fix(adapters,output): repair site drift and markdown/console output bugs

plugins/amazon/README.md

@@ -21,3 +21,7 @@ webcmd plugin install github:agentrhq/webcmd/amazon
 | `webcmd amazon product` | Amazon product page facts for candidate validation |
 | `webcmd amazon search` | Amazon search results for product discovery and coarse filtering |
 | `webcmd amazon whoami` | Show the current logged-in amazon account |
+
+## Notes
+
+- A product or review URL from a sibling marketplace (`amazon.co.uk`, `amazon.de`, `amazon.com.au`, …) is read on that marketplace, and the emitted URLs stay on it. A bare ASIN still defaults to `amazon.com`.

plugins/amazon/discussion.js

@@ -1,6 +1,6 @@
 import { AuthRequiredError, CommandExecutionError } from '@agentrhq/webcmd/errors';
 import { cli, Strategy } from '@agentrhq/webcmd/registry';
-import { buildProductUrl, buildDiscussionUrl, buildProvenance, cleanText, extractAsin, normalizeProductUrl, parseRatingValue, parseReviewCount, trimRatingPrefix, uniqueNonEmpty, assertUsableState, gotoAndReadState, } from './shared.js';
+import { DOMAIN, amazonHostFromInput, buildProductUrl, buildDiscussionUrl, buildProvenance, cleanText, extractAsin, normalizeProductUrl, parseRatingValue, parseReviewCount, trimRatingPrefix, uniqueNonEmpty, assertUsableState, gotoAndReadState, } from './shared.js';
 function normalizeDiscussionPayload(payload) {
     const sourceUrl = cleanText(payload.href) || buildDiscussionUrl(payload.href ?? '');
     const asin = extractAsin(payload.href ?? '') ?? null;
@@ -9,7 +9,7 @@ function normalizeDiscussionPayload(payload) {
     const provenance = buildProvenance(sourceUrl);
     return {
         asin,
-        product_url: asin ? normalizeProductUrl(asin) : null,
+        product_url: asin ? normalizeProductUrl(sourceUrl) : null,
         discussion_url: sourceUrl,
         ...provenance,
         average_rating_text: averageRatingText,
@@ -71,7 +71,7 @@ async function readDiscussionPayload(page, input, limit) {
     const productState = await gotoAndReadState(page, productUrl, 2500, 'discussion');
     assertUsableState(productState, 'discussion');
     if (isSignInState(reviewState) && isSignInState(productState)) {
-        throw new AuthRequiredError('amazon.com', 'Amazon review discussion requires an active signed-in Amazon session in the shared Chrome profile.');
+        throw new AuthRequiredError(amazonHostFromInput(input) ?? DOMAIN, 'Amazon review discussion requires an active signed-in Amazon session in the shared Chrome profile.');
     }
     const productPayload = await readCurrentDiscussionPayload(page, limit);
     if (hasDiscussionSummary(productPayload)) {
@@ -111,7 +111,8 @@ cli({
         const payload = await readDiscussionPayload(page, input, limit);
         const normalized = normalizeDiscussionPayload(payload);
         if (!normalized.average_rating_text && !normalized.total_review_count_text) {
-            throw new CommandExecutionError('amazon discussion page did not expose review summary', 'The review page may have changed or hit a robot check. Open the review page in Chrome and retry.');
+            const landedUrl = cleanText(payload.href) || buildDiscussionUrl(input);
+            throw new CommandExecutionError(`amazon discussion page did not expose review summary (landed on ${landedUrl})`, 'The review page may have changed or hit a robot check. Open the review page in Chrome and retry.');
         }
         return [normalized];
     },

plugins/amazon/product.js

@@ -83,7 +83,8 @@ cli({
         const input = String(kwargs.input ?? '');
         const payload = await readProductPayload(page, input);
         if (!cleanText(payload.product_title)) {
-            throw new CommandExecutionError('amazon product page did not expose product content', 'The product page may have changed or hit a robot check. Open the product page in Chrome and retry.');
+            const landedUrl = cleanText(payload.href) || buildProductUrl(input);
+            throw new CommandExecutionError(`amazon product page did not expose product content (landed on ${landedUrl})`, 'The product page may have changed or hit a robot check. Open the product page in Chrome and retry.');
         }
         return [normalizeProductPayload(payload)];
     },

plugins/amazon/shared.js

@@ -20,6 +20,42 @@ export const PRIMARY_PRICE_SELECTORS = [
     '#priceblock_dealprice',
     '#tp_price_block_total_price_ww',
 ];
+// Keep this explicit because these hosts are navigation targets in the user's
+// signed-in browser. A shape-only `amazon.<tld>` pattern also accepts unrelated
+// registrable domains such as amazon.shop or amazon.zip.
+const MARKETPLACE_DOMAINS = new Set([
+    'amazon.com',
+    'amazon.ca',
+    'amazon.com.mx',
+    'amazon.com.br',
+    'amazon.co.uk',
+    'amazon.de',
+    'amazon.fr',
+    'amazon.it',
+    'amazon.es',
+    'amazon.nl',
+    'amazon.pl',
+    'amazon.se',
+    'amazon.com.be',
+    'amazon.ie',
+    'amazon.com.tr',
+    'amazon.ae',
+    'amazon.sa',
+    'amazon.eg',
+    'amazon.co.za',
+    'amazon.in',
+    'amazon.co.jp',
+    'amazon.com.au',
+    'amazon.sg',
+]);
+function isAmazonMarketplaceHost(hostname) {
+    const normalized = cleanText(hostname).toLowerCase().replace(/\.$/, '');
+    for (const domain of MARKETPLACE_DOMAINS) {
+        if (normalized === domain || normalized.endsWith(`.${domain}`))
+            return true;
+    }
+    return false;
+}
 const ROBOT_TEXT_PATTERNS = [
     'Sorry, we just need to make sure you\'re not a robot',
     'Enter the characters you see below',
@@ -91,19 +127,33 @@ export function extractAsin(input) {
     const match = normalized.match(/\/(?:dp|gp\/product|product-reviews)\/([A-Z0-9]{10})/i);
     return match ? match[1].toUpperCase() : null;
 }
+export function amazonHostFromInput(input) {
+    const normalized = cleanText(input);
+    if (!normalized)
+        return null;
+    try {
+        const url = new URL(normalized);
+        return isAmazonMarketplaceHost(url.hostname) ? url.hostname : null;
+    }
+    catch {
+        return null;
+    }
+}
 export function buildProductUrl(input) {
     const asin = extractAsin(input);
     if (!asin) {
         throw new ArgumentError('amazon product expects an ASIN or product URL', 'Example: webcmd amazon product B0FJS72893');
     }
-    return `${PRODUCT_URL_PREFIX}${asin}`;
+    const host = amazonHostFromInput(input);
+    return host ? `https://${host}/dp/${asin}` : `${PRODUCT_URL_PREFIX}${asin}`;
 }
 export function buildDiscussionUrl(input) {
     const asin = extractAsin(input);
     if (!asin) {
         throw new ArgumentError('amazon discussion expects an ASIN or product URL', 'Example: webcmd amazon discussion B0FJS72893');
     }
-    return `${DISCUSSION_URL_PREFIX}${asin}`;
+    const host = amazonHostFromInput(input);
+    return host ? `https://${host}/product-reviews/${asin}` : `${DISCUSSION_URL_PREFIX}${asin}`;
 }
 function getRankingSpec(listType) {
     return AMAZON_RANKING_SPECS[listType];
@@ -206,7 +256,7 @@ export function resolveBestsellersUrl(input) {
 export function canonicalizeAmazonUrl(input) {
     try {
         const url = new URL(input);
-        if (!url.hostname.endsWith(DOMAIN)) {
+        if (!isAmazonMarketplaceHost(url.hostname)) {
             throw new Error('not-amazon');
         }
         return url.toString();
@@ -230,7 +280,7 @@ export function normalizeProductUrl(value) {
     const normalized = cleanText(value);
     const asin = extractAsin(normalized);
     if (asin)
-        return buildProductUrl(asin);
+        return buildProductUrl(normalized);
     return toAbsoluteAmazonUrl(normalized);
 }
 export function parsePriceText(text) {
@@ -347,8 +397,11 @@ export function assertUsableState(state, action) {
 export const __test__ = {
     buildSearchUrl,
     extractAsin,
+    amazonHostFromInput,
     buildProductUrl,
     buildDiscussionUrl,
+    normalizeProductUrl,
+    canonicalizeAmazonUrl,
     resolveBestsellersUrl,
     resolveRankingUrl,
     isSupportedRankingPath,

plugins/amazon/test/discussion.test.js

@@ -41,6 +41,68 @@ describe('amazon discussion normalization', () => {
     ]);
   });
 
+  it('keeps the review marketplace in every emitted url', () => {
+    const result = __test__.normalizeDiscussionPayload({
+      href: 'https://www.amazon.co.uk/product-reviews/B0FGCPFY9L',
+      average_rating_text: '4.4 out of 5',
+      total_review_count_text: '40 global ratings',
+      qa_links: [],
+      review_samples: [],
+    });
+
+    expect(result.asin).toBe('B0FGCPFY9L');
+    expect(result.discussion_url).toBe('https://www.amazon.co.uk/product-reviews/B0FGCPFY9L');
+    expect(result.product_url).toBe('https://www.amazon.co.uk/dp/B0FGCPFY9L');
+  });
+
+  it('requests the review page on the marketplace the input names', async () => {
+    const command = getRegistry().get('amazon/discussion');
+    const page = createPageMock([
+      {
+        href: 'https://www.amazon.co.uk/product-reviews/B0FGCPFY9L',
+        title: 'Amazon.co.uk: Example product',
+        body_text: 'Customer reviews',
+      },
+      {
+        href: 'https://www.amazon.co.uk/product-reviews/B0FGCPFY9L',
+        average_rating_text: '4.4 out of 5',
+        total_review_count_text: '40 global ratings',
+        review_samples: [],
+      },
+    ]);
+
+    await command.func(page, { input: 'https://www.amazon.co.uk/product-reviews/B0FGCPFY9L', limit: 1 });
+
+    expect(page.goto.mock.calls[0][0]).toBe('https://www.amazon.co.uk/product-reviews/B0FGCPFY9L');
+  });
+
+  it('names the loaded url when neither page exposes a review summary', async () => {
+    const command = getRegistry().get('amazon/discussion');
+    const emptyPayload = { href: 'https://www.amazon.co.uk/dp/B0FGCPFY9L', average_rating_text: '', total_review_count_text: '', review_samples: [] };
+    const page = createPageMock([
+      { href: 'https://www.amazon.co.uk/product-reviews/B0FGCPFY9L', title: 'Amazon.co.uk', body_text: 'Customer reviews' },
+      emptyPayload,
+      { href: 'https://www.amazon.co.uk/dp/B0FGCPFY9L', title: 'Amazon.co.uk', body_text: 'Product' },
+      emptyPayload,
+    ]);
+
+    await expect(command.func(page, { input: 'https://www.amazon.co.uk/dp/B0FGCPFY9L', limit: 1 }))
+      .rejects.toThrow('landed on https://www.amazon.co.uk/dp/B0FGCPFY9L');
+  });
+
+  it('points a gated non-US review page at that marketplace, not the US store', async () => {
+    const command = getRegistry().get('amazon/discussion');
+    const signIn = { href: 'https://www.amazon.co.uk/ap/signin', title: 'Amazon Sign-In', body_text: 'Sign in Create account' };
+    const page = createPageMock([
+      signIn,
+      { href: signIn.href, average_rating_text: '', total_review_count_text: '', review_samples: [] },
+      signIn,
+    ]);
+
+    await expect(command.func(page, { input: 'https://www.amazon.co.uk/dp/B0FGCPFY9L', limit: 1 }))
+      .rejects.toMatchObject({ domain: 'www.amazon.co.uk' });
+  });
+
   it('falls back to the product page when the review page redirects to sign-in', async () => {
     const command = getRegistry().get('amazon/discussion');
     const page = createPageMock([

plugins/amazon/test/shared.test.js

@@ -6,6 +6,35 @@ describe('amazon shared helpers', () => {
         expect(__test__.buildProductUrl('https://www.amazon.com/dp/B0FJS72893/ref=something')).toBe('https://www.amazon.com/dp/B0FJS72893');
         expect(__test__.buildDiscussionUrl('https://www.amazon.com/dp/B0FJS72893')).toBe('https://www.amazon.com/product-reviews/B0FJS72893');
     });
+    it('keeps the input marketplace instead of rewriting it to the US store', () => {
+        expect(__test__.buildProductUrl('https://www.amazon.co.uk/dp/B0FGCPFY9L')).toBe('https://www.amazon.co.uk/dp/B0FGCPFY9L');
+        expect(__test__.buildProductUrl('https://www.amazon.de/dp/B0FJS72893/ref=something')).toBe('https://www.amazon.de/dp/B0FJS72893');
+        expect(__test__.buildProductUrl('https://www.amazon.com.au/dp/B0FJS72893')).toBe('https://www.amazon.com.au/dp/B0FJS72893');
+        expect(__test__.buildDiscussionUrl('https://www.amazon.co.uk/product-reviews/B0FGCPFY9L?pageNumber=1')).toBe('https://www.amazon.co.uk/product-reviews/B0FGCPFY9L');
+        expect(__test__.normalizeProductUrl('https://www.amazon.co.uk/dp/B0FGCPFY9L')).toBe('https://www.amazon.co.uk/dp/B0FGCPFY9L');
+    });
+    it('defaults to the US store for bare ASINs and non-marketplace hosts', () => {
+        expect(__test__.amazonHostFromInput('B0FJS72893')).toBeNull();
+        expect(__test__.buildProductUrl('B0FJS72893')).toBe('https://www.amazon.com/dp/B0FJS72893');
+        expect(__test__.buildDiscussionUrl('B0FJS72893')).toBe('https://www.amazon.com/product-reviews/B0FJS72893');
+        expect(__test__.normalizeProductUrl('B0FJS72893')).toBe('https://www.amazon.com/dp/B0FJS72893');
+    });
+    it('accepts sibling marketplaces but rejects look-alike hosts', () => {
+        expect(__test__.amazonHostFromInput('https://www.amazon.co.uk/dp/B0FJS72893')).toBe('www.amazon.co.uk');
+        expect(__test__.amazonHostFromInput('https://amazon.de/dp/B0FJS72893')).toBe('amazon.de');
+        expect(__test__.amazonHostFromInput('https://amazon.com.au/dp/B0FJS72893')).toBe('amazon.com.au');
+        expect(__test__.amazonHostFromInput('https://smile.amazon.com.be/dp/B0FJS72893')).toBe('smile.amazon.com.be');
+        expect(__test__.amazonHostFromInput('https://evilamazon.com/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://amazon.com.evil.com/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://amazon.evil.com/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://x.amazon.evil.com/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://amazon.attacker.io/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://amazon.shop/dp/B0FJS72893')).toBeNull();
+        expect(__test__.amazonHostFromInput('https://amazon.zip/dp/B0FJS72893')).toBeNull();
+        expect(() => __test__.canonicalizeAmazonUrl('https://amazon.evil.com/gp/bestsellers')).toThrow('Invalid Amazon URL');
+        expect(__test__.canonicalizeAmazonUrl('https://www.amazon.co.uk/gp/bestsellers/books')).toBe('https://www.amazon.co.uk/gp/bestsellers/books');
+        expect(() => __test__.canonicalizeAmazonUrl('https://evilamazon.com/gp/bestsellers')).toThrow('Invalid Amazon URL');
+    });
     it('parses price, rating, and review-count text', () => {
         expect(__test__.parsePriceText('1 offer from $34.11')).toEqual({
             price_text: '$34.11',

plugins/facebook/feed.js

@@ -306,12 +306,35 @@ async function loadFeedPosts(page, limit) {
         + Array.from(document.querySelectorAll('[aria-label]')).filter((el) => /^(Actions for this post)$/i.test((el.getAttribute('aria-label') || '').trim())).length
       : 0;
   })()`;
-  let prev = -1;
+  const extractStep = buildFeedExtractScript(limit);
+  let prevMarkerCount = -1;
+  let prevRowCount = -1;
+  let stalledPasses = 0;
   for (let i = 0; i < 8; i += 1) {
-    let count = 0;
-    try { count = Number(unwrapBrowserResult(await page.evaluate(scrollStep))) || 0; } catch { break; }
-    if (count >= limit || count === prev) break;
-    prev = count;
+    let markerCount = 0;
+    try { markerCount = Number(unwrapBrowserResult(await page.evaluate(scrollStep))) || 0; } catch { break; }
+
+    // Raw article/menu counts include comments, suggestions, and other chrome.
+    // Do not stop merely because those markers reached --limit; stop
+    // when the actual feed extractor has enough valid rows.
+    let rowCount = 0;
+    try {
+      const payload = unwrapBrowserResult(await page.evaluate(extractStep));
+      rowCount = Array.isArray(payload && payload.rows) ? payload.rows.length : 0;
+    } catch {
+      // The final extraction below owns error classification. A transient
+      // observation failure here should only make the bounded scroll continue.
+    }
+    if (rowCount >= limit) break;
+
+    if (markerCount === prevMarkerCount && rowCount === prevRowCount) {
+      stalledPasses += 1;
+      if (stalledPasses >= 2) break;
+    } else {
+      stalledPasses = 0;
+    }
+    prevMarkerCount = markerCount;
+    prevRowCount = rowCount;
   }
 }
 
@@ -387,5 +410,6 @@ export const __test__ = {
   buildFeedExtractScript,
   command,
   getFacebookFeed,
+  loadFeedPosts,
   requireLimit,
 };

plugins/facebook/profile.js

@@ -17,19 +17,48 @@ cli({
     pipeline: [
         { navigate: { url: 'https://www.facebook.com/${{ args.username }}', settleMs: 3000 } },
         { evaluate: `(() => {
-  const h1 = document.querySelector('h1');
-  let name = h1 ? h1.textContent.trim() : '';
+  const username = \${{ args.username | json }};
+  const main = document.querySelector('[role="main"]') || document;
+  const clean = (value) => String(value || '').replace(/\\s+/g, ' ').trim();
+  const h1 = main.querySelector('h1');
+  let name = clean(h1 && h1.textContent);
 
-  // Find friends/followers links
-  const links = Array.from(document.querySelectorAll('a'));
-  const friendsLink = links.find(a => a.href && a.href.includes('/friends'));
-  const followersLink = links.find(a => a.href && a.href.includes('/followers'));
+  // Facebook's current profile header no longer uses an h1. The profile-avatar
+  // link carries the display name on both the link and its role=img child.
+  // Scope this fallback to the requested profile path so page chrome cannot win.
+  if (!name) {
+    const requestedPrefix = '/' + String(username).toLowerCase() + '/';
+    const profileMediaLink = Array.from(main.querySelectorAll('a[aria-label][href]')).find((link) => {
+      const href = link.getAttribute('href') || '';
+      let path = '';
+      try { path = new URL(href, window.location.href).pathname.toLowerCase(); } catch {}
+      const image = link.querySelector('[role="img"][aria-label]');
+      return path.startsWith(requestedPrefix)
+        && image
+        && clean(link.getAttribute('aria-label')) === clean(image.getAttribute('aria-label'));
+    });
+    name = clean(profileMediaLink && profileMediaLink.getAttribute('aria-label'));
+  }
+
+  // Scope relationship links to the profile main region. Global navigation also
+  // contains /friends and previously produced empty or generic chrome text.
+  const links = Array.from(main.querySelectorAll('a[href]'));
+  const hasProfilePath = (link, suffix) => {
+    try {
+      const path = new URL(link.getAttribute('href') || '', window.location.href).pathname
+        .replace(/\\/+$/, '')
+        .toLowerCase();
+      return path === '/' + String(username).toLowerCase() + suffix;
+    } catch { return false; }
+  };
+  const friendsLink = links.find((a) => hasProfilePath(a, '/friends'));
+  const followersLink = links.find((a) => hasProfilePath(a, '/followers'));
 
   return [{
     name: name,
-    username: \${{ args.username | json }},
-    friends: friendsLink ? friendsLink.textContent.trim() : '-',
-    followers: followersLink ? followersLink.textContent.trim() : '-',
+    username,
+    friends: friendsLink ? clean(friendsLink.textContent) : '-',
+    followers: followersLink ? clean(followersLink.textContent) : '-',
     url: window.location.href,
   }];
 })()