab, and all of them use only `chrome.tabs.create` and
-`chrome.tabs.update`, which do not require that permission: the toolbar popup
-(`src/popup/popup.ts`), the omnibox keyword (`src/background.ts`), and the
-welcome tab shown once on install (`src/lib/install.ts`). Neither call can
-read a tab, only point one at a URL.
+## What the extension can see
+
+BunnyLol does not request the `tabs` permission and has no access to your
+browsing history. Three places open a tab, and all of them use only
+`chrome.tabs.create` and `chrome.tabs.update`, which do not require that
+permission: the toolbar popup (`src/popup/popup.ts`), the omnibox keyword
+(`src/background.ts`), and the welcome tab shown once on install
+(`src/lib/install.ts`). Neither call can read a tab, only point one at a URL.
+
+The one content script is `src/content/amazon-goodreads.ts`, injected only on
+`amazon.com` product pages. It reads the page locally to find a labelled ISBN
+(or a JSON-LD `isbn`) and, when it finds one, draws a button. Clicking that
+button navigates your tab to Goodreads. The ISBN never leaves the browser
+except as the path of that navigation you started. Pages without an ISBN are
+untouched. No other site is injected into.
 
 ## Third parties
 

README.md

@@ -21,7 +21,8 @@ It is an independent, unofficial project, inspired by an internal tool at Meta c
 
 Manifest V3. 93 shortcuts ship across 180 keywords, and every one of them can be renamed, re-keyed,
 moved, switched off or deleted. No runtime dependencies, no network requests of its own, and nothing
-leaves your machine. See [PRIVACY.md](PRIVACY.md).
+leaves your machine. On Amazon.com book pages a content script reads the ISBN locally to offer a
+Goodreads button. See [PRIVACY.md](PRIVACY.md).
 
 ## A look inside
 
@@ -38,6 +39,11 @@ before you save:
 
 ![BunnyLol shortcut editor with a live MDN preview](docs/images/editor.png)
 
+On an Amazon.com book page, **View on Goodreads** sits under the title and opens
+the same book on Goodreads from its ISBN. No ISBN, no button.
+
+![View on Goodreads button on an Amazon book page](docs/images/amazon-goodreads-button.png)
+
 The toolbar popup gives you autocomplete when you do not want to leave the current page:
 
 <p align="center">

docs/images/amazon-goodreads-button.png

@@ -38,6 +38,13 @@
     "page": "options.html",
     "open_in_tab": true
   },
+  "content_scripts": [
+    {
+      "matches": ["https://www.amazon.com/*", "https://amazon.com/*", "https://smile.amazon.com/*"],
+      "js": ["amazon-goodreads.js"],
+      "run_at": "document_idle"
+    }
+  ],
   "web_accessible_resources": [
     {
       "resources": ["go.html"],

docs/images/goodreads-from-amazon.png

@@ -0,0 +1,149 @@
+/**
+ * Amazon book pages: a button next to the title that opens the same book on
+ * Goodreads. Isolated world, shadow DOM, textContent only. The page is never
+ * parsed as HTML we did not write (invariant 11).
+ */
+
+import { extractIsbn, goodreadsUrlForIsbn, isAmazonProductPath } from '../lib/amazon-book';
+
+const HOST_ID = 'bunnylol-goodreads';
+
+/**
+ * Copied by hand from design/tokens.css, the same way go.html does. This sheet
+ * lives in a shadow root on amazon.com, so the extension's tokens.css is not
+ * on the page, and Inter is not loaded there either. The sand fill and dark
+ * label are the brand pair that already meets contrast on a white Amazon page.
+ */
+const SHADOW_CSS = `
+:host { all: initial; }
+.wrap { display: block; margin: 8px 0 12px; }
+button {
+  font: 600 13px/1.3 -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, system-ui, sans-serif;
+  background: #e1ab76;
+  color: #1a1a1a;
+  border: 0;
+  border-radius: 6px;
+  padding: 6px 12px;
+  cursor: pointer;
+  display: inline-flex;
+  align-items: center;
+  gap: 6px;
+}
+button:hover { background: #d69b5d; }
+button:focus-visible { outline: 2px solid #895420; outline-offset: 2px; }
+svg { width: 14px; height: 14px; flex: none; }
+`;
+
+let mountedFor: string | null = null;
+let debounce = 0;
+let lastHref = location.href;
+
+function refresh(): void {
+  if (location.href !== lastHref) {
+    lastHref = location.href;
+    unmount();
+    mountedFor = null;
+  }
+
+  if (!isAmazonProductPath(location.pathname)) {
+    unmount();
+    mountedFor = null;
+    return;
+  }
+
+  if (mountedFor && document.getElementById(HOST_ID)) return;
+
+  const isbn = extractIsbn(document.documentElement.innerHTML);
+  if (!isbn) {
+    unmount();
+    mountedFor = null;
+    return;
+  }
+
+  const url = goodreadsUrlForIsbn(isbn);
+  unmount();
+  mount(url);
+  mountedFor = url;
+}
+
+function schedule(): void {
+  window.clearTimeout(debounce);
+  debounce = window.setTimeout(refresh, 120);
+}
+
+function mount(url: string): void {
+  const host = document.createElement('div');
+  host.id = HOST_ID;
+  const shadow = host.attachShadow({ mode: 'open' });
+
+  const style = document.createElement('style');
+  style.textContent = SHADOW_CSS;
+
+  const wrap = document.createElement('div');
+  wrap.className = 'wrap';
+
+  const button = document.createElement('button');
+  button.type = 'button';
+  button.setAttribute('aria-label', 'View this book on Goodreads (BunnyLol)');
+  button.title = 'Open the Goodreads page for this book';
+  button.append(bookIcon(), document.createTextNode('View on Goodreads'));
+  button.addEventListener('click', (event) => {
+    event.preventDefault();
+    event.stopPropagation();
+    location.assign(url);
+  });
+
+  wrap.append(button);
+  shadow.append(style, wrap);
+
+  const target = mountPoint();
+  if (target) {
+    target.insertAdjacentElement('afterend', host);
+    return;
+  }
+  host.style.position = 'fixed';
+  host.style.top = '72px';
+  host.style.right = '16px';
+  host.style.zIndex = '2147483646';
+  document.documentElement.append(host);
+}
+
+function unmount(): void {
+  document.getElementById(HOST_ID)?.remove();
+}
+
+function mountPoint(): Element | null {
+  return (
+    document.querySelector('#productTitle') ??
+    document.querySelector('#title') ??
+    document.querySelector('#titleSection')
+  );
+}
+
+function bookIcon(): SVGSVGElement {
+  const svg = document.createElementNS('http://www.w3.org/2000/svg', 'svg');
+  svg.setAttribute('viewBox', '0 0 16 16');
+  svg.setAttribute('aria-hidden', 'true');
+  svg.setAttribute('focusable', 'false');
+  const path = document.createElementNS('http://www.w3.org/2000/svg', 'path');
+  path.setAttribute('d', 'M3 2.5h6.5A2.5 2.5 0 0 1 12 5v8.5H5.5A2.5 2.5 0 0 0 3 11V2.5zm0 0V11');
+  path.setAttribute('fill', 'none');
+  path.setAttribute('stroke', 'currentColor');
+  path.setAttribute('stroke-width', '1.5');
+  path.setAttribute('stroke-linecap', 'round');
+  path.setAttribute('stroke-linejoin', 'round');
+  svg.append(path);
+  return svg;
+}
+
+function boot(): void {
+  refresh();
+  const observer = new MutationObserver(schedule);
+  observer.observe(document.documentElement, { childList: true, subtree: true });
+}
+
+if (document.readyState === 'loading') {
+  document.addEventListener('DOMContentLoaded', boot, { once: true });
+} else {
+  boot();
+}

public/manifest.json

@@ -0,0 +1,92 @@
+/**
+ * Turn an Amazon product page into a Goodreads destination.
+ *
+ * Pure: HTML in, a URL or null out. The content script is the only runtime
+ * caller; the tests are the other. Matching the labelled ISBN fields (and
+ * JSON-LD `isbn`) is the whole signal. Scanning for any 13-digit run would
+ * follow a carousel ASIN onto the wrong book.
+ */
+
+const ISBN_13_RE = /97[89](?:\D*\d){10}/;
+const ISBN_10_RE = /\b\d(?:\D*\d){8}[\dXx]\b/;
+
+/** Amazon inserts these between "ISBN-13" and the number. */
+const LABEL_WINDOW = 400;
+
+export function goodreadsUrlForIsbn(isbn: string): string {
+  return `https://www.goodreads.com/book/isbn/${isbn}`;
+}
+
+/** `/dp/ASIN` and the older product paths. Search pages are not books. */
+export function isAmazonProductPath(pathname: string): boolean {
+  return /\/(?:dp|gp\/product|gp\/aw\/d)\//i.test(pathname);
+}
+
+/**
+ * ISBN-13 wins when both are on the page, because that is the one Goodreads
+ * indexes most reliably. Null when the page never names a book.
+ */
+export function extractIsbn(html: string): string | null {
+  return isbnFromJsonLd(html) ?? isbnAfterLabel(html, 'ISBN-13') ?? isbnAfterLabel(html, 'ISBN-10');
+}
+
+function isbnAfterLabel(html: string, label: 'ISBN-13' | 'ISBN-10'): string | null {
+  const haystack = html.toUpperCase();
+  const at = haystack.indexOf(label);
+  if (at < 0) return null;
+  const snippet = stripMarkup(html.slice(at + label.length, at + label.length + LABEL_WINDOW));
+  return label === 'ISBN-13' ? findIsbn13(snippet) : findIsbn10(snippet);
+}
+
+function isbnFromJsonLd(html: string): string | null {
+  const re = /<script\b[^>]*type=["']application\/ld\+json["'][^>]*>([\s\S]*?)<\/script>/gi;
+  for (const match of html.matchAll(re)) {
+    const body = match[1];
+    if (!body) continue;
+    try {
+      const found = isbnFromUnknown(JSON.parse(body) as unknown);
+      if (found) return found;
+    } catch {
+      // Amazon's JSON-LD is occasionally not JSON. The labelled fields still
+      // sit in the product details, so a parse miss is not a miss of the book.
+    }
+  }
+  return null;
+}
+
+function isbnFromUnknown(value: unknown): string | null {
+  if (typeof value === 'string') return findIsbn13(value) ?? findIsbn10(value);
+  if (Array.isArray(value)) {
+    for (const entry of value) {
+      const found = isbnFromUnknown(entry);
+      if (found) return found;
+    }
+    return null;
+  }
+  if (!value || typeof value !== 'object') return null;
+  const record = value as Record<string, unknown>;
+  if (Object.hasOwn(record, 'isbn')) {
+    const found = isbnFromUnknown(record.isbn);
+    if (found) return found;
+  }
+  if (Object.hasOwn(record, '@graph')) return isbnFromUnknown(record['@graph']);
+  return null;
+}
+
+function findIsbn13(text: string): string | null {
+  const match = text.match(ISBN_13_RE);
+  if (!match) return null;
+  const digits = match[0].replace(/\D/g, '');
+  return digits.length === 13 ? digits : null;
+}
+
+function findIsbn10(text: string): string | null {
+  const match = text.match(ISBN_10_RE);
+  if (!match) return null;
+  const compact = match[0].toUpperCase().replace(/[^0-9X]/g, '');
+  return compact.length === 10 ? compact : null;
+}
+
+function stripMarkup(html: string): string {
+  return html.replace(/<[^>]+>/g, ' ').replace(/&nbsp;|&#160;|&zwnj;|&#8203;/gi, ' ');
+}