detector: add 5 new antipatterns (em-dash-overuse, broken-image, marketing-buzzword, numbered-section-markers, aphoristic-cadence)

Consolidates eval-side detection logic into the canonical impeccable
detector. Before this change, the eval harness had its own duplicate
implementations of em-dash, copy-slop, and broken-image checks. They
now live alongside the existing 28 antipatterns in the impeccable
registry, available to the CLI, browser extension, critique skill,
and eval (via the existing slop grader child-process call).

New antipatterns:
- em-dash-overuse: 5+ em-dashes in body text content (threshold
  permits legitimate prose use of em-dash; only triggers on AI
  cadence-level density)
- broken-image: <img> with empty src, missing src, or src="#"
- marketing-buzzword: SaaS phrase list (streamline / empower /
  supercharge / enterprise-grade / cutting-edge / etc)
- numbered-section-markers: repeated 01 / 02 / 03 sequence as
  section labels — the AI editorial scaffold one tier deeper than
  tracked eyebrow chips
- aphoristic-cadence: 3+ manufactured-contrast ("Not a X. A Y.")
  or short-rebuttal ("Sentence. No clause." / "Sentence. Just
  clause.") constructions in body text

Engine wiring:
- broken-image runs as a static-html element rule (selector: img)
  and a fallback regex matcher (for non-HTML files)
- em-dash / buzzword / numbered / aphoristic run as regex
  page-analyzers, factored into a new runTextContentAnalyzers()
  helper that both detectText (non-HTML) and detectHtml (HTML)
  call, so .html files get the same coverage as .css/.tsx

Tests: 166 detector + 12 browser + 24 fixture all pass.
Browser detector rebuilt (162.7 KB).

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

cli/engine/detect-antipatterns-browser.js

@@ -247,6 +247,52 @@ const ANTIPATTERNS = [
     skillSection: 'Typography',
     skillGuideline: 'repeated eyebrow or kicker labels as section scaffolding',
   },
+  {
+    id: 'numbered-section-markers',
+    category: 'slop',
+    severity: 'advisory',
+    name: 'Numbered section markers (01 / 02 / 03)',
+    description:
+      'Numbered display markers as section labels (01, 02, 03) are the AI editorial scaffold one tier deeper than tracked eyebrow chips. If you find yourself reaching for them, choose a different section cadence.',
+    skillSection: 'Layout & Space',
+    skillGuideline: 'numbered section markers',
+  },
+  {
+    id: 'em-dash-overuse',
+    category: 'slop',
+    name: 'Em-dash overuse',
+    description:
+      'More than two em-dashes (— or --) in body copy is an AI cadence tell. Use commas, colons, periods, or parentheses instead.',
+    skillSection: 'Copy',
+    skillGuideline: 'no em dashes',
+  },
+  {
+    id: 'marketing-buzzword',
+    category: 'slop',
+    name: 'Marketing buzzword',
+    description:
+      'Generic SaaS phrases (streamline / empower / supercharge / world-class / enterprise-grade / next-generation / cutting-edge / etc) are instant AI tells. Pick a specific verb and noun that says what the product literally does.',
+    skillSection: 'Copy',
+    skillGuideline: 'marketing buzzwords',
+  },
+  {
+    id: 'aphoristic-cadence',
+    category: 'slop',
+    name: 'Aphoristic-cadence copy',
+    description:
+      'Three or more sections landing on a short rebuttal sentence ("X. No Y." / "X. Just Y.") or a manufactured-contrast aphorism ("Not a feature. A platform.") reads as AI cadence, not voice. Once is fine; the pattern is the tell.',
+    skillSection: 'Copy',
+    skillGuideline: 'aphoristic cadence',
+  },
+  {
+    id: 'broken-image',
+    category: 'quality',
+    name: 'Broken or placeholder image',
+    description:
+      '<img> tags with empty src, missing src, or placeholder values ship as broken-image boxes. Use real images, generated assets, or remove the tag.',
+    skillSection: 'Imagery',
+    skillGuideline: 'broken image references',
+  },
 
   // ── Quality: general design and accessibility issues ──
   {

cli/engine/engines/regex/detect-text.mjs

@@ -11,6 +11,17 @@ const hasRounded = (line) => /\brounded(?:-\w+)?\b/.test(line);
 const hasBorderRadius = (line) => /border-radius/i.test(line);
 const isSafeElement = (line) => /<(?:blockquote|nav[\s>]|pre[\s>]|code[\s>]|a\s|input[\s>]|span[\s>])/i.test(line);
 
+/** Strip HTML to plain text — drops script/style/comments/tags so
+ *  content-text analyzers don't false-positive on code or CSS. */
+function stripHtmlToText(html) {
+  return html
+    .replace(/<script\b[^>]*>[\s\S]*?<\/script>/gi, ' ')
+    .replace(/<style\b[^>]*>[\s\S]*?<\/style>/gi, ' ')
+    .replace(/<!--[\s\S]*?-->/g, ' ')
+    .replace(/<[^>]+>/g, ' ')
+    .replace(/\s+/g, ' ');
+}
+
 function isNeutralBorderColor(str) {
   const m = str.match(/solid\s+(#[0-9a-f]{3,8}|rgba?\([^)]+\)|\w+)/i);
   if (!m) return false;
@@ -116,6 +127,14 @@ const REGEX_MATCHERS = [
       const found = m[1].match(/\b(?:(?:max|min)-)?(?:width|height)\b|\bpadding(?:-(?:top|right|bottom|left))?\b|\bmargin(?:-(?:top|right|bottom|left))?\b/gi);
       return `transition-property: ${found ? found.join(', ') : m[1].trim()}`;
     } },
+  // --- Broken image: src="" or src="#" or src=" " ---
+  { id: 'broken-image', regex: /<img\b[^>]*?\bsrc\s*=\s*(?:""|''|"\s+"|'\s+'|"#"|'#')/gi,
+    test: () => true,
+    fmt: (m) => m[0].slice(0, 100) },
+  // --- Broken image: <img> with no src attribute at all ---
+  { id: 'broken-image', regex: /<img\b(?:(?!\bsrc\s*=)[^>])*>/gi,
+    test: (m) => !/\bsrc\s*=/i.test(m[0]),
+    fmt: (m) => m[0].slice(0, 100) },
 ];
 
 const REGEX_ANALYZERS = [
@@ -202,6 +221,87 @@ const REGEX_ANALYZERS = [
     if (total < 5 || centered / total <= 0.7) return [];
     return [finding('everything-centered', filePath, `${centered}/${total} text elements centered (${Math.round(centered / total * 100)}%)`)];
   },
+  // Em-dash overuse: 5+ em-dashes or "--" in body text content
+  // (occasional em-dash use in prose is fine; the pattern fires only
+  // when count crosses into AI-cadence territory).
+  (content, filePath) => {
+    const text = stripHtmlToText(content);
+    let count = 0;
+    const re = /[—]|--(?=\S)/g;
+    while (re.exec(text) !== null) count++;
+    if (count < 5) return [];
+    return [finding('em-dash-overuse', filePath, `${count} em-dashes in body text`)];
+  },
+  // Marketing buzzwords: SaaS phrase list
+  (content, filePath) => {
+    const text = stripHtmlToText(content);
+    const lower = text.toLowerCase();
+    const BUZZWORDS = [
+      'streamline your', 'empower your', 'supercharge your',
+      'unleash your', 'unleash the power', 'leverage the power',
+      'built for the modern', 'trusted by leading', 'trusted by the world',
+      'best-in-class', 'industry-leading', 'world-class', 'enterprise-grade',
+      'next-generation', 'cutting-edge', 'transform your business',
+      'revolutionize', 'game-changer', 'game changing',
+      'mission-critical', 'best of breed', 'future-proof', 'future proof',
+      'seamless experience', 'seamlessly integrate',
+      'drive engagement', 'drive growth', 'drive results',
+      'harness the power',
+    ];
+    let count = 0;
+    let firstSample = '';
+    for (const phrase of BUZZWORDS) {
+      let from = 0;
+      while (true) {
+        const idx = lower.indexOf(phrase, from);
+        if (idx === -1) break;
+        count++;
+        if (!firstSample) {
+          firstSample = text.slice(Math.max(0, idx - 12), Math.min(text.length, idx + phrase.length + 12)).trim();
+        }
+        from = idx + phrase.length;
+      }
+    }
+    if (count === 0) return [];
+    return [finding('marketing-buzzword', filePath, `${count} buzzword phrase${count === 1 ? '' : 's'}: "${firstSample}"`)];
+  },
+  // Numbered section markers (01 / 02 / 03 ...)
+  (content, filePath) => {
+    const text = stripHtmlToText(content);
+    const re = /\b(0[1-9]|1[0-2])\b/g;
+    const seen = new Set();
+    let m;
+    while ((m = re.exec(text)) !== null) seen.add(m[1]);
+    if (seen.size < 3) return [];
+    const sorted = [...seen].sort();
+    let sequential = 0;
+    for (let i = 1; i < sorted.length; i++) {
+      if (parseInt(sorted[i], 10) === parseInt(sorted[i - 1], 10) + 1) sequential++;
+    }
+    if (sequential < 2) return [];
+    return [finding('numbered-section-markers', filePath, `Sequence: ${sorted.slice(0, 6).join(', ')}`)];
+  },
+  // Aphoristic cadence: manufactured-contrast + short-rebuttal
+  (content, filePath) => {
+    const text = stripHtmlToText(content);
+    const NOT_A_RE = /\bNot an? [a-z][^.!?]{1,40}[.!]\s+[A-Z][^.!?]{1,60}[.!]/g;
+    const SHORT_REBUTTAL_RE = /\b[A-Z][^.!?]{4,80}[.!]\s+(No|Just)\s+[a-z][^.!?]{2,60}[.!]/g;
+    let count = 0;
+    let firstSample = '';
+    let m;
+    NOT_A_RE.lastIndex = 0;
+    while ((m = NOT_A_RE.exec(text)) !== null) {
+      count++;
+      if (!firstSample) firstSample = m[0].trim().slice(0, 80);
+    }
+    SHORT_REBUTTAL_RE.lastIndex = 0;
+    while ((m = SHORT_REBUTTAL_RE.exec(text)) !== null) {
+      count++;
+      if (!firstSample) firstSample = m[0].trim().slice(0, 80);
+    }
+    if (count < 3) return [];
+    return [finding('aphoristic-cadence', filePath, `${count} aphoristic constructions: "${firstSample}"`)];
+  },
   // Dark glow (page-level: dark bg + colored box-shadow with blur)
   (content, filePath) => {
     // Check if page has a dark background
@@ -320,6 +420,36 @@ function runRegexMatchers(lines, filePath, lineOffset = 0, blockContext = null,
   return findings;
 }
 
+/** Page-level analyzers that scan rendered text content (em-dash use,
+ *  buzzword phrases, numbered section markers, aphoristic cadence).
+ *  These are detector-agnostic — they work on any HTML/text source
+ *  and don't need a parsed DOM. Exported so detectHtml can call them
+ *  for `.html` files (which otherwise skip the regex engine). */
+const TEXT_CONTENT_ANALYZER_IDS = [
+  'em-dash-overuse',
+  'marketing-buzzword',
+  'numbered-section-markers',
+  'aphoristic-cadence',
+];
+
+function runTextContentAnalyzers(content, filePath, options = {}) {
+  const profile = options?.profile;
+  if (!isFullPage(content)) return [];
+  // The 4 text-content analyzers are at indices 4-7 in REGEX_ANALYZERS.
+  const findings = [];
+  for (let i = 0; i < TEXT_CONTENT_ANALYZER_IDS.length; i++) {
+    const analyzer = REGEX_ANALYZERS[4 + i];
+    const ruleId = TEXT_CONTENT_ANALYZER_IDS[i];
+    findings.push(...profileFindings(profile, {
+      engine: 'regex',
+      phase: 'text-content',
+      ruleId,
+      target: filePath,
+    }, () => analyzer(content, filePath)));
+  }
+  return findings;
+}
+
 function detectText(content, filePath, options = {}) {
   const profile = options?.profile;
   const findings = [];
@@ -386,6 +516,10 @@ function detectText(content, filePath, options = {}) {
       'flat-type-hierarchy',
       'monotonous-spacing',
       'everything-centered',
+      'em-dash-overuse',
+      'marketing-buzzword',
+      'numbered-section-markers',
+      'aphoristic-cadence',
       'dark-glow',
     ];
     for (let i = 0; i < REGEX_ANALYZERS.length; i++) {
@@ -405,8 +539,10 @@ function detectText(content, filePath, options = {}) {
 export {
   REGEX_MATCHERS,
   REGEX_ANALYZERS,
+  TEXT_CONTENT_ANALYZER_IDS,
   extractStyleBlocks,
   extractCSSinJS,
   runRegexMatchers,
+  runTextContentAnalyzers,
   detectText,
 };

cli/engine/engines/static-html/detect-html.mjs

@@ -21,7 +21,7 @@ import {
   resolveBackground,
   resolveBorderRadiusPx,
 } from '../../rules/checks.mjs';
-import { detectText } from '../regex/detect-text.mjs';
+import { detectText, runTextContentAnalyzers } from '../regex/detect-text.mjs';
 import {
   StaticDocument,
   buildStaticStyleMap,
@@ -64,6 +64,20 @@ function checkStaticPageTypography(document, window) {
   return findings;
 }
 
+function checkElementBrokenImage(el) {
+  const src = (el.getAttribute && el.getAttribute('src')) ?? el.attribs?.src;
+  // Missing src attribute entirely
+  if (src === undefined || src === null) {
+    return [{ id: 'broken-image', snippet: '<img> with no src attribute' }];
+  }
+  const trimmed = String(src).trim();
+  // Empty or placeholder-only src values
+  if (trimmed === '' || trimmed === '#') {
+    return [{ id: 'broken-image', snippet: `<img src="${src}">` }];
+  }
+  return [];
+}
+
 const STATIC_ELEMENT_RULES = [
   { id: 'border-rules', selector: '*', run: (el, tag, style, window, customPropMap) => checkElementBorders(tag, style, null, resolveBorderRadiusPx(el, style, parseFloat(style.width) || 0, window)) },
   { id: 'color-rules', selector: '*', run: (el, tag, style, window, customPropMap) => checkElementColors(el, style, tag, window, customPropMap, false) },
@@ -72,6 +86,7 @@ const STATIC_ELEMENT_RULES = [
   { id: 'icon-tile-stack', selector: 'h1,h2,h3,h4,h5,h6', run: (el, tag, _style, window) => checkElementIconTile(el, tag, window) },
   { id: 'italic-serif-display', selector: 'h1,h2', run: (el, tag, style) => checkElementItalicSerif(el, style, tag) },
   { id: 'hero-eyebrow-chip', selector: 'h1', run: (el, tag, style, window, customPropMap) => checkElementHeroEyebrow(el, style, tag, window, customPropMap) },
+  { id: 'broken-image', selector: 'img', run: (el) => checkElementBrokenImage(el) },
   { id: 'quality-rules', selector: '*', run: (el, tag, style, window) => checkElementQuality(el, style, tag, window) },
 ];
 
@@ -166,6 +181,14 @@ async function detectHtml(filePath, options = {}) {
     ))) {
       findings.push(finding(f.id, filePath, f.snippet));
     }
+    // Text-content analyzers (em-dash overuse, marketing buzzwords,
+    // numbered section markers, aphoristic cadence) live in the regex
+    // engine. Call them from here so .html files get the same coverage
+    // as .css/.tsx files. These are scoped to text content only and
+    // don't overlap with static-html's element/page rules.
+    for (const f of runPageCheck('text-content', () => runTextContentAnalyzers(html, filePath, options))) {
+      findings.push(finding(f.antipattern, filePath, f.snippet));
+    }
   }
 
   return findings;

cli/engine/registry/antipatterns.mjs

@@ -145,6 +145,52 @@ const ANTIPATTERNS = [
     skillSection: 'Typography',
     skillGuideline: 'repeated eyebrow or kicker labels as section scaffolding',
   },
+  {
+    id: 'numbered-section-markers',
+    category: 'slop',
+    severity: 'advisory',
+    name: 'Numbered section markers (01 / 02 / 03)',
+    description:
+      'Numbered display markers as section labels (01, 02, 03) are the AI editorial scaffold one tier deeper than tracked eyebrow chips. If you find yourself reaching for them, choose a different section cadence.',
+    skillSection: 'Layout & Space',
+    skillGuideline: 'numbered section markers',
+  },
+  {
+    id: 'em-dash-overuse',
+    category: 'slop',
+    name: 'Em-dash overuse',
+    description:
+      'More than two em-dashes (— or --) in body copy is an AI cadence tell. Use commas, colons, periods, or parentheses instead.',
+    skillSection: 'Copy',
+    skillGuideline: 'no em dashes',
+  },
+  {
+    id: 'marketing-buzzword',
+    category: 'slop',
+    name: 'Marketing buzzword',
+    description:
+      'Generic SaaS phrases (streamline / empower / supercharge / world-class / enterprise-grade / next-generation / cutting-edge / etc) are instant AI tells. Pick a specific verb and noun that says what the product literally does.',
+    skillSection: 'Copy',
+    skillGuideline: 'marketing buzzwords',
+  },
+  {
+    id: 'aphoristic-cadence',
+    category: 'slop',
+    name: 'Aphoristic-cadence copy',
+    description:
+      'Three or more sections landing on a short rebuttal sentence ("X. No Y." / "X. Just Y.") or a manufactured-contrast aphorism ("Not a feature. A platform.") reads as AI cadence, not voice. Once is fine; the pattern is the tell.',
+    skillSection: 'Copy',
+    skillGuideline: 'aphoristic cadence',
+  },
+  {
+    id: 'broken-image',
+    category: 'quality',
+    name: 'Broken or placeholder image',
+    description:
+      '<img> tags with empty src, missing src, or placeholder values ship as broken-image boxes. Use real images, generated assets, or remove the tag.',
+    skillSection: 'Imagery',
+    skillGuideline: 'broken image references',
+  },
 
   // ── Quality: general design and accessibility issues ──
   {