From 7bfaf5aefeb530f7890ad279d7a14c9812fd191b Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Fri, 10 Apr 2026 09:09:35 +0000
Subject: [PATCH 1/5] Initial plan


From 803ae340572e0a068a5d12f1191c00c9bd61ed63 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Fri, 10 Apr 2026 09:33:47 +0000
Subject: [PATCH 2/5] refactor: split shared.ts into focused helpers,
 consolidate types, remove deprecated code
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Split scripts/data-transformers/content-generators/shared.ts (1125 lines → 527 lines)
  into focused sub-modules:
  - doc-type-helpers.ts: DOC_TYPE_DISPLAY, localizeDocType, TITLE_SUFFIX_TEMPLATES
  - event-helpers.ts: findRelatedDocuments, findRelatedQuestions, extractMinister
  - impact-helpers.ts: generateImpactAnalysis, generateConsequencesAnalysis
  - framework-renderers.ts: stub types/functions + PESTLE/stakeholder/risk/implementation
    HTML renderers
  - ai-marker-helpers.ts: detectBannedPatterns + BANNED_PATTERNS

- shared.ts kept as barrel that re-exports all previously-exported symbols for
  full backward compatibility; no existing imports were broken

- Extract URL/text utilities from generators.ts (2234 → 2109 lines) to
  url-utils.ts: extractDocIdFromUrl, isGovernmentUrl, isGitHubUrl,
  toGitHubRawUrl, hashPathSuffix, sanitizePlainText; re-exported for BC

- Types review: scripts/data-transformers/types.ts and
  scripts/generate-news-enhanced/types.ts have no duplicates with
  scripts/types/ — no consolidation needed

- All 4097 tests pass, ESLint clean (no new errors)

Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
---
 .../content-generators/ai-marker-helpers.ts   |  56 ++
 .../content-generators/doc-type-helpers.ts    | 156 +++
 .../content-generators/event-helpers.ts       |  62 ++
 .../content-generators/framework-renderers.ts | 353 +++++++
 .../content-generators/impact-helpers.ts      | 110 +++
 .../content-generators/shared.ts              | 888 +++---------------
 scripts/generate-news-enhanced/generators.ts  | 149 +--
 scripts/generate-news-enhanced/url-utils.ts   | 145 +++
 8 files changed, 1039 insertions(+), 880 deletions(-)
 create mode 100644 scripts/data-transformers/content-generators/ai-marker-helpers.ts
 create mode 100644 scripts/data-transformers/content-generators/doc-type-helpers.ts
 create mode 100644 scripts/data-transformers/content-generators/event-helpers.ts
 create mode 100644 scripts/data-transformers/content-generators/framework-renderers.ts
 create mode 100644 scripts/data-transformers/content-generators/impact-helpers.ts
 create mode 100644 scripts/generate-news-enhanced/url-utils.ts

diff --git a/scripts/data-transformers/content-generators/ai-marker-helpers.ts b/scripts/data-transformers/content-generators/ai-marker-helpers.ts
new file mode 100644
index 000000000..d3d941282
--- /dev/null
+++ b/scripts/data-transformers/content-generators/ai-marker-helpers.ts
@@ -0,0 +1,56 @@
+/**
+ * @module data-transformers/content-generators/ai-marker-helpers
+ * @description Banned content pattern detection.
+ * Per SHARED_PROMPT_PATTERNS.md §BANNED Content Patterns v4.0,
+ * these patterns must never appear in production articles.
+ * AI agents MUST replace all AI_MUST_REPLACE markers with genuine analysis.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+/**
+ * Banned content patterns that indicate low-quality boilerplate text.
+ * Per SHARED_PROMPT_PATTERNS.md §BANNED Content Patterns v4.0, these
+ * must never appear in production articles. AI agents MUST replace them
+ * with genuine, document-specific analysis.
+ */
+const BANNED_PATTERNS: readonly { label: string; pattern: RegExp }[] = [
+  { label: 'neutralText: "The political landscape remains fluid…"', pattern: /The political landscape remains fluid,? with both government and opposition positioning for advantage/i },
+  { label: 'debateAnalysisMarker: "No chamber debate data is available…"', pattern: /No chamber debate data is available for these items,? limiting our ability/i },
+  { label: 'policySignificanceTouches: "Touches on {domains}."', pattern: /Touches on [\p{L}\p{N}][\p{L}\p{N}\s,&/()-]*\./iu },
+  { label: 'analysisOfNDocuments: "Analysis of N documents covering…"', pattern: /Analysis of \d+ documents covering/i },
+  { label: 'policySignificanceGeneric: "Requires committee review and chamber debate…"', pattern: /Requires committee review and chamber debate/i },
+  { label: 'topicInFocusSuffix: "…: {Topic} in Focus"', pattern: /:\s+\w[\w\s]*\bin Focus\b/i },
+  { label: 'briefingOnFieldLabels: "Political intelligence briefing on {Field}: and {Field}:"', pattern: /Political intelligence briefing on \w+:\s+and\s+\w+:/i },
+  // Deep Analysis generic template patterns — AI MUST replace these with specific analysis
+  { label: 'genericTimeline: "The pace of activity signals…"', pattern: /The pace of activity signals the political urgency/i },
+  { label: 'genericTimeline: "define the current legislative landscape"', pattern: /define the current legislative landscape/i },
+  { label: 'genericWhy: "broad legislative push that will shape"', pattern: /broad legislative push that will shape multiple aspects/i },
+  { label: 'genericWhy: "critical period for understanding the government"', pattern: /critical period for understanding the government.s strategic direction/i },
+  { label: 'genericImpact: "culmination of legislative review, with recommendations that guide"', pattern: /culmination of legislative review,? with recommendations that guide/i },
+  { label: 'genericImpact: "interplay between governing ambition and opposition scrutiny"', pattern: /interplay between governing ambition and opposition scrutiny/i },
+  { label: 'genericConsequences: "cascade through committee deliberations"', pattern: /cascade through committee deliberations,? chamber votes/i },
+  { label: 'genericConsequences: "establish the policy alternatives that opposition parties will champion"', pattern: /establish the policy alternatives that opposition parties will champion/i },
+  { label: 'genericCritical: "Standard parliamentary procedures are being followed"', pattern: /Standard parliamentary procedures are being followed/i },
+  { label: 'genericCritical: "gap between legislative intent and implementation"', pattern: /gap between legislative intent and implementation often reveals/i },
+  { label: 'genericPillarTransition: "While parliament deliberates these legislative matters"', pattern: /While parliament deliberates these legislative matters/i },
+];
+
+/**
+ * Detect banned boilerplate patterns in HTML content.
+ * Returns an array of human-readable labels identifying each detected
+ * banned pattern, suitable for quality gate logs and error messages.
+ *
+ * @param html - The HTML string to scan for banned patterns
+ * @returns Array of stable human-readable labels for each detected banned pattern
+ */
+export function detectBannedPatterns(html: string): string[] {
+  const found: string[] = [];
+  for (const { label, pattern } of BANNED_PATTERNS) {
+    if (pattern.test(html)) {
+      found.push(label);
+    }
+  }
+  return found;
+}
diff --git a/scripts/data-transformers/content-generators/doc-type-helpers.ts b/scripts/data-transformers/content-generators/doc-type-helpers.ts
new file mode 100644
index 000000000..056f07fd2
--- /dev/null
+++ b/scripts/data-transformers/content-generators/doc-type-helpers.ts
@@ -0,0 +1,156 @@
+/**
+ * @module data-transformers/content-generators/doc-type-helpers
+ * @description Document type display names (multi-language) and title suffix templates.
+ * Provides DOC_TYPE_DISPLAY lookup table, localizeDocType utility, and
+ * TITLE_SUFFIX_TEMPLATES for inverted-pyramid lede construction.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+import type { Language } from '../../types/language.js';
+
+/** Localized singular/plural display names for a Riksdag document type code. */
+export type DocTypeLocalization = {
+  singular: Partial<Record<Language, string>>;
+  plural: Partial<Record<Language, string>>;
+};
+
+/** Multi-language display names for known Riksdag document type codes. */
+export const DOC_TYPE_DISPLAY: Readonly<Record<string, DocTypeLocalization>> = {
+  prop: {
+    singular: {
+      en: 'Proposition', sv: 'Proposition', da: 'Proposition', no: 'Proposisjon',
+      fi: 'Hallituksen esitys', de: 'Regierungsvorlage', fr: 'Projet de loi', es: 'Proposición',
+      nl: 'Wetsvoorstel', ar: 'مقترح قانون', he: 'הצעת חוק', ja: '法案', ko: '정부 제출 법안', zh: '政府法案',
+    },
+    plural: {
+      en: 'Propositions', sv: 'Propositioner', da: 'Propositioner', no: 'Proposisjoner',
+      fi: 'Hallituksen esitykset', de: 'Regierungsvorlagen', fr: 'Projets de loi', es: 'Proposiciones',
+      nl: 'Wetsvoorstellen', ar: 'مقترحات قوانين', he: 'הצעות חוק', ja: '法案', ko: '정부 제출 법안', zh: '政府法案',
+    },
+  },
+  bet: {
+    singular: {
+      en: 'Committee Report', sv: 'Betänkande', da: 'Udvalgsbetænkning', no: 'Komitéinnstilling',
+      fi: 'Valiokunnan mietintö', de: 'Ausschussbericht', fr: 'Rapport de commission', es: 'Informe de comisión',
+      nl: 'Commissieverslag', ar: 'تقرير لجنة', he: 'דוח ועדה', ja: '委員会報告書', ko: '위원회 보고서', zh: '委员会报告',
+    },
+    plural: {
+      en: 'Committee Reports', sv: 'Betänkanden', da: 'Udvalgsbetænkninger', no: 'Komitéinnstillinger',
+      fi: 'Valiokunnan mietinnöt', de: 'Ausschussberichte', fr: 'Rapports de commission', es: 'Informes de comisión',
+      nl: 'Commissieverslagen', ar: 'تقارير لجان', he: 'דוחות ועדה', ja: '委員会報告書', ko: '위원회 보고서', zh: '委员会报告',
+    },
+  },
+  mot: {
+    singular: {
+      en: 'Motion', sv: 'Motion', da: 'Forslag', no: 'Forslag',
+      fi: 'Aloite', de: 'Antrag', fr: 'Motion', es: 'Moción',
+      nl: 'Motie', ar: 'مقترح', he: 'הצעה', ja: '動議', ko: '동의안', zh: '动议',
+    },
+    plural: {
+      en: 'Motions', sv: 'Motioner', da: 'Forslag', no: 'Forslag',
+      fi: 'Aloitteet', de: 'Anträge', fr: 'Motions', es: 'Mociones',
+      nl: 'Moties', ar: 'مقترحات', he: 'הצעות', ja: '動議', ko: '동의안', zh: '动议',
+    },
+  },
+  skr: {
+    singular: {
+      en: 'Government Communication', sv: 'Skrivelse', da: 'Regeringsskrivelse', no: 'Regjeringsskriv',
+      fi: 'Valtioneuvoston kirjelmä', de: 'Regierungsschreiben', fr: 'Communication du gouvernement', es: 'Comunicación del gobierno',
+      nl: 'Regeringsmededeling', ar: 'مذكرة حكومية', he: 'מכתב ממשלתי', ja: '政府通信文書', ko: '정부 통신문', zh: '政府公文',
+    },
+    plural: {
+      en: 'Government Communications', sv: 'Skrivelser', da: 'Regeringsskrivelser', no: 'Regjeringsskriv',
+      fi: 'Valtioneuvoston kirjelmät', de: 'Regierungsschreiben', fr: 'Communications du gouvernement', es: 'Comunicaciones del gobierno',
+      nl: 'Regeringsmededelingen', ar: 'مذكرات حكومية', he: 'מכתבים ממשלתיים', ja: '政府通信文書', ko: '정부 통신문', zh: '政府公文',
+    },
+  },
+  sfs: {
+    singular: {
+      en: 'Law/Statute', sv: 'Lag/förordning', da: 'Lov/forordning', no: 'Lov/forordning',
+      fi: 'Laki/asetus', de: 'Gesetz/Verordnung', fr: 'Loi/Règlement', es: 'Ley/Reglamento',
+      nl: 'Wet/Verordening', ar: 'قانون / لائحة', he: 'חוק/תקנה', ja: '法律／条例', ko: '법률/법규', zh: '法律/法规',
+    },
+    plural: {
+      en: 'Laws/Statutes', sv: 'Lagar/förordningar', da: 'Love/forordninger', no: 'Lover/forordninger',
+      fi: 'Lait/asetukset', de: 'Gesetze/Verordnungen', fr: 'Lois/Règlements', es: 'Leyes/Reglamentos',
+      nl: 'Wetten/Verordeningen', ar: 'قوانين / لوائح', he: 'חוקים/תקנות', ja: '法律／条例', ko: '법률/법규', zh: '法律/法规',
+    },
+  },
+  fpm: {
+    singular: {
+      en: 'EU Position Paper', sv: 'Faktapromemoria', da: 'EU-faktanota', no: 'EU-faktanotat',
+      fi: 'EU-tietomuistio', de: 'EU-Positionspapier', fr: 'Note de position UE', es: 'Documento de posición de la UE',
+      nl: 'EU-positiepaper', ar: 'ورقة موقف للاتحاد الأوروبي', he: 'מסמך עמדה של האיחוד האירופי', ja: 'EUポジションペーパー', ko: 'EU 입장 문서', zh: '欧盟立场文件',
+    },
+    plural: {
+      en: 'EU Position Papers', sv: 'Faktapromemorior', da: 'EU-faktanotaer', no: 'EU-faktanotater',
+      fi: 'EU-tietomuistiot', de: 'EU-Positionspapiere', fr: 'Notes de position UE', es: 'Documentos de posición de la UE',
+      nl: 'EU-positiepapers', ar: 'أوراق موقف للاتحاد الأوروبي', he: 'מסמכי עמדה של האיחוד האירופי', ja: 'EUポジションペーパー', ko: 'EU 입장 문서', zh: '欧盟立场文件',
+    },
+  },
+  pressm: {
+    singular: {
+      en: 'Press Release', sv: 'Pressmeddelande', da: 'Pressemeddelelse', no: 'Pressemelding',
+      fi: 'Lehdistötiedote', de: 'Pressemitteilung', fr: 'Communiqué de presse', es: 'Comunicado de prensa',
+      nl: 'Persbericht', ar: 'بيان صحفي', he: 'הודעה לעיתונות', ja: 'プレスリリース', ko: '보도자료', zh: '新闻稿',
+    },
+    plural: {
+      en: 'Press Releases', sv: 'Pressmeddelanden', da: 'Pressemeddelelser', no: 'Pressemeldinger',
+      fi: 'Lehdistötiedotteet', de: 'Pressemitteilungen', fr: 'Communiqués de presse', es: 'Comunicados de prensa',
+      nl: 'Persberichten', ar: 'بيانات صحفية', he: 'הודעות לעיתונות', ja: 'プレスリリース', ko: '보도자료', zh: '新闻稿',
+    },
+  },
+  ext: {
+    singular: {
+      en: 'External Reference', sv: 'Extern referens', da: 'Ekstern reference', no: 'Ekstern referanse',
+      fi: 'Ulkoinen viite', de: 'Externe Referenz', fr: 'Référence externe', es: 'Referencia externa',
+      nl: 'Externe referentie', ar: 'مرجع خارجي', he: 'הפניה חיצונית', ja: '外部参照', ko: '외부 참조', zh: '外部参考',
+    },
+    plural: {
+      en: 'External References', sv: 'Externa referenser', da: 'Eksterne referencer', no: 'Eksterne referanser',
+      fi: 'Ulkoiset viitteet', de: 'Externe Referenzen', fr: 'Références externes', es: 'Referencias externas',
+      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참照', zh: '外部参考',
+    },
+  },
+  other: {
+    singular: {
+      en: 'Other Document', sv: 'Övrigt dokument', da: 'Andet dokument', no: 'Annet dokument',
+      fi: 'Muu asiakirja', de: 'Sonstiges Dokument', fr: 'Autre document', es: 'Otro documento',
+      nl: 'Overig document', ar: 'مستند آخر', he: 'מסמך אחר', ja: 'その他の文書', ko: '기타 문서', zh: '其他文件',
+    },
+    plural: {
+      en: 'Other Documents', sv: 'Övriga dokument', da: 'Andre dokumenter', no: 'Andre dokumenter',
+      fi: 'Muut asiakirjat', de: 'Sonstige Dokumente', fr: 'Autres documents', es: 'Otros documentos',
+      nl: 'Overige documenten', ar: 'مستندات أخرى', he: 'מסמכים אחרים', ja: 'その他の文書', ko: '기타 문서', zh: '其他文件',
+    },
+  },
+};
+
+/** Localise raw Riksdag document type codes for display (singular/plural-aware, multi-language). */
+export function localizeDocType(code: string, lang: Language | string, count?: number): string {
+  const entry = DOC_TYPE_DISPLAY[code];
+  if (!entry) return code;
+  const usePlural = count !== 1;
+  const primary = usePlural ? entry.plural : entry.singular;
+  const fallback = usePlural ? entry.singular : entry.plural;
+  return primary[lang as Language] ?? primary.en ?? fallback[lang as Language] ?? fallback.en ?? code;
+}
+
+/** Per-language title-suffix templates for inverted-pyramid lede construction. */
+export const TITLE_SUFFIX_TEMPLATES: Readonly<Record<string, (t: string) => string>> = {
+  sv: t => ` — inklusive "${t}"`,
+  da: t => ` — herunder "${t}"`,
+  no: t => ` — inkludert "${t}"`,
+  fi: t => ` — mukaan lukien "${t}"`,
+  de: t => ` — darunter "${t}"`,
+  fr: t => ` — notamment "${t}"`,
+  es: t => ` — incluyendo "${t}"`,
+  nl: t => ` — inclusief "${t}"`,
+  ar: t => ` — بما فيها "${t}"`,
+  he: t => ` — כולל "${t}"`,
+  ja: t => `、「${t}」を含む`,
+  ko: t => `, "${t}" 포함`,
+  zh: t => `，包括"${t}"`,
+};
diff --git a/scripts/data-transformers/content-generators/event-helpers.ts b/scripts/data-transformers/content-generators/event-helpers.ts
new file mode 100644
index 000000000..caec666f9
--- /dev/null
+++ b/scripts/data-transformers/content-generators/event-helpers.ts
@@ -0,0 +1,62 @@
+/**
+ * @module data-transformers/content-generators/event-helpers
+ * @description Calendar event and document matching helpers.
+ * Provides keyword extraction, related document lookup, and
+ * minister extraction from interpellation summaries.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+import type { RawDocument, RawCalendarEvent } from '../types.js';
+
+/** Extract meaningful keywords from text for cross-reference matching
+ *  (min 2 chars, captures EU, KU, etc.; splits on whitespace, hyphens, and commas) */
+function extractKeywords(text: string): string[] {
+  return text.toLowerCase().split(/[\s,–-]+/u).filter(w => w.length >= 2);
+}
+
+/** Find documents related to a calendar event by organ match or keyword overlap (max 3) */
+export function findRelatedDocuments(event: RawCalendarEvent, documents: RawDocument[]): RawDocument[] {
+  const eventOrgan = event.organ ?? '';
+  const keywords = extractKeywords(event.rubrik ?? event.titel ?? event.title ?? '');
+  return documents.filter(doc => {
+    const docOrgan = doc.organ ?? doc.committee ?? '';
+    if (eventOrgan && docOrgan && eventOrgan.toLowerCase() === docOrgan.toLowerCase()) return true;
+    const docText = (doc.titel ?? doc.title ?? '').toLowerCase();
+    return keywords.some(kw => docText.includes(kw));
+  }).slice(0, 3);
+}
+
+/** Find written questions related to a calendar event by keyword overlap (max 3) */
+export function findRelatedQuestions(event: RawCalendarEvent, questions: RawDocument[]): RawDocument[] {
+  const keywords = extractKeywords(event.rubrik ?? event.titel ?? event.title ?? '');
+  return questions.filter(q => {
+    const qText = (q.titel ?? q.title ?? '').toLowerCase();
+    return keywords.some(kw => qText.includes(kw));
+  }).slice(0, 3);
+}
+
+/** Extract targeted minister name from interpellation summary "till MINISTER" header line.
+ *  Strips trailing topic clauses ("om X", "angående Y", etc.) and punctuation. */
+export function extractMinister(summary: string): string {
+  // Use non-newline whitespace ([^\S\n]+) so we don't cross into the next line
+  const m = summary.match(/\btill[^\S\n]+([^\n]+)/i);
+  if (!m) return '';
+  const raw = m[1].trim();
+  if (!raw) return '';
+
+  // Remove common trailing topic clauses and punctuation
+  const lowerRaw = raw.toLowerCase();
+  const stopPhrases = [' om ', ' angående ', ' rörande ', ' beträffande '];
+  let end = raw.length;
+  for (const phrase of stopPhrases) {
+    const idx = lowerRaw.indexOf(phrase);
+    if (idx !== -1 && idx < end) end = idx;
+  }
+  // Cut at terminating punctuation if it comes earlier
+  const punctIdx = raw.search(/[?:;.,]/);
+  if (punctIdx !== -1 && punctIdx < end) end = punctIdx;
+
+  return raw.slice(0, end).trim();
+}
diff --git a/scripts/data-transformers/content-generators/framework-renderers.ts b/scripts/data-transformers/content-generators/framework-renderers.ts
new file mode 100644
index 000000000..747b17ed0
--- /dev/null
+++ b/scripts/data-transformers/content-generators/framework-renderers.ts
@@ -0,0 +1,353 @@
+/**
+ * @module data-transformers/content-generators/framework-renderers
+ * @description HTML renderers for the document analysis framework sections:
+ * PESTLE analysis, stakeholder impact, risk assessment, and implementation assessment.
+ * Also contains the stub types and stub functions that replace the formerly
+ * AI-generated analysis modules (per ai-driven-analysis-guide.md Rule 2).
+ *
+ * ⚠️ DEPRECATED FOR ANALYSIS GENERATION (v3.0, 2026-04-02):
+ * All render functions produce fallback stub HTML.
+ * AI agents in agentic workflow .md files MUST overwrite this output
+ * with genuine political intelligence.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+import { escapeHtml } from '../../html-utils.js';
+import type { Language } from '../../types/language.js';
+import type { CIAContext } from '../types.js';
+
+// ---------------------------------------------------------------------------
+// Stub types — local to this module; replaced by AI-driven analysis
+// ---------------------------------------------------------------------------
+
+export interface PESTLEDimensions {
+  political: string[];
+  economic: string[];
+  social: string[];
+  technological: string[];
+  legal: string[];
+  environmental: string[];
+}
+
+export interface StakeholderDirectImpact {
+  direction: 'positive' | 'negative' | 'mixed' | 'neutral';
+  magnitude: 'significant' | 'moderate' | 'minor';
+  summary: string;
+}
+
+export interface StakeholderImpact {
+  stakeholder: string;
+  displayName: string;
+  directImpact: StakeholderDirectImpact;
+  confidence: string;
+  implementationBurden: 'high' | 'medium' | 'low';
+}
+
+export interface ImplementationAssessment {
+  feasibility: 'high' | 'medium' | 'low';
+  keyObstacles: string[];
+  agenciesInvolved: string[];
+  timeline: string;
+  estimatedTimeline: string;
+  summary: string;
+}
+
+export interface DocumentAnalysis {
+  pestleDimensions: PESTLEDimensions;
+  stakeholderImpacts: StakeholderImpact[];
+  implementationAssessment: ImplementationAssessment;
+  riskAssessment: RiskAssessment[];
+  [key: string]: unknown;
+}
+
+export type PESTLEAnalysis = PESTLEDimensions;
+
+export interface RiskAssessment {
+  type: 'political' | 'implementation' | 'public-acceptance' | 'legal' | 'financial';
+  severity: 'high' | 'medium' | 'low';
+  description: string;
+}
+
+export interface BatchAnalysisResult { results: unknown[] }
+
+// ---------------------------------------------------------------------------
+// Stub functions — return empty data; real analysis is AI-driven in workflows
+// ---------------------------------------------------------------------------
+
+/** Stub: returns empty analysis. Real analysis is AI-driven in workflows. */
+export function analyzeDocumentsBatch(_docs: unknown[], _lang?: Language | string, _cia?: CIAContext): Map<string, DocumentAnalysis> {
+  return new Map();
+}
+
+/** Stub: returns empty perspectives. Real analysis is AI-driven in workflows. */
+export function analyzeDocumentsPerspectives(_docs: unknown[], _cia?: CIAContext, _lang?: Language | string): BatchAnalysisResult {
+  return { results: [] };
+}
+
+// ---------------------------------------------------------------------------
+// Display constants
+// ---------------------------------------------------------------------------
+
+/** Max items per PESTLE dimension in aggregated display */
+const MAX_PESTLE_ITEMS = 4;
+/** Max stakeholder impacts shown in the summary list */
+const MAX_STAKEHOLDER_IMPACTS = 7;
+/** Max risk items shown in the risk assessment summary */
+const MAX_RISK_ITEMS = 5;
+/** Max perspective insights shown from 6-lens analysis */
+export const MAX_PERSPECTIVE_INSIGHTS = 5;
+/** Max implementation obstacles listed */
+const MAX_IMPLEMENTATION_OBSTACLES = 4;
+/** Max agencies displayed in implementation assessment */
+const MAX_AGENCIES_DISPLAYED = 5;
+
+// ---------------------------------------------------------------------------
+// Multi-language label tables
+// ---------------------------------------------------------------------------
+
+const PESTLE_LABELS: Readonly<Record<string, Record<keyof PESTLEAnalysis, string>>> = {
+  en: { political: 'Political', economic: 'Economic', social: 'Social', technological: 'Technological', legal: 'Legal', environmental: 'Environmental' },
+  sv: { political: 'Politisk', economic: 'Ekonomisk', social: 'Social', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljö' },
+  da: { political: 'Politisk', economic: 'Økonomisk', social: 'Social', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljø' },
+  no: { political: 'Politisk', economic: 'Økonomisk', social: 'Sosial', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljø' },
+  fi: { political: 'Poliittinen', economic: 'Taloudellinen', social: 'Sosiaalinen', technological: 'Teknologinen', legal: 'Oikeudellinen', environmental: 'Ympäristö' },
+  de: { political: 'Politisch', economic: 'Wirtschaftlich', social: 'Sozial', technological: 'Technologisch', legal: 'Rechtlich', environmental: 'Umwelt' },
+  fr: { political: 'Politique', economic: 'Économique', social: 'Social', technological: 'Technologique', legal: 'Juridique', environmental: 'Environnemental' },
+  es: { political: 'Político', economic: 'Económico', social: 'Social', technological: 'Tecnológico', legal: 'Jurídico', environmental: 'Ambiental' },
+  nl: { political: 'Politiek', economic: 'Economisch', social: 'Sociaal', technological: 'Technologisch', legal: 'Juridisch', environmental: 'Milieu' },
+  ar: { political: 'سياسي', economic: 'اقتصادي', social: 'اجتماعي', technological: 'تقني', legal: 'قانوني', environmental: 'بيئي' },
+  he: { political: 'פוליטי', economic: 'כלכלי', social: 'חברתי', technological: 'טכנולוגי', legal: 'משפטי', environmental: 'סביבתי' },
+  ja: { political: '政治', economic: '経済', social: '社会', technological: '技術', legal: '法的', environmental: '環境' },
+  ko: { political: '정치', economic: '경제', social: '사회', technological: '기술', legal: '법률', environmental: '환경' },
+  zh: { political: '政治', economic: '经济', social: '社会', technological: '技术', legal: '法律', environmental: '环境' },
+};
+
+const RISK_TYPE_LABELS: Readonly<Record<string, Record<RiskAssessment['type'], string>>> = {
+  en: { political: 'Political', implementation: 'Implementation', 'public-acceptance': 'Public acceptance', legal: 'Legal', financial: 'Financial' },
+  sv: { political: 'Politisk', implementation: 'Genomförande', 'public-acceptance': 'Offentlig acceptans', legal: 'Juridisk', financial: 'Finansiell' },
+  da: { political: 'Politisk', implementation: 'Implementering', 'public-acceptance': 'Offentlig accept', legal: 'Juridisk', financial: 'Finansiel' },
+  no: { political: 'Politisk', implementation: 'Implementering', 'public-acceptance': 'Offentlig aksept', legal: 'Juridisk', financial: 'Finansiell' },
+  fi: { political: 'Poliittinen', implementation: 'Toteutus', 'public-acceptance': 'Julkinen hyväksyntä', legal: 'Oikeudellinen', financial: 'Taloudellinen' },
+  de: { political: 'Politisch', implementation: 'Umsetzung', 'public-acceptance': 'Öffentliche Akzeptanz', legal: 'Rechtlich', financial: 'Finanziell' },
+  fr: { political: 'Politique', implementation: 'Mise en œuvre', 'public-acceptance': 'Acceptation publique', legal: 'Juridique', financial: 'Financier' },
+  es: { political: 'Político', implementation: 'Implementación', 'public-acceptance': 'Aceptación pública', legal: 'Jurídico', financial: 'Financiero' },
+  nl: { political: 'Politiek', implementation: 'Implementatie', 'public-acceptance': 'Publieke acceptatie', legal: 'Juridisch', financial: 'Financieel' },
+  ar: { political: 'سياسي', implementation: 'تنفيذي', 'public-acceptance': 'القبول العام', legal: 'قانوني', financial: 'مالي' },
+  he: { political: 'פוליטי', implementation: 'יישום', 'public-acceptance': 'קבלה ציבורית', legal: 'משפטי', financial: 'פיננסי' },
+  ja: { political: '政治', implementation: '実装', 'public-acceptance': '世論受容', legal: '法的', financial: '財政' },
+  ko: { political: '정치', implementation: '이행', 'public-acceptance': '대중 수용성', legal: '법률', financial: '재정' },
+  zh: { political: '政治', implementation: '实施', 'public-acceptance': '公众接受度', legal: '法律', financial: '财政' },
+};
+
+const LEVEL_LABELS: Readonly<Record<string, Record<'high' | 'medium' | 'low', string>>> = {
+  en: { high: 'High', medium: 'Medium', low: 'Low' },
+  sv: { high: 'Hög', medium: 'Medel', low: 'Låg' },
+  da: { high: 'Høj', medium: 'Mellem', low: 'Lav' },
+  no: { high: 'Høy', medium: 'Middels', low: 'Lav' },
+  fi: { high: 'Korkea', medium: 'Keskitaso', low: 'Matala' },
+  de: { high: 'Hoch', medium: 'Mittel', low: 'Niedrig' },
+  fr: { high: 'Élevé', medium: 'Moyen', low: 'Faible' },
+  es: { high: 'Alto', medium: 'Medio', low: 'Bajo' },
+  nl: { high: 'Hoog', medium: 'Middel', low: 'Laag' },
+  ar: { high: 'مرتفع', medium: 'متوسط', low: 'منخفض' },
+  he: { high: 'גבוה', medium: 'בינוני', low: 'נמוך' },
+  ja: { high: '高', medium: '中', low: '低' },
+  ko: { high: '높음', medium: '보통', low: '낮음' },
+  zh: { high: '高', medium: '中', low: '低' },
+};
+
+const IMPLEMENTATION_LABELS: Readonly<Record<string, { feasibility: string; obstacles: string; agencies: string; noStakeholderData: string; noImplementationData: string; burden: string }>> = {
+  en: { feasibility: 'Feasibility', obstacles: 'Key obstacles', agencies: 'Agencies involved', noStakeholderData: 'No stakeholder impact data available.', noImplementationData: 'No implementation data available.', burden: 'Burden' },
+  sv: { feasibility: 'Genomförbarhet', obstacles: 'Viktiga hinder', agencies: 'Berörda myndigheter', noStakeholderData: 'Ingen data om intressentpåverkan tillgänglig.', noImplementationData: 'Ingen implementeringsdata tillgänglig.', burden: 'Belastning' },
+  da: { feasibility: 'Gennemførlighed', obstacles: 'Vigtige hindringer', agencies: 'Involverede myndigheder', noStakeholderData: 'Ingen data om interessentpåvirkning tilgængelig.', noImplementationData: 'Ingen implementeringsdata tilgængelig.', burden: 'Byrde' },
+  no: { feasibility: 'Gjennomførbarhet', obstacles: 'Viktige hindringer', agencies: 'Involverte etater', noStakeholderData: 'Ingen data om interessentpåvirkning tilgjengelig.', noImplementationData: 'Ingen implementeringsdata tilgjengelig.', burden: 'Belastning' },
+  fi: { feasibility: 'Toteutettavuus', obstacles: 'Keskeiset esteet', agencies: 'Mukana olevat viranomaiset', noStakeholderData: 'Sidosryhmävaikutustietoa ei saatavilla.', noImplementationData: 'Toteutustietoa ei saatavilla.', burden: 'Rasite' },
+  de: { feasibility: 'Umsetzbarkeit', obstacles: 'Wesentliche Hindernisse', agencies: 'Beteiligte Behörden', noStakeholderData: 'Keine Daten zu Stakeholder-Auswirkungen verfügbar.', noImplementationData: 'Keine Umsetzungsdaten verfügbar.', burden: 'Belastung' },
+  fr: { feasibility: 'Faisabilité', obstacles: 'Obstacles clés', agencies: 'Agences impliquées', noStakeholderData: 'Aucune donnée d\'impact des parties prenantes disponible.', noImplementationData: 'Aucune donnée de mise en œuvre disponible.', burden: 'Charge' },
+  es: { feasibility: 'Viabilidad', obstacles: 'Obstáculos clave', agencies: 'Organismos implicados', noStakeholderData: 'No hay datos de impacto en partes interesadas.', noImplementationData: 'No hay datos de implementación disponibles.', burden: 'Carga' },
+  nl: { feasibility: 'Haalbaarheid', obstacles: 'Belangrijkste obstakels', agencies: 'Betrokken instanties', noStakeholderData: 'Geen gegevens over impact op belanghebbenden beschikbaar.', noImplementationData: 'Geen implementatiegegevens beschikbaar.', burden: 'Last' },
+  ar: { feasibility: 'قابلية التنفيذ', obstacles: 'العقبات الرئيسية', agencies: 'الجهات المعنية', noStakeholderData: 'لا تتوفر بيانات تأثير أصحاب المصلحة.', noImplementationData: 'لا تتوفر بيانات تنفيذ.', burden: 'العبء' },
+  he: { feasibility: 'ישימות', obstacles: 'חסמים מרכזיים', agencies: 'גורמים מעורבים', noStakeholderData: 'אין נתוני השפעה על בעלי עניין.', noImplementationData: 'אין נתוני יישום.', burden: 'נטל' },
+  ja: { feasibility: '実現可能性', obstacles: '主な障害', agencies: '関係機関', noStakeholderData: 'ステークホルダー影響データはありません。', noImplementationData: '実施データはありません。', burden: '負担' },
+  ko: { feasibility: '실행 가능성', obstacles: '주요 장애 요인', agencies: '관여 기관', noStakeholderData: '이해관계자 영향 데이터가 없습니다.', noImplementationData: '이행 데이터가 없습니다.', burden: '부담' },
+  zh: { feasibility: '可实施性', obstacles: '关键障碍', agencies: '涉及机构', noStakeholderData: '暂无利益相关方影响数据。', noImplementationData: '暂无实施数据。', burden: '负担' },
+};
+
+// ---------------------------------------------------------------------------
+// Localization helpers
+// ---------------------------------------------------------------------------
+
+function localizeLevel(level: 'high' | 'medium' | 'low', lang: Language | string): string {
+  return LEVEL_LABELS[lang as string]?.[level] ?? LEVEL_LABELS.en[level];
+}
+
+function localizeRiskType(type: RiskAssessment['type'], lang: Language | string): string {
+  return RISK_TYPE_LABELS[lang as string]?.[type] ?? RISK_TYPE_LABELS.en[type];
+}
+
+function localizedImplementationLabels(lang: Language | string): { feasibility: string; obstacles: string; agencies: string; noStakeholderData: string; noImplementationData: string; burden: string } {
+  return IMPLEMENTATION_LABELS[lang as string] ?? IMPLEMENTATION_LABELS.en;
+}
+
+// ---------------------------------------------------------------------------
+// Ranking helpers
+// ---------------------------------------------------------------------------
+
+function severityRank(s: string): number {
+  return s === 'high' ? 3 : s === 'medium' ? 2 : 1;
+}
+
+function magnitudeRank(magnitude: 'significant' | 'moderate' | 'minor'): number {
+  return magnitude === 'significant' ? 3 : magnitude === 'moderate' ? 2 : 1;
+}
+
+function feasibilityRank(f: string): number {
+  return f === 'high' ? 3 : f === 'medium' ? 2 : 1;
+}
+
+// ---------------------------------------------------------------------------
+// Framework analysis HTML renderers
+// ---------------------------------------------------------------------------
+
+/**
+ * Aggregate PESTLE dimensions across multiple document analyses into a
+ * deduplicated list per dimension and render as an HTML description list.
+ */
+export function renderAggregatedPestle(analyses: DocumentAnalysis[], lang: Language | string): string {
+  const merged: PESTLEAnalysis = {
+    political: [], economic: [], social: [],
+    technological: [], legal: [], environmental: [],
+  };
+
+  for (const a of analyses) {
+    const p = a.pestleDimensions;
+    merged.political.push(...p.political);
+    merged.economic.push(...p.economic);
+    merged.social.push(...p.social);
+    merged.technological.push(...p.technological);
+    merged.legal.push(...p.legal);
+    merged.environmental.push(...p.environmental);
+  }
+
+  // Deduplicate per dimension
+  const dedup = (arr: string[]): string[] => [...new Set(arr)].slice(0, MAX_PESTLE_ITEMS);
+
+  const labels = PESTLE_LABELS[lang as string] ?? PESTLE_LABELS.en;
+  const dims: Array<[string, string[]]> = [
+    [labels.political, dedup(merged.political)],
+    [labels.economic, dedup(merged.economic)],
+    [labels.social, dedup(merged.social)],
+    [labels.technological, dedup(merged.technological)],
+    [labels.legal, dedup(merged.legal)],
+    [labels.environmental, dedup(merged.environmental)],
+  ];
+
+  const items = dims
+    .filter(([, items]) => items.length > 0)
+    .map(([label, items]) =>
+      `      <dt><strong>${escapeHtml(label)}</strong></dt>\n      <dd>${items.map(i => escapeHtml(i)).join(' ')}</dd>`,
+    )
+    .join('\n');
+
+  return `    <dl class="pestle-analysis">\n${items}\n    </dl>`;
+}
+
+/**
+ * Render a summary of stakeholder impacts across all analysed documents.
+ * Shows up to 7 stakeholder groups with impact direction, confidence, and burden.
+ */
+export function renderStakeholderImpactSummary(analyses: DocumentAnalysis[], lang: Language | string): string {
+  const labels = localizedImplementationLabels(lang);
+  // Collect all stakeholder impacts, deduplicated by stakeholder name
+  const impactMap = new Map<string, StakeholderImpact>();
+  for (const a of analyses) {
+    for (const impact of a.stakeholderImpacts) {
+      // Keep the higher-magnitude impact per stakeholder
+      const existing = impactMap.get(impact.stakeholder);
+      if (!existing || magnitudeRank(impact.directImpact.magnitude) > magnitudeRank(existing.directImpact.magnitude)) {
+        impactMap.set(impact.stakeholder, impact);
+      }
+    }
+  }
+
+  const impacts = [...impactMap.values()].slice(0, MAX_STAKEHOLDER_IMPACTS);
+  if (impacts.length === 0) return `    <p>${escapeHtml(labels.noStakeholderData)}</p>`;
+
+  const rows = impacts.map(i => {
+    const directionIcon =
+      i.directImpact.direction === 'positive' ? '↑'
+      : i.directImpact.direction === 'negative' ? '↓'
+      : i.directImpact.direction === 'mixed' ? '↕'
+      : '→';
+    const burdenText = localizeLevel(i.implementationBurden, lang);
+    return `      <li><strong>${escapeHtml(i.displayName)}</strong>: ${directionIcon} ${escapeHtml(i.directImpact.summary)} (${escapeHtml(i.confidence)}; ${escapeHtml(labels.burden)}: ${escapeHtml(burdenText)})</li>`;
+  });
+
+  return `    <ul class="stakeholder-impact-list">\n${rows.join('\n')}\n    </ul>`;
+}
+
+/**
+ * Render a risk assessment summary. Groups risks by type and keeps the
+ * highest-severity risk per type.
+ */
+export function renderRiskAssessment(risks: RiskAssessment[], lang: Language | string): string {
+  // Deduplicate by type, preferring higher severity
+  const byType = new Map<string, RiskAssessment>();
+  for (const r of risks) {
+    const key = r.type;
+    const existing = byType.get(key);
+    if (!existing || severityRank(r.severity) > severityRank(existing.severity)) {
+      byType.set(key, r);
+    }
+  }
+
+  const top = [...byType.values()].slice(0, MAX_RISK_ITEMS);
+  const rows = top.map(r => {
+    const icon = r.severity === 'high' ? '🔴' : r.severity === 'medium' ? '🟡' : '🟢';
+    return `      <li>${icon} <strong>${escapeHtml(localizeRiskType(r.type, lang))}</strong> (${escapeHtml(localizeLevel(r.severity, lang))}): ${escapeHtml(r.description)}</li>`;
+  });
+
+  return `    <ul class="risk-assessment-list">\n${rows.join('\n')}\n    </ul>`;
+}
+
+/**
+ * Render implementation assessment summary from framework analyses.
+ */
+export function renderImplementationAssessment(analyses: DocumentAnalysis[], lang: Language | string): string {
+  const labels = localizedImplementationLabels(lang);
+  const assessments: ImplementationAssessment[] = analyses.map(a => a.implementationAssessment);
+  if (assessments.length === 0) return `    <p>${escapeHtml(labels.noImplementationData)}</p>`;
+
+  // Aggregate obstacles and agencies across all documents
+  const allObstacles = new Set<string>();
+  const allAgencies = new Set<string>();
+  let highestFeasibility: ImplementationAssessment['feasibility'] = 'high';
+  let selectedAssessment: ImplementationAssessment = assessments[0];
+
+  for (const ia of assessments) {
+    ia.keyObstacles.forEach(o => allObstacles.add(o));
+    ia.agenciesInvolved.forEach(a => allAgencies.add(a));
+    if (feasibilityRank(ia.feasibility) < feasibilityRank(highestFeasibility)) {
+      highestFeasibility = ia.feasibility;
+      selectedAssessment = ia;
+    }
+  }
+
+  const parts: string[] = [];
+  const fIcon = highestFeasibility === 'high' ? '🟢' : highestFeasibility === 'medium' ? '🟡' : '🔴';
+  const timeline = selectedAssessment.estimatedTimeline;
+  parts.push(`    <p>${fIcon} <strong>${escapeHtml(labels.feasibility)}:</strong> ${escapeHtml(localizeLevel(highestFeasibility, lang))}. ${escapeHtml(timeline)}</p>`);
+
+  if (allObstacles.size > 0) {
+    const obstacleList = [...allObstacles].slice(0, MAX_IMPLEMENTATION_OBSTACLES).map(o => `<li>${escapeHtml(o)}</li>`).join('');
+    parts.push(`    <p><strong>${escapeHtml(labels.obstacles)}:</strong></p>\n    <ul>${obstacleList}</ul>`);
+  }
+
+  if (allAgencies.size > 0) {
+    parts.push(`    <p><strong>${escapeHtml(labels.agencies)}:</strong> ${[...allAgencies].slice(0, MAX_AGENCIES_DISPLAYED).map(a => escapeHtml(a)).join(', ')}</p>`);
+  }
+
+  return parts.join('\n');
+}
diff --git a/scripts/data-transformers/content-generators/impact-helpers.ts b/scripts/data-transformers/content-generators/impact-helpers.ts
new file mode 100644
index 000000000..4cc1c1f71
--- /dev/null
+++ b/scripts/data-transformers/content-generators/impact-helpers.ts
@@ -0,0 +1,110 @@
+/**
+ * @module data-transformers/content-generators/impact-helpers
+ * @description Political impact and legislative consequences analysis helpers.
+ * Generates AI_MUST_REPLACE stub paragraphs for political impact,
+ * implementation consequences, and related analysis sections.
+ *
+ * ⚠️ DEPRECATED FOR ANALYSIS GENERATION (v3.0, 2026-04-02):
+ * All functions in this module produce fallback stub text.
+ * AI agents in agentic workflow .md files MUST overwrite this
+ * template-generated text with genuine political intelligence.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+import type { RawDocument } from '../types.js';
+import type { CIAContext } from '../types.js';
+import type { Language } from '../../types/language.js';
+
+// ---------------------------------------------------------------------------
+// Political impact analysis (AI_MUST_REPLACE stubs)
+// ---------------------------------------------------------------------------
+
+function propImpactText(_lang: Language | string, n: number): string {
+  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} government proposition(s) filed. Analyse the specific political impact: which propositions face opposition, what coalition dynamics are at play, which policy areas face the strongest resistance, and what the vote arithmetic looks like. Output MUST be in the article's language. -->`;
+}
+
+function betImpactText(_lang: Language | string, n: number): string {
+  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} committee report(s) issued. Analyse the political significance of these specific committee recommendations: key votes, reservation patterns, which parties are aligned or divided, and implications for chamber votes. Output MUST be in the article's language. -->`;
+}
+
+function motImpactText(_lang: Language | string, n: number): string {
+  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} opposition motion(s) filed. Analyse the strategic purpose of these specific motions: which target government vulnerabilities, which signal election campaign themes, and what they reveal about opposition coordination. Output MUST be in the article's language. -->`;
+}
+
+function thinMajorityImpactText(_lang: Language | string, margin: number): string {
+  return `<!-- AI_MUST_REPLACE: majority_impact — DATA: majority margin ${margin} seat(s). Analyse how this thin margin affects these specific legislative items and which measures are most vulnerable to defection. Output MUST be in the article's language. -->`;
+}
+
+function genericImpactText(_lang: Language | string): string {
+  return '<!-- AI_MUST_REPLACE: political_impact — Write specific analysis of the political impact of these items, naming parties, citing vote margins, and identifying which measures will pass or face challenges. Output MUST be in the article\'s language. -->';
+}
+
+/**
+ * Generate a political impact analysis paragraph for the deep analysis section.
+ * Returns an AI_MUST_REPLACE stub that AI agents must overwrite.
+ */
+export function generateImpactAnalysis(docs: RawDocument[], lang: Language | string, cia: CIAContext | undefined): string {
+  const parts: string[] = [];
+
+  const propCount = docs.filter(d => d.doktyp === 'prop').length;
+  const motCount = docs.filter(d => d.doktyp === 'mot').length;
+  const betCount = docs.filter(d => d.doktyp === 'bet').length;
+
+  if (propCount > 0) {
+    parts.push(propImpactText(lang, propCount));
+  }
+  if (betCount > 0) {
+    parts.push(betImpactText(lang, betCount));
+  }
+  if (motCount > 0) {
+    parts.push(motImpactText(lang, motCount));
+  }
+
+  if (cia) {
+    const margin = cia.coalitionStability?.majorityMargin ?? 0;
+    if (margin <= 5) {
+      parts.push(thinMajorityImpactText(lang, margin));
+    }
+  }
+
+  return parts.join(' ') || genericImpactText(lang);
+}
+
+// ---------------------------------------------------------------------------
+// Legislative consequences analysis (AI_MUST_REPLACE stubs)
+// ---------------------------------------------------------------------------
+
+function propConsequencesText(_lang: Language | string, n: number): string {
+  return `<!-- AI_MUST_REPLACE: consequences — DATA: ${n} proposition(s) pending. Analyse the specific implementation consequences: which agencies must act, what regulatory changes are required, budget implications, and timeline for each proposition. Output MUST be in the article's language. -->`;
+}
+
+function motConsequencesText(_lang: Language | string, n: number): string {
+  return `<!-- AI_MUST_REPLACE: consequences — DATA: ${n} opposition motion(s) pending. Analyse the strategic consequences: how will rejection/acceptance affect party positioning, which motions establish new policy alternatives, and what campaign value do they carry. Output MUST be in the article's language. -->`;
+}
+
+function genericConsequencesText(_lang: Language | string): string {
+  return '<!-- AI_MUST_REPLACE: consequences — Write specific analysis of the consequences and next steps for these items, including committee timelines, expected vote dates, implementation requirements, and political ramifications. Output MUST be in the article\'s language. -->';
+}
+
+/**
+ * Generate a legislative consequences analysis paragraph for the deep analysis section.
+ * Returns AI_MUST_REPLACE stubs that AI agents must overwrite.
+ */
+export function generateConsequencesAnalysis(docs: RawDocument[], lang: Language | string, _articleType: string): string {
+  const propCount = docs.filter(d => d.doktyp === 'prop').length;
+  const motCount = docs.filter(d => d.doktyp === 'mot').length;
+  const parts: string[] = [];
+
+  if (propCount > 0) {
+    parts.push(propConsequencesText(lang, propCount));
+  }
+  if (motCount > 0) {
+    parts.push(motConsequencesText(lang, motCount));
+  }
+  if (parts.length === 0) {
+    parts.push(genericConsequencesText(lang));
+  }
+  return parts.join(' ');
+}
diff --git a/scripts/data-transformers/content-generators/shared.ts b/scripts/data-transformers/content-generators/shared.ts
index a6258bf0d..268fadc21 100644
--- a/scripts/data-transformers/content-generators/shared.ts
+++ b/scripts/data-transformers/content-generators/shared.ts
@@ -4,6 +4,13 @@
  * Contains TITLE_SUFFIX_TEMPLATES, keyword extraction, event/document matching helpers,
  * and the deep analysis section generator (5W framework).
  *
+ * Implementation is split into focused sub-modules:
+ * - doc-type-helpers.ts    — DOC_TYPE_DISPLAY, localizeDocType, TITLE_SUFFIX_TEMPLATES
+ * - event-helpers.ts       — findRelatedDocuments, findRelatedQuestions, extractMinister
+ * - impact-helpers.ts      — generateImpactAnalysis, generateConsequencesAnalysis
+ * - framework-renderers.ts — PESTLE/stakeholder/risk/implementation HTML renderers
+ * - ai-marker-helpers.ts   — detectBannedPatterns
+ *
  * ⚠️ DEPRECATED FOR ANALYSIS GENERATION (v3.0, 2026-04-02):
  * Per analysis/methodologies/ai-driven-analysis-guide.md Rule 2, the following
  * functions are DEPRECATED for generating analysis content:
@@ -26,265 +33,26 @@
 
 import { escapeHtml } from '../../html-utils.js';
 import type { Language } from '../../types/language.js';
-import type { RawDocument, RawCalendarEvent, CIAContext } from '../types.js';
+import type { RawDocument, CIAContext } from '../types.js';
 import { L, normalizePartyKey } from '../helpers.js';
 import { detectPolicyDomains } from '../policy-analysis.js';
 
-/* ── Stub types/functions for deleted analysis modules ── */
-/* Per ai-driven-analysis-guide.md Rule 2: scripts must NOT generate analysis */
-
-interface PESTLEDimensions {
-  political: string[];
-  economic: string[];
-  social: string[];
-  technological: string[];
-  legal: string[];
-  environmental: string[];
-}
-
-interface StakeholderDirectImpact {
-  direction: 'positive' | 'negative' | 'mixed' | 'neutral';
-  magnitude: 'significant' | 'moderate' | 'minor';
-  summary: string;
-}
-
-interface StakeholderImpact {
-  stakeholder: string;
-  displayName: string;
-  directImpact: StakeholderDirectImpact;
-  confidence: string;
-  implementationBurden: 'high' | 'medium' | 'low';
-}
-
-interface ImplementationAssessment {
-  feasibility: 'high' | 'medium' | 'low';
-  keyObstacles: string[];
-  agenciesInvolved: string[];
-  timeline: string;
-  estimatedTimeline: string;
-  summary: string;
-}
-
-interface DocumentAnalysis {
-  pestleDimensions: PESTLEDimensions;
-  stakeholderImpacts: StakeholderImpact[];
-  implementationAssessment: ImplementationAssessment;
-  riskAssessment: RiskAssessment[];
-  [key: string]: unknown;
-}
-
-type PESTLEAnalysis = PESTLEDimensions;
-
-interface RiskAssessment {
-  type: 'political' | 'implementation' | 'public-acceptance' | 'legal' | 'financial';
-  severity: 'high' | 'medium' | 'low';
-  description: string;
-}
-
-interface BatchAnalysisResult { results: unknown[] }
-
-/** Stub: returns empty analysis. Real analysis is AI-driven in workflows. */
-function analyzeDocumentsBatch(_docs: unknown[], _lang?: Language | string, _cia?: CIAContext): Map<string, DocumentAnalysis> {
-  return new Map();
-}
-
-/** Stub: returns empty perspectives. Real analysis is AI-driven in workflows. */
-function analyzeDocumentsPerspectives(_docs: unknown[], _cia?: CIAContext, _lang?: Language | string): BatchAnalysisResult {
-  return { results: [] };
-}
-
-/** Localise raw Riksdag document type codes for display (singular/plural-aware, multi-language). */
-export type DocTypeLocalization = {
-  singular: Partial<Record<Language, string>>;
-  plural: Partial<Record<Language, string>>;
-};
-
-export const DOC_TYPE_DISPLAY: Readonly<Record<string, DocTypeLocalization>> = {
-  prop: {
-    singular: {
-      en: 'Proposition', sv: 'Proposition', da: 'Proposition', no: 'Proposisjon',
-      fi: 'Hallituksen esitys', de: 'Regierungsvorlage', fr: 'Projet de loi', es: 'Proposición',
-      nl: 'Wetsvoorstel', ar: 'مقترح قانون', he: 'הצעת חוק', ja: '法案', ko: '정부 제출 법안', zh: '政府法案',
-    },
-    plural: {
-      en: 'Propositions', sv: 'Propositioner', da: 'Propositioner', no: 'Proposisjoner',
-      fi: 'Hallituksen esitykset', de: 'Regierungsvorlagen', fr: 'Projets de loi', es: 'Proposiciones',
-      nl: 'Wetsvoorstellen', ar: 'مقترحات قوانين', he: 'הצעות חוק', ja: '法案', ko: '정부 제출 법안', zh: '政府法案',
-    },
-  },
-  bet: {
-    singular: {
-      en: 'Committee Report', sv: 'Betänkande', da: 'Udvalgsbetænkning', no: 'Komitéinnstilling',
-      fi: 'Valiokunnan mietintö', de: 'Ausschussbericht', fr: 'Rapport de commission', es: 'Informe de comisión',
-      nl: 'Commissieverslag', ar: 'تقرير لجنة', he: 'דוח ועדה', ja: '委員会報告書', ko: '위원회 보고서', zh: '委员会报告',
-    },
-    plural: {
-      en: 'Committee Reports', sv: 'Betänkanden', da: 'Udvalgsbetænkninger', no: 'Komitéinnstillinger',
-      fi: 'Valiokunnan mietinnöt', de: 'Ausschussberichte', fr: 'Rapports de commission', es: 'Informes de comisión',
-      nl: 'Commissieverslagen', ar: 'تقارير لجان', he: 'דוחות ועדה', ja: '委員会報告書', ko: '위원회 보고서', zh: '委员会报告',
-    },
-  },
-  mot: {
-    singular: {
-      en: 'Motion', sv: 'Motion', da: 'Forslag', no: 'Forslag',
-      fi: 'Aloite', de: 'Antrag', fr: 'Motion', es: 'Moción',
-      nl: 'Motie', ar: 'مقترح', he: 'הצעה', ja: '動議', ko: '동의안', zh: '动议',
-    },
-    plural: {
-      en: 'Motions', sv: 'Motioner', da: 'Forslag', no: 'Forslag',
-      fi: 'Aloitteet', de: 'Anträge', fr: 'Motions', es: 'Mociones',
-      nl: 'Moties', ar: 'مقترحات', he: 'הצעות', ja: '動議', ko: '동의안', zh: '动议',
-    },
-  },
-  skr: {
-    singular: {
-      en: 'Government Communication', sv: 'Skrivelse', da: 'Regeringsskrivelse', no: 'Regjeringsskriv',
-      fi: 'Valtioneuvoston kirjelmä', de: 'Regierungsschreiben', fr: 'Communication du gouvernement', es: 'Comunicación del gobierno',
-      nl: 'Regeringsmededeling', ar: 'مذكرة حكومية', he: 'מכתב ממשלתי', ja: '政府通信文書', ko: '정부 통신문', zh: '政府公文',
-    },
-    plural: {
-      en: 'Government Communications', sv: 'Skrivelser', da: 'Regeringsskrivelser', no: 'Regjeringsskriv',
-      fi: 'Valtioneuvoston kirjelmät', de: 'Regierungsschreiben', fr: 'Communications du gouvernement', es: 'Comunicaciones del gobierno',
-      nl: 'Regeringsmededelingen', ar: 'مذكرات حكومية', he: 'מכתבים ממשלתיים', ja: '政府通信文書', ko: '정부 통신문', zh: '政府公文',
-    },
-  },
-  sfs: {
-    singular: {
-      en: 'Law/Statute', sv: 'Lag/förordning', da: 'Lov/forordning', no: 'Lov/forordning',
-      fi: 'Laki/asetus', de: 'Gesetz/Verordnung', fr: 'Loi/Règlement', es: 'Ley/Reglamento',
-      nl: 'Wet/Verordening', ar: 'قانون / لائحة', he: 'חוק/תקנה', ja: '法律／条例', ko: '법률/법규', zh: '法律/法规',
-    },
-    plural: {
-      en: 'Laws/Statutes', sv: 'Lagar/förordningar', da: 'Love/forordninger', no: 'Lover/forordninger',
-      fi: 'Lait/asetukset', de: 'Gesetze/Verordnungen', fr: 'Lois/Règlements', es: 'Leyes/Reglamentos',
-      nl: 'Wetten/Verordeningen', ar: 'قوانين / لوائح', he: 'חוקים/תקנות', ja: '法律／条例', ko: '법률/법규', zh: '法律/法规',
-    },
-  },
-  fpm: {
-    singular: {
-      en: 'EU Position Paper', sv: 'Faktapromemoria', da: 'EU-faktanota', no: 'EU-faktanotat',
-      fi: 'EU-tietomuistio', de: 'EU-Positionspapier', fr: 'Note de position UE', es: 'Documento de posición de la UE',
-      nl: 'EU-positiepaper', ar: 'ورقة موقف للاتحاد الأوروبي', he: 'מסמך עמדה של האיחוד האירופי', ja: 'EUポジションペーパー', ko: 'EU 입장 문서', zh: '欧盟立场文件',
-    },
-    plural: {
-      en: 'EU Position Papers', sv: 'Faktapromemorior', da: 'EU-faktanotaer', no: 'EU-faktanotater',
-      fi: 'EU-tietomuistiot', de: 'EU-Positionspapiere', fr: 'Notes de position UE', es: 'Documentos de posición de la UE',
-      nl: 'EU-positiepapers', ar: 'أوراق موقف للاتحاد الأوروبي', he: 'מסמכי עמדה של האיחוד האירופי', ja: 'EUポジションペーパー', ko: 'EU 입장 문서', zh: '欧盟立场文件',
-    },
-  },
-  pressm: {
-    singular: {
-      en: 'Press Release', sv: 'Pressmeddelande', da: 'Pressemeddelelse', no: 'Pressemelding',
-      fi: 'Lehdistötiedote', de: 'Pressemitteilung', fr: 'Communiqué de presse', es: 'Comunicado de prensa',
-      nl: 'Persbericht', ar: 'بيان صحفي', he: 'הודעה לעיתונות', ja: 'プレスリリース', ko: '보도자료', zh: '新闻稿',
-    },
-    plural: {
-      en: 'Press Releases', sv: 'Pressmeddelanden', da: 'Pressemeddelelser', no: 'Pressemeldinger',
-      fi: 'Lehdistötiedotteet', de: 'Pressemitteilungen', fr: 'Communiqués de presse', es: 'Comunicados de prensa',
-      nl: 'Persberichten', ar: 'بيانات صحفية', he: 'הודעות לעיתונות', ja: 'プレスリリース', ko: '보도자료', zh: '新闻稿',
-    },
-  },
-  ext: {
-    singular: {
-      en: 'External Reference', sv: 'Extern referens', da: 'Ekstern reference', no: 'Ekstern referanse',
-      fi: 'Ulkoinen viite', de: 'Externe Referenz', fr: 'Référence externe', es: 'Referencia externa',
-      nl: 'Externe referentie', ar: 'مرجع خارجي', he: 'הפניה חיצונית', ja: '外部参照', ko: '외부 참조', zh: '外部参考',
-    },
-    plural: {
-      en: 'External References', sv: 'Externa referenser', da: 'Eksterne referencer', no: 'Eksterne referanser',
-      fi: 'Ulkoiset viitteet', de: 'Externe Referenzen', fr: 'Références externes', es: 'Referencias externas',
-      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참조', zh: '外部参考',
-    },
-  },
-  other: {
-    singular: {
-      en: 'Other Document', sv: 'Övrigt dokument', da: 'Andet dokument', no: 'Annet dokument',
-      fi: 'Muu asiakirja', de: 'Sonstiges Dokument', fr: 'Autre document', es: 'Otro documento',
-      nl: 'Overig document', ar: 'مستند آخر', he: 'מסמך אחר', ja: 'その他の文書', ko: '기타 문서', zh: '其他文件',
-    },
-    plural: {
-      en: 'Other Documents', sv: 'Övriga dokument', da: 'Andre dokumenter', no: 'Andre dokumenter',
-      fi: 'Muut asiakirjat', de: 'Sonstige Dokumente', fr: 'Autres documents', es: 'Otros documentos',
-      nl: 'Overige documenten', ar: 'مستندات أخرى', he: 'מסמכים אחרים', ja: 'その他の文書', ko: '기타 문서', zh: '其他文件',
-    },
-  },
-};
-
-export function localizeDocType(code: string, lang: Language | string, count?: number): string {
-  const entry = DOC_TYPE_DISPLAY[code];
-  if (!entry) return code;
-  const usePlural = count !== 1;
-  const primary = usePlural ? entry.plural : entry.singular;
-  const fallback = usePlural ? entry.singular : entry.plural;
-  return primary[lang as Language] ?? primary.en ?? fallback[lang as Language] ?? fallback.en ?? code;
-}
-
-/** Per-language title-suffix templates for inverted-pyramid lede construction. */
-export const TITLE_SUFFIX_TEMPLATES: Readonly<Record<string, (t: string) => string>> = {
-  sv: t => ` — inklusive "${t}"`,
-  da: t => ` — herunder "${t}"`,
-  no: t => ` — inkludert "${t}"`,
-  fi: t => ` — mukaan lukien "${t}"`,
-  de: t => ` — darunter "${t}"`,
-  fr: t => ` — notamment "${t}"`,
-  es: t => ` — incluyendo "${t}"`,
-  nl: t => ` — inclusief "${t}"`,
-  ar: t => ` — بما فيها "${t}"`,
-  he: t => ` — כולל "${t}"`,
-  ja: t => `、「${t}」を含む`,
-  ko: t => `, "${t}" 포함`,
-  zh: t => `，包括"${t}"`,
-};
-
-/** Extract meaningful keywords from text for cross-reference matching (min 2 chars, captures EU, KU, etc.; splits on whitespace, hyphens, and commas) */
-function extractKeywords(text: string): string[] {
-  return text.toLowerCase().split(/[\s,–-]+/u).filter(w => w.length >= 2);
-}
-
-/** Find documents related to a calendar event by organ match or keyword overlap (max 3) */
-export function findRelatedDocuments(event: RawCalendarEvent, documents: RawDocument[]): RawDocument[] {
-  const eventOrgan = event.organ ?? '';
-  const keywords = extractKeywords(event.rubrik ?? event.titel ?? event.title ?? '');
-  return documents.filter(doc => {
-    const docOrgan = doc.organ ?? doc.committee ?? '';
-    if (eventOrgan && docOrgan && eventOrgan.toLowerCase() === docOrgan.toLowerCase()) return true;
-    const docText = (doc.titel ?? doc.title ?? '').toLowerCase();
-    return keywords.some(kw => docText.includes(kw));
-  }).slice(0, 3);
-}
-
-/** Find written questions related to a calendar event by keyword overlap (max 3) */
-export function findRelatedQuestions(event: RawCalendarEvent, questions: RawDocument[]): RawDocument[] {
-  const keywords = extractKeywords(event.rubrik ?? event.titel ?? event.title ?? '');
-  return questions.filter(q => {
-    const qText = (q.titel ?? q.title ?? '').toLowerCase();
-    return keywords.some(kw => qText.includes(kw));
-  }).slice(0, 3);
-}
-
-/** Extract targeted minister name from interpellation summary "till MINISTER" header line.
- *  Strips trailing topic clauses ("om X", "angående Y", etc.) and punctuation. */
-export function extractMinister(summary: string): string {
-  // Use non-newline whitespace ([^\S\n]+) so we don't cross into the next line
-  const m = summary.match(/\btill[^\S\n]+([^\n]+)/i);
-  if (!m) return '';
-  const raw = m[1].trim();
-  if (!raw) return '';
-
-  // Remove common trailing topic clauses and punctuation
-  const lowerRaw = raw.toLowerCase();
-  const stopPhrases = [' om ', ' angående ', ' rörande ', ' beträffande '];
-  let end = raw.length;
-  for (const phrase of stopPhrases) {
-    const idx = lowerRaw.indexOf(phrase);
-    if (idx !== -1 && idx < end) end = idx;
-  }
-  // Cut at terminating punctuation if it comes earlier
-  const punctIdx = raw.search(/[?:;.,]/);
-  if (punctIdx !== -1 && punctIdx < end) end = punctIdx;
-
-  return raw.slice(0, end).trim();
-}
+// ── Re-exports from sub-modules (backward-compatible) ──────────────────────
+export type { DocTypeLocalization } from './doc-type-helpers.js';
+export { DOC_TYPE_DISPLAY, localizeDocType, TITLE_SUFFIX_TEMPLATES } from './doc-type-helpers.js';
+export { findRelatedDocuments, findRelatedQuestions, extractMinister } from './event-helpers.js';
+export { detectBannedPatterns } from './ai-marker-helpers.js';
+
+// ── Sub-module imports used by the deep analysis section ───────────────────
+import { generateImpactAnalysis, generateConsequencesAnalysis } from './impact-helpers.js';
+import {
+  type DocumentAnalysis, type BatchAnalysisResult, type RiskAssessment,
+  analyzeDocumentsBatch, analyzeDocumentsPerspectives,
+  MAX_PERSPECTIVE_INSIGHTS,
+  renderAggregatedPestle, renderStakeholderImpactSummary,
+  renderRiskAssessment, renderImplementationAssessment,
+} from './framework-renderers.js';
+import { localizeDocType } from './doc-type-helpers.js';
 
 // ---------------------------------------------------------------------------
 // Deep Analysis Section (5W Framework)
@@ -339,6 +107,10 @@ export function analyzeDocumentsForContent(
   return { frameworkAnalysis, perspectiveAnalysis };
 }
 
+// ---------------------------------------------------------------------------
+// Deep Analysis section private helpers
+// ---------------------------------------------------------------------------
+
 /**
  * Extract unique party names from a set of documents for "Who" analysis.
  */
@@ -453,150 +225,6 @@ function coalitionRiskText(lang: Language | string, margin: number): string {
   return templates[lang as string] ?? templates.en;
 }
 
-/**
- * Generate a comprehensive Deep Analysis section following the 5W framework
- * (Who, What, When, Why, Winners/Losers) plus impact, consequences, and critical
- * assessment subsections. This section is designed for highly analytical readers
- * who seek multi-perspective intelligence on parliamentary developments.
- *
- * @returns HTML string for the deep analysis section, or empty string if insufficient data
- */
-export function generateDeepAnalysisSection(opts: DeepAnalysisOptions): string {
-  const { documents, lang, cia, articleType, whyContext, frameworkAnalysis, perspectiveAnalysis } = opts;
-
-  // Deep analysis requires at least 2 documents for cross-document insights
-  // in standard article types. For deep-inspection articles, allow single-
-  // document analysis since the whole article is dedicated to in-depth review.
-  const minDocs = articleType === 'deep-inspection' ? 1 : 2;
-  if (!documents || documents.length < minDocs) return '';
-
-  const lbl = (key: string): string => {
-    const val = L(lang, key);
-    return typeof val === 'string' ? val : key;
-  };
-
-  const parts: string[] = [];
-  parts.push(`\n    <section class="deep-analysis" aria-label="${escapeHtml(lbl('deepAnalysis'))}">`);
-  parts.push(`    <h2>${escapeHtml(lbl('deepAnalysis'))}</h2>`);
-
-  // ── WHO: Key Actors ────────────────────────────────────────────────────────
-  const { parties, authors } = extractKeyActors(documents);
-  if (parties.size > 0 || authors.length > 0) {
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWho'))}</h3>`);
-    if (parties.size > 0) {
-      const sortedParties = [...parties.entries()].sort((a, b) => b[1] - a[1]);
-      const partyList = sortedParties
-        .map(([p, count]) => `<strong>${escapeHtml(p)}</strong> (${count})`)
-        .join(', ');
-      parts.push(`    <p>${partyList}</p>`);
-    }
-    if (authors.length > 0) {
-      parts.push(`    <p>${authors.map(a => escapeHtml(a)).join(', ')}</p>`);
-    }
-  }
-
-  // ── WHAT: What Happened ────────────────────────────────────────────────────
-  const domains = aggregateDomains(documents, lang);
-  if (domains.size > 0) {
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhat'))}</h3>`);
-    const sortedDomains = [...domains.entries()].sort((a, b) => b[1] - a[1]);
-    const domainItems = sortedDomains.slice(0, 6)
-      .map(([d, c]) => `${escapeHtml(d)} (${c})`)
-      .join(', ');
-    const docTypes = new Map<string, number>();
-    for (const doc of documents) {
-      const t = doc.doktyp || doc.documentType || 'other';
-      docTypes.set(t, (docTypes.get(t) ?? 0) + 1);
-    }
-    const typeList = [...docTypes.entries()]
-      .sort((a, b) => b[1] - a[1])
-      .map(([t, c]) => `${escapeHtml(localizeDocType(t, lang, c))}: ${c}`)
-      .join(', ');
-    parts.push(`    <p>${domainItems}</p>`);
-    parts.push(`    <p><em>${typeList}</em></p>`);
-  }
-
-  // ── WHEN: Timeline & Context ───────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhen'))}</h3>`);
-  const timelineContext = generateTimelineContext(documents, lang, articleType);
-  parts.push(`    <p>${timelineContext}</p>`);
-
-  // ── WHY: Why This Matters ──────────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhy'))}</h3>`);
-  const whyText = generateWhyAnalysis(documents, lang, cia, whyContext);
-  parts.push(`    <p>${whyText}</p>`);
-
-  // ── WINNERS & LOSERS ───────────────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWinners'))}</h3>`);
-  const winnersText = analyseWinnersLosers(documents, cia, lang);
-  parts.push(`    <p>${winnersText}</p>`);
-
-  // ── POLITICAL IMPACT ───────────────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisImpact'))}</h3>`);
-  const impactText = generateImpactAnalysis(documents, lang, cia);
-  parts.push(`    <p>${impactText}</p>`);
-
-  // ── ACTIONS & CONSEQUENCES ─────────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisConsequences'))}</h3>`);
-  const consequencesText = generateConsequencesAnalysis(documents, lang, articleType);
-  parts.push(`    <p>${consequencesText}</p>`);
-
-  // ── CRITICAL ASSESSMENT ────────────────────────────────────────────────────
-  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisCritical'))}</h3>`);
-  const criticalText = generateCriticalAssessment(documents, lang, cia);
-  parts.push(`    <p>${criticalText}</p>`);
-
-  // ── MULTIPLE PERSPECTIVES ──────────────────────────────────────────────────
-  if (parties.size >= 2) {
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisPerspectives'))}</h3>`);
-    const perspectivesText = generatePerspectivesAnalysis(documents, lang, parties);
-    parts.push(`    <p>${perspectivesText}</p>`);
-  }
-
-  // ── FRAMEWORK ANALYSIS SECTIONS ────────────────────────────────────────────
-  // When the document analysis framework has been run, inject its richer
-  // PESTLE, stakeholder impact, risk, and implementation assessment data.
-  if (frameworkAnalysis && frameworkAnalysis.size > 0) {
-    const analyses = [...frameworkAnalysis.values()];
-
-    // PESTLE Analysis — aggregate across all analysed documents
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisPestle'))}</h3>`);
-    parts.push(renderAggregatedPestle(analyses, lang));
-
-    // Stakeholder Impact — summarise stakeholder impacts from the framework
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisStakeholderImpact'))}</h3>`);
-    parts.push(renderStakeholderImpactSummary(analyses, lang));
-
-    // Risk Assessment — aggregate risk factors across documents
-    const allRisks = analyses.flatMap(a => a.riskAssessment);
-    if (allRisks.length > 0) {
-      parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisRisk'))}</h3>`);
-      parts.push(renderRiskAssessment(allRisks, lang));
-    }
-
-    // Implementation Assessment — summarise implementation feasibility
-    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisImplementation'))}</h3>`);
-    parts.push(renderImplementationAssessment(analyses, lang));
-  }
-
-  // ── MULTI-PERSPECTIVE INSIGHTS (6 lenses) ────────────────────────────────
-  // When the analysis-framework has been run, inject key insights from the
-  // government, opposition, citizen, economic, international, and media lenses.
-  if (perspectiveAnalysis && perspectiveAnalysis.results.length > 0) {
-    const allInsights = perspectiveAnalysis.results.flatMap((r: unknown) => ((r as { keyInsights?: string[] }).keyInsights ?? []));
-    if (allInsights.length > 0) {
-      const uniqueInsights = [...new Set(allInsights)].slice(0, MAX_PERSPECTIVE_INSIGHTS);
-      parts.push(`    <div class="perspective-insights">`);
-      const insightItems = uniqueInsights.map(i => `      <li>${escapeHtml(i)}</li>`).join('\n');
-      parts.push(`    <ul>\n${insightItems}\n    </ul>`);
-      parts.push(`    </div>`);
-    }
-  }
-
-  parts.push('    </section>\n');
-  return parts.join('\n');
-}
-
 // ---------------------------------------------------------------------------
 // Deep Analysis subsection generators
 // ---------------------------------------------------------------------------
@@ -671,82 +299,6 @@ function defaultWhyText(_lang: Language | string): string {
   return '<!-- AI_MUST_REPLACE: why_matters — Write specific analysis of why these particular parliamentary developments matter, citing concrete political dynamics, stakeholder impacts, and strategic implications. Output MUST be in the article\'s language. -->';
 }
 
-function generateImpactAnalysis(docs: RawDocument[], lang: Language | string, cia: CIAContext | undefined): string {
-  const parts: string[] = [];
-
-  const propCount = docs.filter(d => d.doktyp === 'prop').length;
-  const motCount = docs.filter(d => d.doktyp === 'mot').length;
-  const betCount = docs.filter(d => d.doktyp === 'bet').length;
-
-  if (propCount > 0) {
-    parts.push(propImpactText(lang, propCount));
-  }
-  if (betCount > 0) {
-    parts.push(betImpactText(lang, betCount));
-  }
-  if (motCount > 0) {
-    parts.push(motImpactText(lang, motCount));
-  }
-
-  if (cia) {
-    const margin = cia.coalitionStability?.majorityMargin ?? 0;
-    if (margin <= 5) {
-      parts.push(thinMajorityImpactText(lang, margin));
-    }
-  }
-
-  return parts.join(' ') || genericImpactText(lang);
-}
-
-function propImpactText(_lang: Language | string, n: number): string {
-  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} government proposition(s) filed. Analyse the specific political impact: which propositions face opposition, what coalition dynamics are at play, which policy areas face the strongest resistance, and what the vote arithmetic looks like. Output MUST be in the article's language. -->`;
-}
-
-function betImpactText(_lang: Language | string, n: number): string {
-  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} committee report(s) issued. Analyse the political significance of these specific committee recommendations: key votes, reservation patterns, which parties are aligned or divided, and implications for chamber votes. Output MUST be in the article's language. -->`;
-}
-
-function motImpactText(_lang: Language | string, n: number): string {
-  return `<!-- AI_MUST_REPLACE: political_impact — DATA: ${n} opposition motion(s) filed. Analyse the strategic purpose of these specific motions: which target government vulnerabilities, which signal election campaign themes, and what they reveal about opposition coordination. Output MUST be in the article's language. -->`;
-}
-
-function thinMajorityImpactText(_lang: Language | string, margin: number): string {
-  return `<!-- AI_MUST_REPLACE: majority_impact — DATA: majority margin ${margin} seat(s). Analyse how this thin margin affects these specific legislative items and which measures are most vulnerable to defection. Output MUST be in the article's language. -->`;
-}
-
-function genericImpactText(_lang: Language | string): string {
-  return '<!-- AI_MUST_REPLACE: political_impact — Write specific analysis of the political impact of these items, naming parties, citing vote margins, and identifying which measures will pass or face challenges. Output MUST be in the article\'s language. -->';
-}
-
-function generateConsequencesAnalysis(docs: RawDocument[], lang: Language | string, _articleType: string): string {
-  const propCount = docs.filter(d => d.doktyp === 'prop').length;
-  const motCount = docs.filter(d => d.doktyp === 'mot').length;
-  const parts: string[] = [];
-
-  if (propCount > 0) {
-    parts.push(propConsequencesText(lang, propCount));
-  }
-  if (motCount > 0) {
-    parts.push(motConsequencesText(lang, motCount));
-  }
-  if (parts.length === 0) {
-    parts.push(genericConsequencesText(lang));
-  }
-  return parts.join(' ');
-}
-
-function propConsequencesText(_lang: Language | string, n: number): string {
-  return `<!-- AI_MUST_REPLACE: consequences — DATA: ${n} proposition(s) pending. Analyse the specific implementation consequences: which agencies must act, what regulatory changes are required, budget implications, and timeline for each proposition. Output MUST be in the article's language. -->`;
-}
-
-function motConsequencesText(_lang: Language | string, n: number): string {
-  return `<!-- AI_MUST_REPLACE: consequences — DATA: ${n} opposition motion(s) pending. Analyse the strategic consequences: how will rejection/acceptance affect party positioning, which motions establish new policy alternatives, and what campaign value do they carry. Output MUST be in the article's language. -->`;
-}
-
-function genericConsequencesText(_lang: Language | string): string {
-  return '<!-- AI_MUST_REPLACE: consequences — Write specific analysis of the consequences and next steps for these items, including committee timelines, expected vote dates, implementation requirements, and political ramifications. Output MUST be in the article\'s language. -->';
-}
-
 function generateCriticalAssessment(docs: RawDocument[], lang: Language | string, cia: CIAContext | undefined): string {
   const parts: string[] = [];
 
@@ -827,299 +379,149 @@ function generatePerspectivesAnalysis(docs: RawDocument[], lang: Language | stri
 }
 
 // ---------------------------------------------------------------------------
-// Framework analysis section renderers
+// Main deep analysis section generator
 // ---------------------------------------------------------------------------
 
-/** Max items per PESTLE dimension in aggregated display */
-const MAX_PESTLE_ITEMS = 4;
-/** Max stakeholder impacts shown in the summary list */
-const MAX_STAKEHOLDER_IMPACTS = 7;
-/** Max risk items shown in the risk assessment summary */
-const MAX_RISK_ITEMS = 5;
-/** Max perspective insights shown from 6-lens analysis */
-const MAX_PERSPECTIVE_INSIGHTS = 5;
-/** Max implementation obstacles listed */
-const MAX_IMPLEMENTATION_OBSTACLES = 4;
-/** Max agencies displayed in implementation assessment */
-const MAX_AGENCIES_DISPLAYED = 5;
-
-const PESTLE_LABELS: Readonly<Record<string, Record<keyof PESTLEAnalysis, string>>> = {
-  en: { political: 'Political', economic: 'Economic', social: 'Social', technological: 'Technological', legal: 'Legal', environmental: 'Environmental' },
-  sv: { political: 'Politisk', economic: 'Ekonomisk', social: 'Social', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljö' },
-  da: { political: 'Politisk', economic: 'Økonomisk', social: 'Social', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljø' },
-  no: { political: 'Politisk', economic: 'Økonomisk', social: 'Sosial', technological: 'Teknologisk', legal: 'Juridisk', environmental: 'Miljø' },
-  fi: { political: 'Poliittinen', economic: 'Taloudellinen', social: 'Sosiaalinen', technological: 'Teknologinen', legal: 'Oikeudellinen', environmental: 'Ympäristö' },
-  de: { political: 'Politisch', economic: 'Wirtschaftlich', social: 'Sozial', technological: 'Technologisch', legal: 'Rechtlich', environmental: 'Umwelt' },
-  fr: { political: 'Politique', economic: 'Économique', social: 'Social', technological: 'Technologique', legal: 'Juridique', environmental: 'Environnemental' },
-  es: { political: 'Político', economic: 'Económico', social: 'Social', technological: 'Tecnológico', legal: 'Jurídico', environmental: 'Ambiental' },
-  nl: { political: 'Politiek', economic: 'Economisch', social: 'Sociaal', technological: 'Technologisch', legal: 'Juridisch', environmental: 'Milieu' },
-  ar: { political: 'سياسي', economic: 'اقتصادي', social: 'اجتماعي', technological: 'تقني', legal: 'قانوني', environmental: 'بيئي' },
-  he: { political: 'פוליטי', economic: 'כלכלי', social: 'חברתי', technological: 'טכנולוגי', legal: 'משפטי', environmental: 'סביבתי' },
-  ja: { political: '政治', economic: '経済', social: '社会', technological: '技術', legal: '法的', environmental: '環境' },
-  ko: { political: '정치', economic: '경제', social: '사회', technological: '기술', legal: '법률', environmental: '환경' },
-  zh: { political: '政治', economic: '经济', social: '社会', technological: '技术', legal: '法律', environmental: '环境' },
-};
-
-const RISK_TYPE_LABELS: Readonly<Record<string, Record<RiskAssessment['type'], string>>> = {
-  en: { political: 'Political', implementation: 'Implementation', 'public-acceptance': 'Public acceptance', legal: 'Legal', financial: 'Financial' },
-  sv: { political: 'Politisk', implementation: 'Genomförande', 'public-acceptance': 'Offentlig acceptans', legal: 'Juridisk', financial: 'Finansiell' },
-  da: { political: 'Politisk', implementation: 'Implementering', 'public-acceptance': 'Offentlig accept', legal: 'Juridisk', financial: 'Finansiel' },
-  no: { political: 'Politisk', implementation: 'Implementering', 'public-acceptance': 'Offentlig aksept', legal: 'Juridisk', financial: 'Finansiell' },
-  fi: { political: 'Poliittinen', implementation: 'Toteutus', 'public-acceptance': 'Julkinen hyväksyntä', legal: 'Oikeudellinen', financial: 'Taloudellinen' },
-  de: { political: 'Politisch', implementation: 'Umsetzung', 'public-acceptance': 'Öffentliche Akzeptanz', legal: 'Rechtlich', financial: 'Finanziell' },
-  fr: { political: 'Politique', implementation: 'Mise en œuvre', 'public-acceptance': 'Acceptation publique', legal: 'Juridique', financial: 'Financier' },
-  es: { political: 'Político', implementation: 'Implementación', 'public-acceptance': 'Aceptación pública', legal: 'Jurídico', financial: 'Financiero' },
-  nl: { political: 'Politiek', implementation: 'Implementatie', 'public-acceptance': 'Publieke acceptatie', legal: 'Juridisch', financial: 'Financieel' },
-  ar: { political: 'سياسي', implementation: 'تنفيذي', 'public-acceptance': 'القبول العام', legal: 'قانوني', financial: 'مالي' },
-  he: { political: 'פוליטי', implementation: 'יישום', 'public-acceptance': 'קבלה ציבורית', legal: 'משפטי', financial: 'פיננסי' },
-  ja: { political: '政治', implementation: '実装', 'public-acceptance': '世論受容', legal: '法的', financial: '財政' },
-  ko: { political: '정치', implementation: '이행', 'public-acceptance': '대중 수용성', legal: '법률', financial: '재정' },
-  zh: { political: '政治', implementation: '实施', 'public-acceptance': '公众接受度', legal: '法律', financial: '财政' },
-};
-
-const LEVEL_LABELS: Readonly<Record<string, Record<'high' | 'medium' | 'low', string>>> = {
-  en: { high: 'High', medium: 'Medium', low: 'Low' },
-  sv: { high: 'Hög', medium: 'Medel', low: 'Låg' },
-  da: { high: 'Høj', medium: 'Mellem', low: 'Lav' },
-  no: { high: 'Høy', medium: 'Middels', low: 'Lav' },
-  fi: { high: 'Korkea', medium: 'Keskitaso', low: 'Matala' },
-  de: { high: 'Hoch', medium: 'Mittel', low: 'Niedrig' },
-  fr: { high: 'Élevé', medium: 'Moyen', low: 'Faible' },
-  es: { high: 'Alto', medium: 'Medio', low: 'Bajo' },
-  nl: { high: 'Hoog', medium: 'Middel', low: 'Laag' },
-  ar: { high: 'مرتفع', medium: 'متوسط', low: 'منخفض' },
-  he: { high: 'גבוה', medium: 'בינוני', low: 'נמוך' },
-  ja: { high: '高', medium: '中', low: '低' },
-  ko: { high: '높음', medium: '보통', low: '낮음' },
-  zh: { high: '高', medium: '中', low: '低' },
-};
-
-const IMPLEMENTATION_LABELS: Readonly<Record<string, { feasibility: string; obstacles: string; agencies: string; noStakeholderData: string; noImplementationData: string; burden: string }>> = {
-  en: { feasibility: 'Feasibility', obstacles: 'Key obstacles', agencies: 'Agencies involved', noStakeholderData: 'No stakeholder impact data available.', noImplementationData: 'No implementation data available.', burden: 'Burden' },
-  sv: { feasibility: 'Genomförbarhet', obstacles: 'Viktiga hinder', agencies: 'Berörda myndigheter', noStakeholderData: 'Ingen data om intressentpåverkan tillgänglig.', noImplementationData: 'Ingen implementeringsdata tillgänglig.', burden: 'Belastning' },
-  da: { feasibility: 'Gennemførlighed', obstacles: 'Vigtige hindringer', agencies: 'Involverede myndigheder', noStakeholderData: 'Ingen data om interessentpåvirkning tilgængelig.', noImplementationData: 'Ingen implementeringsdata tilgængelig.', burden: 'Byrde' },
-  no: { feasibility: 'Gjennomførbarhet', obstacles: 'Viktige hindringer', agencies: 'Involverte etater', noStakeholderData: 'Ingen data om interessentpåvirkning tilgjengelig.', noImplementationData: 'Ingen implementeringsdata tilgjengelig.', burden: 'Belastning' },
-  fi: { feasibility: 'Toteutettavuus', obstacles: 'Keskeiset esteet', agencies: 'Mukana olevat viranomaiset', noStakeholderData: 'Sidosryhmävaikutustietoa ei saatavilla.', noImplementationData: 'Toteutustietoa ei saatavilla.', burden: 'Rasite' },
-  de: { feasibility: 'Umsetzbarkeit', obstacles: 'Wesentliche Hindernisse', agencies: 'Beteiligte Behörden', noStakeholderData: 'Keine Daten zu Stakeholder-Auswirkungen verfügbar.', noImplementationData: 'Keine Umsetzungsdaten verfügbar.', burden: 'Belastung' },
-  fr: { feasibility: 'Faisabilité', obstacles: 'Obstacles clés', agencies: 'Agences impliquées', noStakeholderData: 'Aucune donnée d’impact des parties prenantes disponible.', noImplementationData: 'Aucune donnée de mise en œuvre disponible.', burden: 'Charge' },
-  es: { feasibility: 'Viabilidad', obstacles: 'Obstáculos clave', agencies: 'Organismos implicados', noStakeholderData: 'No hay datos de impacto en partes interesadas.', noImplementationData: 'No hay datos de implementación disponibles.', burden: 'Carga' },
-  nl: { feasibility: 'Haalbaarheid', obstacles: 'Belangrijkste obstakels', agencies: 'Betrokken instanties', noStakeholderData: 'Geen gegevens over impact op belanghebbenden beschikbaar.', noImplementationData: 'Geen implementatiegegevens beschikbaar.', burden: 'Last' },
-  ar: { feasibility: 'قابلية التنفيذ', obstacles: 'العقبات الرئيسية', agencies: 'الجهات المعنية', noStakeholderData: 'لا تتوفر بيانات تأثير أصحاب المصلحة.', noImplementationData: 'لا تتوفر بيانات تنفيذ.', burden: 'العبء' },
-  he: { feasibility: 'ישימות', obstacles: 'חסמים מרכזיים', agencies: 'גורמים מעורבים', noStakeholderData: 'אין נתוני השפעה על בעלי עניין.', noImplementationData: 'אין נתוני יישום.', burden: 'נטל' },
-  ja: { feasibility: '実現可能性', obstacles: '主な障害', agencies: '関係機関', noStakeholderData: 'ステークホルダー影響データはありません。', noImplementationData: '実施データはありません。', burden: '負担' },
-  ko: { feasibility: '실행 가능성', obstacles: '주요 장애 요인', agencies: '관여 기관', noStakeholderData: '이해관계자 영향 데이터가 없습니다.', noImplementationData: '이행 데이터가 없습니다.', burden: '부담' },
-  zh: { feasibility: '可实施性', obstacles: '关键障碍', agencies: '涉及机构', noStakeholderData: '暂无利益相关方影响数据。', noImplementationData: '暂无实施数据。', burden: '负担' },
-};
-
-function localizeLevel(level: 'high' | 'medium' | 'low', lang: Language | string): string {
-  return LEVEL_LABELS[lang as string]?.[level] ?? LEVEL_LABELS.en[level];
-}
-
-function localizeRiskType(type: RiskAssessment['type'], lang: Language | string): string {
-  return RISK_TYPE_LABELS[lang as string]?.[type] ?? RISK_TYPE_LABELS.en[type];
-}
-
-function localizedImplementationLabels(lang: Language | string): { feasibility: string; obstacles: string; agencies: string; noStakeholderData: string; noImplementationData: string; burden: string } {
-  return IMPLEMENTATION_LABELS[lang as string] ?? IMPLEMENTATION_LABELS.en;
-}
-
 /**
- * Aggregate PESTLE dimensions across multiple document analyses into a
- * deduplicated list per dimension and render as an HTML description list.
+ * Generate a comprehensive Deep Analysis section following the 5W framework
+ * (Who, What, When, Why, Winners/Losers) plus impact, consequences, and critical
+ * assessment subsections. This section is designed for highly analytical readers
+ * who seek multi-perspective intelligence on parliamentary developments.
+ *
+ * @returns HTML string for the deep analysis section, or empty string if insufficient data
  */
-function renderAggregatedPestle(analyses: DocumentAnalysis[], lang: Language | string): string {
-  const merged: PESTLEAnalysis = {
-    political: [], economic: [], social: [],
-    technological: [], legal: [], environmental: [],
-  };
+export function generateDeepAnalysisSection(opts: DeepAnalysisOptions): string {
+  const { documents, lang, cia, articleType, whyContext, frameworkAnalysis, perspectiveAnalysis } = opts;
 
-  for (const a of analyses) {
-    const p = a.pestleDimensions;
-    merged.political.push(...p.political);
-    merged.economic.push(...p.economic);
-    merged.social.push(...p.social);
-    merged.technological.push(...p.technological);
-    merged.legal.push(...p.legal);
-    merged.environmental.push(...p.environmental);
-  }
+  // Deep analysis requires at least 2 documents for cross-document insights
+  // in standard article types. For deep-inspection articles, allow single-
+  // document analysis since the whole article is dedicated to in-depth review.
+  const minDocs = articleType === 'deep-inspection' ? 1 : 2;
+  if (!documents || documents.length < minDocs) return '';
 
-  // Deduplicate per dimension
-  const dedup = (arr: string[]): string[] => [...new Set(arr)].slice(0, MAX_PESTLE_ITEMS);
-
-  const labels = PESTLE_LABELS[lang as string] ?? PESTLE_LABELS.en;
-  const dims: Array<[string, string[]]> = [
-    [labels.political, dedup(merged.political)],
-    [labels.economic, dedup(merged.economic)],
-    [labels.social, dedup(merged.social)],
-    [labels.technological, dedup(merged.technological)],
-    [labels.legal, dedup(merged.legal)],
-    [labels.environmental, dedup(merged.environmental)],
-  ];
-
-  const items = dims
-    .filter(([, items]) => items.length > 0)
-    .map(([label, items]) =>
-      `      <dt><strong>${escapeHtml(label)}</strong></dt>\n      <dd>${items.map(i => escapeHtml(i)).join(' ')}</dd>`,
-    )
-    .join('\n');
-
-  return `    <dl class="pestle-analysis">\n${items}\n    </dl>`;
-}
+  const lbl = (key: string): string => {
+    const val = L(lang, key);
+    return typeof val === 'string' ? val : key;
+  };
 
-/**
- * Render a summary of stakeholder impacts across all analysed documents.
- * Shows up to 7 stakeholder groups with impact direction, confidence, and burden.
- */
-function renderStakeholderImpactSummary(analyses: DocumentAnalysis[], lang: Language | string): string {
-  const labels = localizedImplementationLabels(lang);
-  // Collect all stakeholder impacts, deduplicated by stakeholder name
-  const impactMap = new Map<string, StakeholderImpact>();
-  for (const a of analyses) {
-    for (const impact of a.stakeholderImpacts) {
-      // Keep the higher-magnitude impact per stakeholder
-      const existing = impactMap.get(impact.stakeholder);
-      if (!existing || magnitudeRank(impact.directImpact.magnitude) > magnitudeRank(existing.directImpact.magnitude)) {
-        impactMap.set(impact.stakeholder, impact);
-      }
+  const parts: string[] = [];
+  parts.push(`\n    <section class="deep-analysis" aria-label="${escapeHtml(lbl('deepAnalysis'))}">`);
+  parts.push(`    <h2>${escapeHtml(lbl('deepAnalysis'))}</h2>`);
+
+  // ── WHO: Key Actors ────────────────────────────────────────────────────────
+  const { parties, authors } = extractKeyActors(documents);
+  if (parties.size > 0 || authors.length > 0) {
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWho'))}</h3>`);
+    if (parties.size > 0) {
+      const sortedParties = [...parties.entries()].sort((a, b) => b[1] - a[1]);
+      const partyList = sortedParties
+        .map(([p, count]) => `<strong>${escapeHtml(p)}</strong> (${count})`)
+        .join(', ');
+      parts.push(`    <p>${partyList}</p>`);
+    }
+    if (authors.length > 0) {
+      parts.push(`    <p>${authors.map(a => escapeHtml(a)).join(', ')}</p>`);
     }
   }
 
-  const impacts = [...impactMap.values()].slice(0, MAX_STAKEHOLDER_IMPACTS);
-  if (impacts.length === 0) return `    <p>${escapeHtml(labels.noStakeholderData)}</p>`;
-
-  const rows = impacts.map(i => {
-    const directionIcon =
-      i.directImpact.direction === 'positive' ? '↑'
-      : i.directImpact.direction === 'negative' ? '↓'
-      : i.directImpact.direction === 'mixed' ? '↕'
-      : '→';
-    const burdenText = localizeLevel(i.implementationBurden, lang);
-    return `      <li><strong>${escapeHtml(i.displayName)}</strong>: ${directionIcon} ${escapeHtml(i.directImpact.summary)} (${escapeHtml(i.confidence)}; ${escapeHtml(labels.burden)}: ${escapeHtml(burdenText)})</li>`;
-  });
-
-  return `    <ul class="stakeholder-impact-list">\n${rows.join('\n')}\n    </ul>`;
-}
-
-/**
- * Render a risk assessment summary. Groups risks by type and keeps the
- * highest-severity risk per type.
- */
-function renderRiskAssessment(risks: RiskAssessment[], lang: Language | string): string {
-  // Deduplicate by type, preferring higher severity
-  const byType = new Map<string, RiskAssessment>();
-  for (const r of risks) {
-    const key = r.type;
-    const existing = byType.get(key);
-    if (!existing || severityRank(r.severity) > severityRank(existing.severity)) {
-      byType.set(key, r);
+  // ── WHAT: What Happened ────────────────────────────────────────────────────
+  const domains = aggregateDomains(documents, lang);
+  if (domains.size > 0) {
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhat'))}</h3>`);
+    const sortedDomains = [...domains.entries()].sort((a, b) => b[1] - a[1]);
+    const domainItems = sortedDomains.slice(0, 6)
+      .map(([d, c]) => `${escapeHtml(d)} (${c})`)
+      .join(', ');
+    const docTypes = new Map<string, number>();
+    for (const doc of documents) {
+      const t = doc.doktyp || doc.documentType || 'other';
+      docTypes.set(t, (docTypes.get(t) ?? 0) + 1);
     }
+    const typeList = [...docTypes.entries()]
+      .sort((a, b) => b[1] - a[1])
+      .map(([t, c]) => `${escapeHtml(localizeDocType(t, lang, c))}: ${c}`)
+      .join(', ');
+    parts.push(`    <p>${domainItems}</p>`);
+    parts.push(`    <p><em>${typeList}</em></p>`);
   }
 
-  const top = [...byType.values()].slice(0, MAX_RISK_ITEMS);
-  const rows = top.map(r => {
-    const icon = r.severity === 'high' ? '🔴' : r.severity === 'medium' ? '🟡' : '🟢';
-    return `      <li>${icon} <strong>${escapeHtml(localizeRiskType(r.type, lang))}</strong> (${escapeHtml(localizeLevel(r.severity, lang))}): ${escapeHtml(r.description)}</li>`;
-  });
+  // ── WHEN: Timeline & Context ───────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhen'))}</h3>`);
+  const timelineContext = generateTimelineContext(documents, lang, articleType);
+  parts.push(`    <p>${timelineContext}</p>`);
 
-  return `    <ul class="risk-assessment-list">\n${rows.join('\n')}\n    </ul>`;
-}
+  // ── WHY: Why This Matters ──────────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWhy'))}</h3>`);
+  const whyText = generateWhyAnalysis(documents, lang, cia, whyContext);
+  parts.push(`    <p>${whyText}</p>`);
 
-function severityRank(s: string): number {
-  return s === 'high' ? 3 : s === 'medium' ? 2 : 1;
-}
+  // ── WINNERS & LOSERS ───────────────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisWinners'))}</h3>`);
+  const winnersText = analyseWinnersLosers(documents, cia, lang);
+  parts.push(`    <p>${winnersText}</p>`);
 
-function magnitudeRank(magnitude: 'significant' | 'moderate' | 'minor'): number {
-  return magnitude === 'significant' ? 3 : magnitude === 'moderate' ? 2 : 1;
-}
+  // ── POLITICAL IMPACT ───────────────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisImpact'))}</h3>`);
+  const impactText = generateImpactAnalysis(documents, lang, cia);
+  parts.push(`    <p>${impactText}</p>`);
 
-/**
- * Render implementation assessment summary from framework analyses.
- */
-function renderImplementationAssessment(analyses: DocumentAnalysis[], lang: Language | string): string {
-  const labels = localizedImplementationLabels(lang);
-  const assessments: ImplementationAssessment[] = analyses.map(a => a.implementationAssessment);
-  if (assessments.length === 0) return `    <p>${escapeHtml(labels.noImplementationData)}</p>`;
-
-  // Aggregate obstacles and agencies across all documents
-  const allObstacles = new Set<string>();
-  const allAgencies = new Set<string>();
-  let highestFeasibility: ImplementationAssessment['feasibility'] = 'high';
-  let selectedAssessment: ImplementationAssessment = assessments[0];
-
-  for (const ia of assessments) {
-    ia.keyObstacles.forEach(o => allObstacles.add(o));
-    ia.agenciesInvolved.forEach(a => allAgencies.add(a));
-    if (feasibilityRank(ia.feasibility) < feasibilityRank(highestFeasibility)) {
-      highestFeasibility = ia.feasibility;
-      selectedAssessment = ia;
-    }
-  }
+  // ── ACTIONS & CONSEQUENCES ─────────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisConsequences'))}</h3>`);
+  const consequencesText = generateConsequencesAnalysis(documents, lang, articleType);
+  parts.push(`    <p>${consequencesText}</p>`);
 
-  const parts: string[] = [];
-  const fIcon = highestFeasibility === 'high' ? '🟢' : highestFeasibility === 'medium' ? '🟡' : '🔴';
-  const timeline = selectedAssessment.estimatedTimeline;
-  parts.push(`    <p>${fIcon} <strong>${escapeHtml(labels.feasibility)}:</strong> ${escapeHtml(localizeLevel(highestFeasibility, lang))}. ${escapeHtml(timeline)}</p>`);
+  // ── CRITICAL ASSESSMENT ────────────────────────────────────────────────────
+  parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisCritical'))}</h3>`);
+  const criticalText = generateCriticalAssessment(documents, lang, cia);
+  parts.push(`    <p>${criticalText}</p>`);
 
-  if (allObstacles.size > 0) {
-    const obstacleList = [...allObstacles].slice(0, MAX_IMPLEMENTATION_OBSTACLES).map(o => `<li>${escapeHtml(o)}</li>`).join('');
-    parts.push(`    <p><strong>${escapeHtml(labels.obstacles)}:</strong></p>\n    <ul>${obstacleList}</ul>`);
+  // ── MULTIPLE PERSPECTIVES ──────────────────────────────────────────────────
+  if (parties.size >= 2) {
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisPerspectives'))}</h3>`);
+    const perspectivesText = generatePerspectivesAnalysis(documents, lang, parties);
+    parts.push(`    <p>${perspectivesText}</p>`);
   }
 
-  if (allAgencies.size > 0) {
-    parts.push(`    <p><strong>${escapeHtml(labels.agencies)}:</strong> ${[...allAgencies].slice(0, MAX_AGENCIES_DISPLAYED).map(a => escapeHtml(a)).join(', ')}</p>`);
-  }
+  // ── FRAMEWORK ANALYSIS SECTIONS ────────────────────────────────────────────
+  // When the document analysis framework has been run, inject its richer
+  // PESTLE, stakeholder impact, risk, and implementation assessment data.
+  if (frameworkAnalysis && frameworkAnalysis.size > 0) {
+    const analyses = [...frameworkAnalysis.values()];
 
-  return parts.join('\n');
-}
+    // PESTLE Analysis — aggregate across all analysed documents
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisPestle'))}</h3>`);
+    parts.push(renderAggregatedPestle(analyses, lang));
 
-function feasibilityRank(f: string): number {
-  return f === 'high' ? 3 : f === 'medium' ? 2 : 1;
-}
+    // Stakeholder Impact — summarise stakeholder impacts from the framework
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisStakeholderImpact'))}</h3>`);
+    parts.push(renderStakeholderImpactSummary(analyses, lang));
 
-/* ── Banned pattern detection ─────────────────────────────────────────────── */
+    // Risk Assessment — aggregate risk factors across documents
+    const allRisks = analyses.flatMap(a => a.riskAssessment as RiskAssessment[]);
+    if (allRisks.length > 0) {
+      parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisRisk'))}</h3>`);
+      parts.push(renderRiskAssessment(allRisks, lang));
+    }
 
-/**
- * Banned content patterns that indicate low-quality boilerplate text.
- * Per SHARED_PROMPT_PATTERNS.md §BANNED Content Patterns v4.0, these
- * must never appear in production articles. AI agents MUST replace them
- * with genuine, document-specific analysis.
- */
-const BANNED_PATTERNS: readonly { label: string; pattern: RegExp }[] = [
-  { label: 'neutralText: "The political landscape remains fluid…"', pattern: /The political landscape remains fluid,? with both government and opposition positioning for advantage/i },
-  { label: 'debateAnalysisMarker: "No chamber debate data is available…"', pattern: /No chamber debate data is available for these items,? limiting our ability/i },
-  { label: 'policySignificanceTouches: "Touches on {domains}."', pattern: /Touches on [\p{L}\p{N}][\p{L}\p{N}\s,&/()-]*\./iu },
-  { label: 'analysisOfNDocuments: "Analysis of N documents covering…"', pattern: /Analysis of \d+ documents covering/i },
-  { label: 'policySignificanceGeneric: "Requires committee review and chamber debate…"', pattern: /Requires committee review and chamber debate/i },
-  { label: 'topicInFocusSuffix: "…: {Topic} in Focus"', pattern: /:\s+\w[\w\s]*\bin Focus\b/i },
-  { label: 'briefingOnFieldLabels: "Political intelligence briefing on {Field}: and {Field}:"', pattern: /Political intelligence briefing on \w+:\s+and\s+\w+:/i },
-  // Deep Analysis generic template patterns — AI MUST replace these with specific analysis
-  { label: 'genericTimeline: "The pace of activity signals…"', pattern: /The pace of activity signals the political urgency/i },
-  { label: 'genericTimeline: "define the current legislative landscape"', pattern: /define the current legislative landscape/i },
-  { label: 'genericWhy: "broad legislative push that will shape"', pattern: /broad legislative push that will shape multiple aspects/i },
-  { label: 'genericWhy: "critical period for understanding the government"', pattern: /critical period for understanding the government.s strategic direction/i },
-  { label: 'genericImpact: "culmination of legislative review, with recommendations that guide"', pattern: /culmination of legislative review,? with recommendations that guide/i },
-  { label: 'genericImpact: "interplay between governing ambition and opposition scrutiny"', pattern: /interplay between governing ambition and opposition scrutiny/i },
-  { label: 'genericConsequences: "cascade through committee deliberations"', pattern: /cascade through committee deliberations,? chamber votes/i },
-  { label: 'genericConsequences: "establish the policy alternatives that opposition parties will champion"', pattern: /establish the policy alternatives that opposition parties will champion/i },
-  { label: 'genericCritical: "Standard parliamentary procedures are being followed"', pattern: /Standard parliamentary procedures are being followed/i },
-  { label: 'genericCritical: "gap between legislative intent and implementation"', pattern: /gap between legislative intent and implementation often reveals/i },
-  { label: 'genericPillarTransition: "While parliament deliberates these legislative matters"', pattern: /While parliament deliberates these legislative matters/i },
-];
+    // Implementation Assessment — summarise implementation feasibility
+    parts.push(`    <h3>${escapeHtml(lbl('deepAnalysisImplementation'))}</h3>`);
+    parts.push(renderImplementationAssessment(analyses, lang));
+  }
 
-/**
- * Detect banned boilerplate patterns in HTML content.
- * Returns an array of human-readable labels identifying each detected
- * banned pattern, suitable for quality gate logs and error messages.
- *
- * @param html - The HTML string to scan for banned patterns
- * @returns Array of stable human-readable labels for each detected banned pattern
- */
-export function detectBannedPatterns(html: string): string[] {
-  const found: string[] = [];
-  for (const { label, pattern } of BANNED_PATTERNS) {
-    if (pattern.test(html)) {
-      found.push(label);
+  // ── MULTI-PERSPECTIVE INSIGHTS (6 lenses) ────────────────────────────────
+  // When the analysis-framework has been run, inject key insights from the
+  // government, opposition, citizen, economic, international, and media lenses.
+  if (perspectiveAnalysis && perspectiveAnalysis.results.length > 0) {
+    const allInsights = perspectiveAnalysis.results.flatMap((r: unknown) => ((r as { keyInsights?: string[] }).keyInsights ?? []));
+    if (allInsights.length > 0) {
+      const uniqueInsights = [...new Set(allInsights)].slice(0, MAX_PERSPECTIVE_INSIGHTS);
+      parts.push(`    <div class="perspective-insights">`);
+      const insightItems = uniqueInsights.map(i => `      <li>${escapeHtml(i)}</li>`).join('\n');
+      parts.push(`    <ul>\n${insightItems}\n    </ul>`);
+      parts.push(`    </div>`);
     }
   }
-  return found;
+
+  parts.push('    </section>\n');
+  return parts.join('\n');
 }
diff --git a/scripts/generate-news-enhanced/generators.ts b/scripts/generate-news-enhanced/generators.ts
index d29a81365..347e5e0fc 100644
--- a/scripts/generate-news-enhanced/generators.ts
+++ b/scripts/generate-news-enhanced/generators.ts
@@ -662,143 +662,18 @@ export async function generateInterpellations(): Promise<GenerationResult> {
   }
 }
 
-/**
- * Extract a Riksdag document ID (dok_id) from a known URL pattern.
- * Supports:
- *   - https://riksdagen.se/sv/dokument-och-lagar/dokument/{type}/{dok_id}/
- *   - https://data.riksdagen.se/dokument/{dok_id}
- *   - https://data.riksdagen.se/dokument/{dok_id}.json
- *
- * @returns The extracted dok_id, or null if the URL doesn't match a known pattern.
- */
-export function extractDocIdFromUrl(url: string): string | null {
-  try {
-    const parsed = new URL(url);
-    const hostname = parsed.hostname.toLowerCase();
-    const segments = parsed.pathname.split('/').filter(Boolean);
-
-    // https://riksdagen.se/sv/dokument-och-lagar/dokument/{type}/{dok_id}
-    if (hostname === 'riksdagen.se' || hostname === 'www.riksdagen.se') {
-      // Path: /sv/dokument-och-lagar/dokument/{type}/{dok_id}
-      const dokIdx = segments.indexOf('dokument');
-      if (dokIdx >= 0 && segments.length > dokIdx + 2) {
-        return segments[dokIdx + 2];
-      }
-    }
-
-    // https://data.riksdagen.se/dokument/{dok_id}[.json|.xml|.html]
-    if (hostname === 'data.riksdagen.se') {
-      const dokIdx = segments.indexOf('dokument');
-      if (dokIdx >= 0 && segments.length > dokIdx + 1) {
-        return segments[dokIdx + 1].replace(/\.(json|xml|html|pdf)$/i, ''); // strip known file extensions
-      }
-    }
-
-    return null;
-  } catch {
-    return null;
-  }
-}
-
-/**
- * Determine whether a URL points to a government (regeringen.se) resource
- * that can be fetched via the get_g0v_document_content MCP tool.
- */
-export function isGovernmentUrl(url: string): boolean {
-  try {
-    const parsed = new URL(url);
-    const hostname = parsed.hostname.toLowerCase();
-    return hostname === 'regeringen.se' || hostname === 'www.regeringen.se';
-  } catch {
-    return false;
-  }
-}
-
-/**
- * Determine whether a URL points to a GitHub repository resource
- * (github.com or raw.githubusercontent.com) that can be fetched as raw content.
- */
-export function isGitHubUrl(url: string): boolean {
-  try {
-    const parsed = new URL(url);
-    const hostname = parsed.hostname.toLowerCase();
-    return hostname === 'github.com'
-      || hostname === 'www.github.com'
-      || hostname === 'raw.githubusercontent.com';
-  } catch {
-    return false;
-  }
-}
-
-/**
- * Convert a GitHub blob/tree URL to a raw.githubusercontent.com URL.
- * Handles patterns like:
- *   - https://github.com/{owner}/{repo}/blob/{branch}/{path}
- *   - https://github.com/{owner}/{repo}/raw/{branch}/{path}
- *   - https://raw.githubusercontent.com/{owner}/{repo}/{branch}/{path} (returned as-is)
- *
- * @returns The raw URL, or null if the URL cannot be converted.
- */
-export function toGitHubRawUrl(url: string): string | null {
-  try {
-    const parsed = new URL(url);
-    const hostname = parsed.hostname.toLowerCase();
-
-    // Already a raw URL — return as-is
-    if (hostname === 'raw.githubusercontent.com') {
-      return url;
-    }
-
-    if (hostname !== 'github.com' && hostname !== 'www.github.com') {
-      return null;
-    }
-
-    // Path: /{owner}/{repo}/blob/{branch}/{...path}
-    // or:   /{owner}/{repo}/raw/{branch}/{...path}
-    const segments = parsed.pathname.split('/').filter(Boolean);
-    if (segments.length < 4) return null;
-
-    const [owner, repo, refType, ...rest] = segments;
-    if (refType !== 'blob' && refType !== 'raw') return null;
-
-    // rest = [branch, ...pathParts]
-    return `https://raw.githubusercontent.com/${owner}/${repo}/${rest.join('/')}`;
-  } catch {
-    return null;
-  }
-}
-
-/**
- * Compute a short, deterministic hash suffix from a URL path string.
- * Used to generate collision-resistant `dok_id` values for documents
- * fetched from government or GitHub URLs.
- *
- * The hash is a simple DJB2-style left-shift-and-add over each character,
- * rendered in base-36.  A leading `-` (from negative ints) is replaced with `n`.
- */
-export function hashPathSuffix(path: string): string {
-  return path
-    .split('')
-    .reduce((a, c) => ((a << 5) - a + c.charCodeAt(0)) | 0, 0)
-    .toString(36)
-    .replace(/^-/, 'n');
-}
-
-/**
- * Strip HTML tags from a user-supplied string to prevent XSS.
- * Uses a multi-pass loop to handle nested tag reconstruction attempts
- * (e.g. `<scr<script>ipt>`).  Returns **plain text** — callers must
- * apply `escapeHtml()` at their render sites so escaping happens exactly once.
- */
-export function sanitizePlainText(text: string): string {
-  let cleaned = text;
-  let prev: string;
-  do {
-    prev = cleaned;
-    cleaned = cleaned.replace(/<[^>]*>/g, '');
-  } while (cleaned !== prev);
-  return cleaned;
-}
+// ---------------------------------------------------------------------------
+// URL & text utilities — implementation extracted to url-utils.ts
+// Re-exported here for backward compatibility.
+// ---------------------------------------------------------------------------
+export {
+  extractDocIdFromUrl,
+  isGovernmentUrl,
+  isGitHubUrl,
+  toGitHubRawUrl,
+  hashPathSuffix,
+  sanitizePlainText,
+} from './url-utils.js';
 
 // ---------------------------------------------------------------------------
 // Deep-Inspection content generator (topic-focused, comprehensive)
diff --git a/scripts/generate-news-enhanced/url-utils.ts b/scripts/generate-news-enhanced/url-utils.ts
new file mode 100644
index 000000000..f864fecf0
--- /dev/null
+++ b/scripts/generate-news-enhanced/url-utils.ts
@@ -0,0 +1,145 @@
+/**
+ * @module generate-news-enhanced/url-utils
+ * @description URL parsing and text sanitization utilities for the
+ * deep-inspection article generator. Provides Riksdag/government URL
+ * extraction, GitHub raw URL conversion, and XSS-safe text cleaning.
+ *
+ * @author Hack23 AB
+ * @license Apache-2.0
+ */
+
+/**
+ * Extract a `dok_id` from a Riksdag or data.riksdagen.se document URL.
+ * Returns `null` if the URL is not a recognised Riksdag document URL.
+ *
+ * Supported patterns:
+ * - `https://riksdagen.se/sv/dokument-och-lagar/dokument/{type}/{dok_id}`
+ * - `https://data.riksdagen.se/dokument/{dok_id}[.json|.xml|.html]`
+ */
+export function extractDocIdFromUrl(url: string): string | null {
+  try {
+    const parsed = new URL(url);
+    const hostname = parsed.hostname.toLowerCase();
+    const segments = parsed.pathname.split('/').filter(Boolean);
+
+    // https://riksdagen.se/sv/dokument-och-lagar/dokument/{type}/{dok_id}
+    if (hostname === 'riksdagen.se' || hostname === 'www.riksdagen.se') {
+      const dokIdx = segments.indexOf('dokument');
+      if (dokIdx >= 0 && segments.length > dokIdx + 2) {
+        return segments[dokIdx + 2];
+      }
+    }
+
+    // https://data.riksdagen.se/dokument/{dok_id}[.json|.xml|.html]
+    if (hostname === 'data.riksdagen.se') {
+      const dokIdx = segments.indexOf('dokument');
+      if (dokIdx >= 0 && segments.length > dokIdx + 1) {
+        return segments[dokIdx + 1].replace(/\.(json|xml|html|pdf)$/i, ''); // strip known file extensions
+      }
+    }
+
+    return null;
+  } catch {
+    return null;
+  }
+}
+
+/**
+ * Determine whether a URL points to a government (regeringen.se) resource
+ * that can be fetched via the get_g0v_document_content MCP tool.
+ */
+export function isGovernmentUrl(url: string): boolean {
+  try {
+    const parsed = new URL(url);
+    const hostname = parsed.hostname.toLowerCase();
+    return hostname === 'regeringen.se' || hostname === 'www.regeringen.se';
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Determine whether a URL points to a GitHub repository resource
+ * (github.com or raw.githubusercontent.com) that can be fetched as raw content.
+ */
+export function isGitHubUrl(url: string): boolean {
+  try {
+    const parsed = new URL(url);
+    const hostname = parsed.hostname.toLowerCase();
+    return hostname === 'github.com'
+      || hostname === 'www.github.com'
+      || hostname === 'raw.githubusercontent.com';
+  } catch {
+    return false;
+  }
+}
+
+/**
+ * Convert a GitHub blob/tree URL to a raw.githubusercontent.com URL.
+ * Handles patterns like:
+ *   - https://github.com/{owner}/{repo}/blob/{branch}/{path}
+ *   - https://github.com/{owner}/{repo}/raw/{branch}/{path}
+ *   - https://raw.githubusercontent.com/{owner}/{repo}/{branch}/{path} (returned as-is)
+ *
+ * @returns The raw URL, or null if the URL cannot be converted.
+ */
+export function toGitHubRawUrl(url: string): string | null {
+  try {
+    const parsed = new URL(url);
+    const hostname = parsed.hostname.toLowerCase();
+
+    // Already a raw URL — return as-is
+    if (hostname === 'raw.githubusercontent.com') {
+      return url;
+    }
+
+    if (hostname !== 'github.com' && hostname !== 'www.github.com') {
+      return null;
+    }
+
+    // Path: /{owner}/{repo}/blob/{branch}/{...path}
+    // or:   /{owner}/{repo}/raw/{branch}/{...path}
+    const segments = parsed.pathname.split('/').filter(Boolean);
+    if (segments.length < 4) return null;
+
+    const [owner, repo, refType, ...rest] = segments;
+    if (refType !== 'blob' && refType !== 'raw') return null;
+
+    // rest = [branch, ...pathParts]
+    return `https://raw.githubusercontent.com/${owner}/${repo}/${rest.join('/')}`;
+  } catch {
+    return null;
+  }
+}
+
+/**
+ * Compute a short, deterministic hash suffix from a URL path string.
+ * Used to generate collision-resistant `dok_id` values for documents
+ * fetched from government or GitHub URLs.
+ *
+ * The hash is a simple DJB2-style left-shift-and-add over each character,
+ * rendered in base-36. A leading `-` (from negative ints) is replaced with `n`.
+ */
+export function hashPathSuffix(path: string): string {
+  return path
+    .split('')
+    .reduce((a, c) => ((a << 5) - a + c.charCodeAt(0)) | 0, 0)
+    .toString(36)
+    .replace(/^-/, 'n');
+}
+
+/**
+ * Strip HTML tags from a user-supplied string to prevent XSS.
+ * Uses a multi-pass loop to handle nested tag reconstruction attempts
+ * (e.g. `<scr<script>ipt>`). Returns **plain text** — callers must
+ * apply `escapeHtml()` at their render sites so escaping happens exactly once.
+ */
+export function sanitizePlainText(text: string): string {
+  let cleaned = text;
+  let prev: string;
+  do {
+    prev = cleaned;
+    cleaned = cleaned.replace(/<[^>]*>/g, '');
+  } while (cleaned !== prev);
+  return cleaned;
+}

From 694112801f662bf358cb78b442d2aa9a30423c4e Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Fri, 10 Apr 2026 09:37:53 +0000
Subject: [PATCH 3/5] fix: use double quotes for French string containing
 apostrophe in framework-renderers.ts

Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
---
 .../data-transformers/content-generators/framework-renderers.ts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/scripts/data-transformers/content-generators/framework-renderers.ts b/scripts/data-transformers/content-generators/framework-renderers.ts
index 747b17ed0..00e835ed6 100644
--- a/scripts/data-transformers/content-generators/framework-renderers.ts
+++ b/scripts/data-transformers/content-generators/framework-renderers.ts
@@ -165,7 +165,7 @@ const IMPLEMENTATION_LABELS: Readonly<Record<string, { feasibility: string; obst
   no: { feasibility: 'Gjennomførbarhet', obstacles: 'Viktige hindringer', agencies: 'Involverte etater', noStakeholderData: 'Ingen data om interessentpåvirkning tilgjengelig.', noImplementationData: 'Ingen implementeringsdata tilgjengelig.', burden: 'Belastning' },
   fi: { feasibility: 'Toteutettavuus', obstacles: 'Keskeiset esteet', agencies: 'Mukana olevat viranomaiset', noStakeholderData: 'Sidosryhmävaikutustietoa ei saatavilla.', noImplementationData: 'Toteutustietoa ei saatavilla.', burden: 'Rasite' },
   de: { feasibility: 'Umsetzbarkeit', obstacles: 'Wesentliche Hindernisse', agencies: 'Beteiligte Behörden', noStakeholderData: 'Keine Daten zu Stakeholder-Auswirkungen verfügbar.', noImplementationData: 'Keine Umsetzungsdaten verfügbar.', burden: 'Belastung' },
-  fr: { feasibility: 'Faisabilité', obstacles: 'Obstacles clés', agencies: 'Agences impliquées', noStakeholderData: 'Aucune donnée d\'impact des parties prenantes disponible.', noImplementationData: 'Aucune donnée de mise en œuvre disponible.', burden: 'Charge' },
+  fr: { feasibility: 'Faisabilité', obstacles: 'Obstacles clés', agencies: 'Agences impliquées', noStakeholderData: "Aucune donnée d'impact des parties prenantes disponible.", noImplementationData: 'Aucune donnée de mise en œuvre disponible.', burden: 'Charge' },
   es: { feasibility: 'Viabilidad', obstacles: 'Obstáculos clave', agencies: 'Organismos implicados', noStakeholderData: 'No hay datos de impacto en partes interesadas.', noImplementationData: 'No hay datos de implementación disponibles.', burden: 'Carga' },
   nl: { feasibility: 'Haalbaarheid', obstacles: 'Belangrijkste obstakels', agencies: 'Betrokken instanties', noStakeholderData: 'Geen gegevens over impact op belanghebbenden beschikbaar.', noImplementationData: 'Geen implementatiegegevens beschikbaar.', burden: 'Last' },
   ar: { feasibility: 'قابلية التنفيذ', obstacles: 'العقبات الرئيسية', agencies: 'الجهات المعنية', noStakeholderData: 'لا تتوفر بيانات تأثير أصحاب المصلحة.', noImplementationData: 'لا تتوفر بيانات تنفيذ.', burden: 'العبء' },

From ae267a9b17ae20b9879017e17cb08d0f6e059fc8 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Fri, 10 Apr 2026 11:34:13 +0000
Subject: [PATCH 4/5] fix: import URL utils locally in generators.ts and fix
 Korean typo in doc-type-helpers.ts

Agent-Logs-Url: https://github.com/Hack23/riksdagsmonitor/sessions/dff320eb-d844-4f76-8ec0-f968ed2780c5

Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
---
 .../content-generators/doc-type-helpers.ts          |  2 +-
 scripts/generate-news-enhanced/generators.ts        | 13 +++++++++++--
 2 files changed, 12 insertions(+), 3 deletions(-)

diff --git a/scripts/data-transformers/content-generators/doc-type-helpers.ts b/scripts/data-transformers/content-generators/doc-type-helpers.ts
index 056f07fd2..375f7a952 100644
--- a/scripts/data-transformers/content-generators/doc-type-helpers.ts
+++ b/scripts/data-transformers/content-generators/doc-type-helpers.ts
@@ -111,7 +111,7 @@ export const DOC_TYPE_DISPLAY: Readonly<Record<string, DocTypeLocalization>> = {
     plural: {
       en: 'External References', sv: 'Externa referenser', da: 'Eksterne referencer', no: 'Eksterne referanser',
       fi: 'Ulkoiset viitteet', de: 'Externe Referenzen', fr: 'Références externes', es: 'Referencias externas',
-      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참照', zh: '外部参考',
+      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참조', zh: '외부参考',
     },
   },
   other: {
diff --git a/scripts/generate-news-enhanced/generators.ts b/scripts/generate-news-enhanced/generators.ts
index 347e5e0fc..75061a04a 100644
--- a/scripts/generate-news-enhanced/generators.ts
+++ b/scripts/generate-news-enhanced/generators.ts
@@ -664,9 +664,9 @@ export async function generateInterpellations(): Promise<GenerationResult> {
 
 // ---------------------------------------------------------------------------
 // URL & text utilities — implementation extracted to url-utils.ts
-// Re-exported here for backward compatibility.
+// Imported for local use and re-exported for backward compatibility.
 // ---------------------------------------------------------------------------
-export {
+import {
   extractDocIdFromUrl,
   isGovernmentUrl,
   isGitHubUrl,
@@ -675,6 +675,15 @@ export {
   sanitizePlainText,
 } from './url-utils.js';
 
+export {
+  extractDocIdFromUrl,
+  isGovernmentUrl,
+  isGitHubUrl,
+  toGitHubRawUrl,
+  hashPathSuffix,
+  sanitizePlainText,
+};
+
 // ---------------------------------------------------------------------------
 // Deep-Inspection content generator (topic-focused, comprehensive)
 // ---------------------------------------------------------------------------

From 1b86b312ae85d70a547fc3458b90c75036a95fc2 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Fri, 10 Apr 2026 13:15:25 +0000
Subject: [PATCH 5/5] fix: correct zh translation typo and extend
 toGitHubRawUrl to handle tree URLs

Agent-Logs-Url: https://github.com/Hack23/riksdagsmonitor/sessions/4ac9d786-f068-4ebc-9b61-40abb31e209f

Co-authored-by: pethers <1726836+pethers@users.noreply.github.com>
---
 .../content-generators/doc-type-helpers.ts                  | 2 +-
 scripts/generate-news-enhanced/url-utils.ts                 | 6 ++++--
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/scripts/data-transformers/content-generators/doc-type-helpers.ts b/scripts/data-transformers/content-generators/doc-type-helpers.ts
index 375f7a952..dcc4f646c 100644
--- a/scripts/data-transformers/content-generators/doc-type-helpers.ts
+++ b/scripts/data-transformers/content-generators/doc-type-helpers.ts
@@ -111,7 +111,7 @@ export const DOC_TYPE_DISPLAY: Readonly<Record<string, DocTypeLocalization>> = {
     plural: {
       en: 'External References', sv: 'Externa referenser', da: 'Eksterne referencer', no: 'Eksterne referanser',
       fi: 'Ulkoiset viitteet', de: 'Externe Referenzen', fr: 'Références externes', es: 'Referencias externas',
-      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참조', zh: '외부参考',
+      nl: 'Externe referenties', ar: 'مراجع خارجية', he: 'הפניות חיצוניות', ja: '外部参照', ko: '외부 참조', zh: '外部参考',
     },
   },
   other: {
diff --git a/scripts/generate-news-enhanced/url-utils.ts b/scripts/generate-news-enhanced/url-utils.ts
index f864fecf0..7d9aed9c6 100644
--- a/scripts/generate-news-enhanced/url-utils.ts
+++ b/scripts/generate-news-enhanced/url-utils.ts
@@ -75,9 +75,10 @@ export function isGitHubUrl(url: string): boolean {
 }
 
 /**
- * Convert a GitHub blob/tree URL to a raw.githubusercontent.com URL.
+ * Convert a GitHub blob/tree/raw URL to a raw.githubusercontent.com URL.
  * Handles patterns like:
  *   - https://github.com/{owner}/{repo}/blob/{branch}/{path}
+ *   - https://github.com/{owner}/{repo}/tree/{branch}/{path}
  *   - https://github.com/{owner}/{repo}/raw/{branch}/{path}
  *   - https://raw.githubusercontent.com/{owner}/{repo}/{branch}/{path} (returned as-is)
  *
@@ -98,12 +99,13 @@ export function toGitHubRawUrl(url: string): string | null {
     }
 
     // Path: /{owner}/{repo}/blob/{branch}/{...path}
+    // or:   /{owner}/{repo}/tree/{branch}/{...path}
     // or:   /{owner}/{repo}/raw/{branch}/{...path}
     const segments = parsed.pathname.split('/').filter(Boolean);
     if (segments.length < 4) return null;
 
     const [owner, repo, refType, ...rest] = segments;
-    if (refType !== 'blob' && refType !== 'raw') return null;
+    if (refType !== 'blob' && refType !== 'raw' && refType !== 'tree') return null;
 
     // rest = [branch, ...pathParts]
     return `https://raw.githubusercontent.com/${owner}/${repo}/${rest.join('/')}`;