aiseo-audit 1.4.6 → 1.4.7

package/README.md

@@ -6,7 +6,7 @@
 [![License: MIT](https://img.shields.io/badge/License-MIT-7EB6D7.svg)](https://opensource.org/licenses/MIT)
 [![Node.js](https://img.shields.io/badge/node-%3E%3D20-7EB6D7.svg)](https://nodejs.org)
 [![TypeScript](https://img.shields.io/badge/TypeScript-5.9-7EB6D7?logo=typescript&logoColor=white)](https://www.typescriptlang.org/)
-[![Tests](https://img.shields.io/badge/tests-454%20passed-8FBC8F)](https://github.com/agencyenterprise/aiseo-audit)
+[![Tests](https://img.shields.io/badge/tests-455%20passed-8FBC8F)](https://github.com/agencyenterprise/aiseo-audit)
 [![Coverage](https://img.shields.io/codecov/c/github/agencyenterprise/aiseo-audit?color=8FBC8F&label=coverage)](https://codecov.io/gh/agencyenterprise/aiseo-audit)
 [![GitHub Stars](https://img.shields.io/github/stars/agencyenterprise/aiseo-audit?style=flat&color=8FBC8F)](https://github.com/agencyenterprise/aiseo-audit/stargazers)
 

package/dist/cli.js

@@ -28,7 +28,7 @@ var import_zod3 = require("zod");
 
 // src/modules/analyzer/constants.ts
 var DOMAIN_SIGNAL_TIMEOUT_CAP = 5e3;
-var VERSION = true ? "1.4.6" : "0.0.0";
+var VERSION = true ? "1.4.7" : "0.0.0";
 
 // src/modules/fetcher/constants.ts
 var MAX_RESPONSE_SIZE = 10 * 1024 * 1024;
@@ -191,223 +191,10 @@ function getDomain(url) {
   }
 }
 
-// src/modules/audits/constants.ts
-var CATEGORY_DISPLAY_NAMES = {
-  contentExtractability: "Content Extractability",
-  contentStructure: "Content Structure for Reuse",
-  answerability: "Answerability",
-  entityClarity: "Entity Clarity",
-  groundingSignals: "Grounding Signals",
-  authorityContext: "Authority Context",
-  readabilityForCompression: "Readability for Compression"
-};
-
-// src/modules/audits/support/patterns.ts
-var DEFINITION_PATTERNS = [
-  /\bis\s+defined\s+as\b/gi,
-  /\brefers?\s+to\b/gi,
-  /\bmeans?\s+that\b/gi,
-  /\bis\s+a\s+type\s+of\b/gi,
-  /\bcan\s+be\s+described\s+as\b/gi,
-  /\balso\s+known\s+as\b/gi
-];
-var CITATION_PATTERNS = [
-  /\[\d+\]/g,
-  /\([\w\s]+,?\s*\d{4}\)/g,
-  /according\s+to/gi,
-  /research\s+(?:shows|indicates|suggests)/gi,
-  /studies?\s+(?:show|indicate|suggest|found)/gi,
-  /data\s+from/gi,
-  /as\s+reported\s+by/gi
-];
-var ATTRIBUTION_PATTERNS = [
-  /according\s+to/gi,
-  /\bsaid\b/gi,
-  /\bstated\b/gi,
-  /\breported\b/gi,
-  /\bcited\s+by\b/gi
-];
-var NUMERIC_CLAIM_PATTERNS = [
-  /\d+(?:\.\d+)?\s*%/g,
-  /\d+(?:\.\d+)?\s*(?:million|billion|thousand|trillion)/gi,
-  /\$[\d,.]+/g,
-  /increased\s+by/gi,
-  /decreased\s+by/gi,
-  /grew\s+by/gi
-];
-var STEP_PATTERNS = [
-  /step\s+\d+/gi,
-  /^\s*\d+\.\s+\w/gm,
-  /\bfirst(?:ly)?,?\s/gi,
-  /\bsecond(?:ly)?,?\s/gi,
-  /\bfinally,?\s/gi,
-  /\bhow\s+to\b/gi
-];
-var SUMMARY_MARKERS = [
-  /\bin\s+summary\b/gi,
-  /\bin\s+conclusion\b/gi,
-  /\bto\s+summarize\b/gi,
-  /\bkey\s+takeaways?\b/gi,
-  /\bbottom\s+line\b/gi,
-  /\btl;?dr\b/gi
-];
-var QUESTION_PATTERNS = [
-  /what\s+is/gi,
-  /what\s+are/gi,
-  /how\s+to/gi,
-  /how\s+do/gi,
-  /why\s+is/gi,
-  /why\s+do/gi,
-  /when\s+to/gi,
-  /where\s+to/gi,
-  /which\s+is/gi,
-  /who\s+is/gi
-];
-var DIRECT_ANSWER_PATTERNS = [
-  /^The\s+\w+\s+is\b/gm,
-  /^It\s+is\b/gm,
-  /^This\s+is\b/gm,
-  /^They\s+are\b/gm,
-  /\bsimply\s+put\b/gi,
-  /\bin\s+short\b/gi
-];
-var TRANSITION_WORDS = [
-  "however",
-  "therefore",
-  "moreover",
-  "furthermore",
-  "consequently",
-  "additionally",
-  "in contrast",
-  "similarly",
-  "as a result",
-  "for example",
-  "for instance",
-  "on the other hand",
-  "nevertheless",
-  "meanwhile",
-  "likewise",
-  "in addition",
-  "specifically",
-  "in particular",
-  "notably",
-  "importantly"
-];
-var AUTHOR_SELECTORS = [
-  '[rel="author"]',
-  ".author",
-  ".byline",
-  '[itemprop="author"]',
-  ".post-author",
-  ".entry-author",
-  'meta[name="author"]'
-];
-var DATE_SELECTORS = [
-  "time[datetime]",
-  '[itemprop="datePublished"]',
-  '[itemprop="dateModified"]',
-  ".published",
-  ".post-date",
-  ".entry-date",
-  'meta[property="article:published_time"]',
-  'meta[property="article:modified_time"]'
-];
-var QUESTION_HEADING_PATTERN = /^(?:what|how|why|when|where|which|who|can|do|does|is|are|should|will)\b/i;
-var QUOTED_ATTRIBUTION_PATTERNS = [
-  /"[^"]{10,}"\s*[-\u2013\u2014]\s*[A-Z][a-z]+/g,
-  /"[^"]{10,}",?\s+said\s+[A-Z]/g,
-  /"[^"]{10,}",?\s+according\s+to\s+[A-Z]/g,
-  /according\s+to\s+[A-Z][a-z]+[^,]*,\s*"[^"]{10,}"/g,
-  /\u201c[^\u201d]{10,}\u201d\s*[-\u2013\u2014]\s*[A-Z][a-z]+/g,
-  /\u201c[^\u201d]{10,}\u201d,?\s+said\s+[A-Z]/g
-];
-var AI_CRAWLERS = [
-  "GPTBot",
-  "ChatGPT-User",
-  "ClaudeBot",
-  "PerplexityBot",
-  "Google-Extended"
-];
-var MODIFIED_DATE_SELECTORS = [
-  '[itemprop="dateModified"]',
-  'meta[property="article:modified_time"]'
-];
-var PUBLISH_DATE_SELECTORS = [
-  "time[datetime]",
-  '[itemprop="datePublished"]',
-  'meta[property="article:published_time"]'
-];
-
-// src/modules/audits/support/dom.ts
-function detectAnswerCapsules($) {
-  let total = 0;
-  let withCapsule = 0;
-  $("h2").each((_, el) => {
-    const headingText = $(el).text().trim();
-    const isQuestion = headingText.includes("?") || QUESTION_HEADING_PATTERN.test(headingText);
-    if (!isQuestion) return;
-    total++;
-    const nextP = $(el).nextAll("p").first();
-    if (!nextP.length) return;
-    const pText = nextP.text().trim();
-    const firstSentence = pText.split(/[.!?]/)[0] || "";
-    if (firstSentence.length > 0 && firstSentence.length <= 200) {
-      withCapsule++;
-    }
-  });
-  return { total, withCapsule };
-}
-function measureSectionLengths($) {
-  const headings = $("h1, h2, h3, h4, h5, h6");
-  if (headings.length === 0)
-    return { sectionCount: 0, avgWordsPerSection: 0, sections: [] };
-  const sections = [];
-  headings.each((_, el) => {
-    let words = 0;
-    let sibling = $(el).next();
-    while (sibling.length && !sibling.is("h1, h2, h3, h4, h5, h6")) {
-      const text = sibling.text().trim();
-      words += text.split(/\s+/).filter((w) => w.length > 0).length;
-      sibling = sibling.next();
-    }
-    if (words > 0) sections.push(words);
-  });
-  const avg = sections.length > 0 ? Math.round(sections.reduce((a, b) => a + b, 0) / sections.length) : 0;
-  return { sectionCount: sections.length, avgWordsPerSection: avg, sections };
-}
-function parseJsonLdObjects($) {
-  const objects = [];
-  $('script[type="application/ld+json"]').each((_, el) => {
-    try {
-      const data = JSON.parse($(el).html() || "{}");
-      if (Array.isArray(data)) objects.push(...data);
-      else objects.push(data);
-    } catch {
-    }
-  });
-  return objects;
-}
-
-// src/modules/audits/support/nlp.ts
+// src/modules/nlp/service.ts
 var import_compromise = __toESM(require("compromise"));
 
-// src/utils/strings.ts
-function countWords(text) {
-  return text.split(/\s+/).filter((w) => w.length > 0).length;
-}
-function countSentences(text) {
-  return text.split(/[.!?]+/).filter((s) => s.trim().length > 5).length;
-}
-function countSyllables(word) {
-  word = word.toLowerCase().replace(/[^a-z]/g, "");
-  if (word.length <= 3) return 1;
-  word = word.replace(/(?:[^laeiouy]es|ed|[^laeiouy]e)$/, "");
-  word = word.replace(/^y/, "");
-  const matches = word.match(/[aeiouy]{1,2}/g);
-  return matches ? Math.max(matches.length, 1) : 1;
-}
-
-// src/modules/audits/support/nlp.ts
+// src/modules/nlp/constants.ts
 var STOPWORDS = /* @__PURE__ */ new Set([
   "a",
   "an",
@@ -680,6 +467,8 @@ var ACRONYM_STOPLIST = /* @__PURE__ */ new Set([
 ]);
 var ORG_SUFFIXES = /\b(?:Inc|Corp|Corporation|LLC|Ltd|Limited|Co|Company|Group|Foundation|Institute|University|Association|Society|Agency|Authority|Bureau|Commission|Council|Department|Board|Trust|Fund|Partners|Ventures|Labs|Technologies|Solutions|Systems|Services|Consulting|Media|Network|Studios|Entertainment|Healthcare|Pharmaceuticals|Dynamics|Holdings|Capital|Enterprises|International)\b/i;
 var PERSON_HONORIFICS = /\b(?:Mr|Mrs|Ms|Miss|Dr|Prof|Professor|Rev|Reverend|Sen|Senator|Rep|Representative|Gov|Governor|Pres|President|Gen|General|Col|Colonel|Sgt|Sergeant|Cpl|Corporal|Pvt|Private|Adm|Admiral|Capt|Captain|Lt|Lieutenant|Maj|Major|Sir|Dame|Lord|Lady|Hon|Honorable|Judge|Justice|Chancellor|Dean|Provost)\.\s*/;
+
+// src/modules/nlp/support/entities.ts
 function extractAcronymEntities(text) {
   const matches = text.match(/\b[A-Z]{2,6}\b/g);
   if (!matches) return [];
@@ -731,6 +520,33 @@ function isPersonByHonorific(text, entity) {
   );
   return pattern.test(text);
 }
+function smartDedup(entities) {
+  if (entities.length === 0) return [];
+  const sorted = [...entities].sort((a, b) => b.length - a.length);
+  const result = [];
+  const lowerSeen = /* @__PURE__ */ new Set();
+  for (const entity of sorted) {
+    const lower = entity.toLowerCase();
+    if (lowerSeen.has(lower)) continue;
+    let isSubstring = false;
+    for (const accepted of lowerSeen) {
+      if (accepted.includes(lower)) {
+        isSubstring = true;
+        break;
+      }
+    }
+    if (isSubstring) continue;
+    result.push(entity);
+    lowerSeen.add(lower);
+  }
+  return result;
+}
+function mergeEntityLists(compromiseList, supplementalList, limit) {
+  const combined = [...compromiseList, ...supplementalList];
+  return smartDedup(combined).slice(0, limit);
+}
+
+// src/modules/nlp/support/topics.ts
 function extractTopicsByTfIdf(text, limit) {
   const lower = text.toLowerCase();
   const words = lower.replace(/[^a-z0-9\s'-]/g, " ").split(/\s+/).filter((w) => w.length > 2 && !STOPWORDS.has(w));
@@ -754,31 +570,60 @@ function extractTopicsByTfIdf(text, limit) {
   candidates.sort((a, b) => b[1] - a[1]);
   return candidates.slice(0, limit).map(([term]) => term);
 }
-function smartDedup(entities) {
-  if (entities.length === 0) return [];
-  const sorted = [...entities].sort((a, b) => b.length - a.length);
-  const result = [];
-  const lowerSeen = /* @__PURE__ */ new Set();
-  for (const entity of sorted) {
-    const lower = entity.toLowerCase();
-    if (lowerSeen.has(lower)) continue;
-    let isSubstring = false;
-    for (const accepted of lowerSeen) {
-      if (accepted.includes(lower)) {
-        isSubstring = true;
-        break;
-      }
-    }
-    if (isSubstring) continue;
-    result.push(entity);
-    lowerSeen.add(lower);
+
+// src/modules/nlp/support/patterns.ts
+function countPatternMatches(text, patterns) {
+  let count = 0;
+  for (const pattern of patterns) {
+    const re = new RegExp(pattern.source, pattern.flags);
+    const matches = text.match(re);
+    if (matches) count += matches.length;
   }
-  return result;
+  return count;
+}
+function countTransitionWords(text, words) {
+  const lower = text.toLowerCase();
+  return words.filter((w) => lower.includes(w)).length;
+}
+
+// src/utils/strings.ts
+function countWords(text) {
+  return text.split(/\s+/).filter((w) => w.length > 0).length;
+}
+function countSentences(text) {
+  return text.split(/[.!?]+/).filter((s) => s.trim().length > 5).length;
+}
+function countSyllables(word) {
+  word = word.toLowerCase().replace(/[^a-z]/g, "");
+  if (word.length <= 3) return 1;
+  word = word.replace(/(?:[^laeiouy]es|ed|[^laeiouy]e)$/, "");
+  word = word.replace(/^y/, "");
+  const matches = word.match(/[aeiouy]{1,2}/g);
+  return matches ? Math.max(matches.length, 1) : 1;
 }
-function mergeEntityLists(compromiseList, supplementalList, limit) {
-  const combined = [...compromiseList, ...supplementalList];
-  return smartDedup(combined).slice(0, limit);
+
+// src/modules/nlp/support/readability.ts
+function computeFleschReadingEase(text) {
+  const words = text.split(/\s+/).filter((w) => w.length > 0);
+  const sentences = text.split(/[.!?]+/).filter((s) => s.trim().length > 5);
+  const totalSyllables = words.reduce((sum, w) => sum + countSyllables(w), 0);
+  if (words.length === 0 || sentences.length === 0) return 0;
+  const avgSentenceLength2 = words.length / sentences.length;
+  const avgSyllablesPerWord = totalSyllables / words.length;
+  return 206.835 - 1.015 * avgSentenceLength2 - 84.6 * avgSyllablesPerWord;
+}
+function countComplexWords(text) {
+  const words = text.split(/\s+/).filter((w) => w.length > 0);
+  return words.filter((w) => countSyllables(w) >= 4).length;
 }
+function avgSentenceLength(text) {
+  const words = text.split(/\s+/).filter((w) => w.length > 0);
+  const sentences = text.split(/[.!?]+/).filter((s) => s.trim().length > 5);
+  if (sentences.length === 0) return 0;
+  return Math.round(words.length / sentences.length);
+}
+
+// src/modules/nlp/service.ts
 function extractEntities(text) {
   const doc = (0, import_compromise.default)(text);
   const compromisePeople = [...new Set(doc.people().out("array"))];
@@ -820,37 +665,202 @@ function extractEntities(text) {
     numberCount
   };
 }
-function computeFleschReadingEase(text) {
-  const words = text.split(/\s+/).filter((w) => w.length > 0);
-  const sentences = text.split(/[.!?]+/).filter((s) => s.trim().length > 5);
-  const totalSyllables = words.reduce((sum, w) => sum + countSyllables(w), 0);
-  if (words.length === 0 || sentences.length === 0) return 0;
-  const avgSentenceLength2 = words.length / sentences.length;
-  const avgSyllablesPerWord = totalSyllables / words.length;
-  return 206.835 - 1.015 * avgSentenceLength2 - 84.6 * avgSyllablesPerWord;
-}
-function countComplexWords(text) {
-  const words = text.split(/\s+/).filter((w) => w.length > 0);
-  return words.filter((w) => countSyllables(w) >= 4).length;
-}
-function countPatternMatches(text, patterns) {
-  let count = 0;
-  for (const pattern of patterns) {
-    const re = new RegExp(pattern.source, pattern.flags);
-    const matches = text.match(re);
-    if (matches) count += matches.length;
-  }
-  return count;
+
+// src/modules/audits/constants.ts
+var CATEGORY_DISPLAY_NAMES = {
+  contentExtractability: "Content Extractability",
+  contentStructure: "Content Structure for Reuse",
+  answerability: "Answerability",
+  entityClarity: "Entity Clarity",
+  groundingSignals: "Grounding Signals",
+  authorityContext: "Authority Context",
+  readabilityForCompression: "Readability for Compression"
+};
+
+// src/modules/audits/support/patterns.ts
+var DEFINITION_PATTERNS = [
+  /\bis\s+defined\s+as\b/gi,
+  /\brefers?\s+to\b/gi,
+  /\bmeans?\s+that\b/gi,
+  /\bis\s+a\s+type\s+of\b/gi,
+  /\bcan\s+be\s+described\s+as\b/gi,
+  /\balso\s+known\s+as\b/gi
+];
+var CITATION_PATTERNS = [
+  /\[\d+\]/g,
+  /\([\w\s]+,?\s*\d{4}\)/g,
+  /according\s+to/gi,
+  /research\s+(?:shows|indicates|suggests)/gi,
+  /studies?\s+(?:show|indicate|suggest|found)/gi,
+  /data\s+from/gi,
+  /as\s+reported\s+by/gi
+];
+var ATTRIBUTION_PATTERNS = [
+  /according\s+to/gi,
+  /\bsaid\b/gi,
+  /\bstated\b/gi,
+  /\breported\b/gi,
+  /\bcited\s+by\b/gi
+];
+var NUMERIC_CLAIM_PATTERNS = [
+  /\d+(?:\.\d+)?\s*%/g,
+  /\d+(?:\.\d+)?\s*(?:million|billion|thousand|trillion)/gi,
+  /\$[\d,.]+/g,
+  /increased\s+by/gi,
+  /decreased\s+by/gi,
+  /grew\s+by/gi
+];
+var STEP_PATTERNS = [
+  /step\s+\d+/gi,
+  /^\s*\d+\.\s+\w/gm,
+  /\bfirst(?:ly)?,?\s/gi,
+  /\bsecond(?:ly)?,?\s/gi,
+  /\bfinally,?\s/gi,
+  /\bhow\s+to\b/gi
+];
+var SUMMARY_MARKERS = [
+  /\bin\s+summary\b/gi,
+  /\bin\s+conclusion\b/gi,
+  /\bto\s+summarize\b/gi,
+  /\bkey\s+takeaways?\b/gi,
+  /\bbottom\s+line\b/gi,
+  /\btl;?dr\b/gi
+];
+var QUESTION_PATTERNS = [
+  /what\s+is/gi,
+  /what\s+are/gi,
+  /how\s+to/gi,
+  /how\s+do/gi,
+  /why\s+is/gi,
+  /why\s+do/gi,
+  /when\s+to/gi,
+  /where\s+to/gi,
+  /which\s+is/gi,
+  /who\s+is/gi
+];
+var DIRECT_ANSWER_PATTERNS = [
+  /^The\s+\w+\s+is\b/gm,
+  /^It\s+is\b/gm,
+  /^This\s+is\b/gm,
+  /^They\s+are\b/gm,
+  /\bsimply\s+put\b/gi,
+  /\bin\s+short\b/gi
+];
+var TRANSITION_WORDS = [
+  "however",
+  "therefore",
+  "moreover",
+  "furthermore",
+  "consequently",
+  "additionally",
+  "in contrast",
+  "similarly",
+  "as a result",
+  "for example",
+  "for instance",
+  "on the other hand",
+  "nevertheless",
+  "meanwhile",
+  "likewise",
+  "in addition",
+  "specifically",
+  "in particular",
+  "notably",
+  "importantly"
+];
+var AUTHOR_SELECTORS = [
+  '[rel="author"]',
+  ".author",
+  ".byline",
+  '[itemprop="author"]',
+  ".post-author",
+  ".entry-author",
+  'meta[name="author"]'
+];
+var DATE_SELECTORS = [
+  "time[datetime]",
+  '[itemprop="datePublished"]',
+  '[itemprop="dateModified"]',
+  ".published",
+  ".post-date",
+  ".entry-date",
+  'meta[property="article:published_time"]',
+  'meta[property="article:modified_time"]'
+];
+var QUESTION_HEADING_PATTERN = /^(?:what|how|why|when|where|which|who|can|do|does|is|are|should|will)\b/i;
+var QUOTED_ATTRIBUTION_PATTERNS = [
+  /"[^"]{10,}"\s*[-\u2013\u2014]\s*[A-Z][a-z]+/g,
+  /"[^"]{10,}",?\s+said\s+[A-Z]/g,
+  /"[^"]{10,}",?\s+according\s+to\s+[A-Z]/g,
+  /according\s+to\s+[A-Z][a-z]+[^,]*,\s*"[^"]{10,}"/g,
+  /\u201c[^\u201d]{10,}\u201d\s*[-\u2013\u2014]\s*[A-Z][a-z]+/g,
+  /\u201c[^\u201d]{10,}\u201d,?\s+said\s+[A-Z]/g
+];
+var AI_CRAWLERS = [
+  "GPTBot",
+  "ChatGPT-User",
+  "ClaudeBot",
+  "PerplexityBot",
+  "Google-Extended"
+];
+var MODIFIED_DATE_SELECTORS = [
+  '[itemprop="dateModified"]',
+  'meta[property="article:modified_time"]'
+];
+var PUBLISH_DATE_SELECTORS = [
+  "time[datetime]",
+  '[itemprop="datePublished"]',
+  'meta[property="article:published_time"]'
+];
+
+// src/modules/audits/support/dom.ts
+function detectAnswerCapsules($) {
+  let total = 0;
+  let withCapsule = 0;
+  $("h2").each((_, el) => {
+    const headingText = $(el).text().trim();
+    const isQuestion = headingText.includes("?") || QUESTION_HEADING_PATTERN.test(headingText);
+    if (!isQuestion) return;
+    total++;
+    const nextP = $(el).nextAll("p").first();
+    if (!nextP.length) return;
+    const pText = nextP.text().trim();
+    const firstSentence = pText.split(/[.!?]/)[0] || "";
+    if (firstSentence.length > 0 && firstSentence.length <= 200) {
+      withCapsule++;
+    }
+  });
+  return { total, withCapsule };
 }
-function countTransitionWords(text, words) {
-  const lower = text.toLowerCase();
-  return words.filter((w) => lower.includes(w)).length;
+function measureSectionLengths($) {
+  const headings = $("h1, h2, h3, h4, h5, h6");
+  if (headings.length === 0)
+    return { sectionCount: 0, avgWordsPerSection: 0, sections: [] };
+  const sections = [];
+  headings.each((_, el) => {
+    let words = 0;
+    let sibling = $(el).next();
+    while (sibling.length && !sibling.is("h1, h2, h3, h4, h5, h6")) {
+      const text = sibling.text().trim();
+      words += text.split(/\s+/).filter((w) => w.length > 0).length;
+      sibling = sibling.next();
+    }
+    if (words > 0) sections.push(words);
+  });
+  const avg = sections.length > 0 ? Math.round(sections.reduce((a, b) => a + b, 0) / sections.length) : 0;
+  return { sectionCount: sections.length, avgWordsPerSection: avg, sections };
 }
-function avgSentenceLength(text) {
-  const words = text.split(/\s+/).filter((w) => w.length > 0);
-  const sentences = text.split(/[.!?]+/).filter((s) => s.trim().length > 5);
-  if (sentences.length === 0) return 0;
-  return Math.round(words.length / sentences.length);
+function parseJsonLdObjects($) {
+  const objects = [];
+  $('script[type="application/ld+json"]').each((_, el) => {
+    try {
+      const data = JSON.parse($(el).html() || "{}");
+      if (Array.isArray(data)) objects.push(...data);
+      else objects.push(data);
+    } catch {
+    }
+  });
+  return objects;
 }
 
 // src/modules/audits/support/scoring.ts
@@ -4104,13 +4114,13 @@ function renderSitemapMarkdown(result) {
 var import_chalk = __toESM(require("chalk"));
 function scoreColor(score, max) {
   const pct = max > 0 ? score / max * 100 : 0;
-  if (pct >= 70) return import_chalk.default.green;
-  if (pct >= 40) return import_chalk.default.yellow;
+  if (pct >= 90) return import_chalk.default.green;
+  if (pct >= 50) return import_chalk.default.yellow;
   return import_chalk.default.red;
 }
 function gradeColor(grade) {
   if (grade.startsWith("A")) return import_chalk.default.green;
-  if (grade.startsWith("B")) return import_chalk.default.yellow;
+  if (grade.startsWith("B") || grade.startsWith("C")) return import_chalk.default.yellow;
   return import_chalk.default.red;
 }
 function pad(str, len) {
@@ -4257,7 +4267,7 @@ function renderSitemapPretty(result) {
     lines.push(import_chalk.default.bold("  Site-Wide Category Averages:"));
     lines.push("");
     for (const avg of Object.values(result.categoryAverages)) {
-      const color = avg.averagePct >= 70 ? import_chalk.default.green : avg.averagePct >= 40 ? import_chalk.default.yellow : import_chalk.default.red;
+      const color = scoreColor(avg.averagePct, 100);
       const name = pad(avg.name, 38);
       const dts = dots(Math.max(2, 40 - avg.name.length));
       lines.push(`  ${import_chalk.default.bold(name)} ${dts} ${color(`${avg.averagePct}%`)}`);
@@ -4337,6 +4347,36 @@ function renderSitemapReport(result, options) {
 }
 
 // src/modules/sitemap/service.ts
+var import_xml_to_html_converter = require("xml-to-html-converter");
+function stripCdata(raw) {
+  const trimmed = raw.trim();
+  if (trimmed.startsWith("<![CDATA[") && trimmed.endsWith("]]>")) {
+    return trimmed.slice(9, -3);
+  }
+  return trimmed;
+}
+function collectLocText(nodes, urls) {
+  for (const node of nodes) {
+    if (node.xmlTag === "loc" && node.children) {
+      const text = node.children.filter((c) => c.role === "textLeaf").map((c) => stripCdata(c.raw)).join("").trim();
+      if (text) urls.push(text);
+    }
+    if (node.children) collectLocText(node.children, urls);
+  }
+}
+function extractLocUrls(xml) {
+  const nodes = (0, import_xml_to_html_converter.scaffold)(xml);
+  const urls = [];
+  collectLocText(nodes, urls);
+  return urls;
+}
+function hasSitemapIndexNode(nodes) {
+  for (const node of nodes) {
+    if (node.xmlTag === "sitemapindex") return true;
+    if (node.children && hasSitemapIndexNode(node.children)) return true;
+  }
+  return false;
+}
 async function fetchSitemapUrls(sitemapUrl, timeout, userAgent) {
   const response = await httpGet({
     url: sitemapUrl,
@@ -4346,11 +4386,13 @@ async function fetchSitemapUrls(sitemapUrl, timeout, userAgent) {
   if (response.status !== 200) {
     throw new Error(`Failed to fetch sitemap: HTTP ${response.status}`);
   }
-  const isSitemapIndex = response.data.includes("<sitemapindex");
-  if (isSitemapIndex) {
+  const nodes = (0, import_xml_to_html_converter.scaffold)(response.data);
+  if (hasSitemapIndexNode(nodes)) {
     return fetchSitemapIndexUrls(response.data, timeout, userAgent);
   }
-  return extractLocUrls(response.data);
+  const urls = [];
+  collectLocText(nodes, urls);
+  return urls;
 }
 async function fetchSitemapIndexUrls(xml, timeout, userAgent) {
   const childSitemapUrls = extractLocUrls(xml);
@@ -4367,10 +4409,6 @@ async function fetchSitemapIndexUrls(xml, timeout, userAgent) {
   }
   return allUrls;
 }
-function extractLocUrls(xml) {
-  const matches = xml.matchAll(/<loc>\s*(.*?)\s*<\/loc>/g);
-  return Array.from(matches, (m) => m[1]).filter(Boolean);
-}
 function computeCategoryAverages(urlResults) {
   const successResults = urlResults.filter((r) => r.status === "success").map(
     (r) => r.result