🗳️ Riksdagsmonitor

July 8, 2026 · View on GitHub

Swedish Political Intelligence Platform — democratic transparency, evidence-based analysis and AI-generated political news, powered by official open data and a fully autonomous agentic newsroom.

🕵️ Political intelligence · 🔍 Democratic transparency · 🤖 AI-generated news · 📊 50+ years of evidence

Riksdagsmonitor monitors Sweden's Riksdag (Parliament), the Government (Regeringskansliet) and public agencies (Myndigheter) with structured intelligence techniques — ACH, SWOT, PESTLE, STRIDE, political-risk scoring and OSINT/INTOP tradecraft — applied to 349 current MPs, 2,494 historical politicians (1971–2024), 3.5M+ votes and 109,000+ parliamentary documents.

An autonomous AI newsroom — 14 agentic workflows, Claude Opus 4.8 (Claude Sonnet 4.6 for translation), zero human editors — turns this evidence into publication-ready intelligence articles in 14 languages, every day.

🌐 Live Platform • 🕵️ Political Intelligence • 📰 AI Newsroom • 📊 Intelligence Dashboard • 🗺️ Sitemap

🎯 Mission

Strengthen Swedish democracy through systematic transparency.

Riksdagsmonitor exists to put rigorous, evidence-based political intelligence in the hands of every citizen, journalist, researcher and policymaker. We combine:

🕵️ Structured intelligence tradecraft — ACH, SWOT, PESTLE, STRIDE, political-risk scoring, OSINT/INTOP methodology, ICD-203 Key Judgments
🔍 Democratic transparency — every claim traceable to a primary source (dok_id, vote count, named MP, government document) or it does not get published
🤖 AI-generated political news — a fully autonomous newsroom turning open data into 14-language analysis articles, daily
📊 50+ years of historical evidence — 1971–2024 voting records, document corpus, and party-evolution analytics from the Citizen Intelligence Agency (CIA) platform
⚖️ Neutrality & GDPR by design — equal treatment of all 8 parliamentary parties, public-data only, privacy-by-design architecture, explicit democratic safeguards

The platform is non-partisan, open-source (Apache-2.0), and operated under the Hack23 ISMS with full ISO 27001:2022 / NIST CSF 2.0 / CIS Controls v8.1 alignment. It does not collect user data, does not run ads, does not push narratives, and is architecturally designed so it cannot be weaponised for partisan influence.

🌐 Explore the Platform

Five flagship pages anchor the public site. Each is multilingual (14 languages), WCAG 2.1 AA accessible, SEO-optimised with JSON-LD structured data and hreflang alternates, and CSP-hardened with Subresource Integrity (SRI) on every CDN asset.

Icon	Page	What it does
🌐	Live Platform — riksdagsmonitor.com	The public homepage and primary entry point. Front-loads the current coalition status (e.g. Tidö Agreement — 176/349 seats, fragility indicators, CIA risk alerts), then deep-dives into the full intelligence stack: 🗳️ Election-cycle intelligence — 40 years (1994–2034) 🏛️ Party performance & effectiveness — 1990–2026, 8 parties 🤝 Committee networks & productivity — 15 committees 📊 Coalition & voting-pattern analysis 🌡️ Seasonal activity with Z-score anomaly detection — 2002–2025 🔮 Pre-election monitoring 🎖️ Ministry risk & influence 🚨 45-rule × 349-MP risk heat map (live) Updated daily at 03:00 CET.
🕵️	Political Intelligence Index	The single canonical entry point for OSINT/INTOP researchers. Catalogues every methodology that governs analysis on the platform: 📐 AI-Driven Analysis Guide 🔍 OSINT Tradecraft Standards ⚠️ Political Risk Methodology 💼 Political SWOT Framework 🎯 Political Threat Framework 🏷️ Political Classification Guide 🗳️ Electoral Domain Methodology 🧬 Synthesis Methodology 📈 IMF & 🌍 World Bank Indicator Mappings 📏 Reference-Quality Thresholds · ✍️ Political Style Guide Plus the full daily artifact catalogue (Family A baseline · Family B · Family C synthesis · Family D extension). Each item links straight to the source-of-truth on GitHub. Available in 14 languages (append `_sv`, `_de`, `_ar`, …).
📰	AI-Generated News & Analysis	The world's first fully AI-driven political-intelligence newsroom for parliamentary monitoring. 14 specialised agentic workflows (scheduled daily, also manually dispatchable, powered by Claude Opus 4.8 — Claude Sonnet 4.6 for the translation workflow — via GitHub Copilot Coding Agent) autonomously produce daily intelligence articles: 🌅 Committee reports · 🏛️ propositions · ✊ motions · ❓ interpellations 🔮 Week-ahead · 📅 month-ahead · 🔍 real-time monitor 🌆 Evening analysis · 📊 weekly review · 📈 monthly review Source verification, multi-party balance and GDPR-compliant OSINT methodology baked in. Every article carries machine-readable provenance through JSON-LD `NewsArticle.isBasedOn`, links back to the analysis artifacts that produced it, and is published in 14 languages.
📊	CIA Intelligence Dashboard	Interactive Chart.js / D3.js intelligence dashboard powered by the Citizen Intelligence Agency (CIA) data products: 📋 Overview metrics — MPs, parties, risk rules, coalition seats 🚨 Risk alerts — Critical / Major / Minor (last 90 days) 🏛️ Party performance analysis 🗳️ Swedish Election 2026 predictions — coalition scenarios & key factors 👥 Top-10 most influential MPs — network analysis 🔥 Voting-patterns heat map 🤝 Committee network analysis 🎖️ Ministry performance 📊 Parliamentary demographics · 📄 document activity · 📉 risk-score evolution Local-first data loading with 1-hour cache, keyboard navigable, CSP-compliant.
🗺️	Sitemap (Human + Machine)	Human-readable navigation of every page on the platform plus the machine-readable `sitemap.xml` and `rss.xml` feeds. Daily refreshed; honours `robots.txt` and search-engine indexing best practices.

🌍 14 supported languages: English · Swedish · Danish · Norwegian · Finnish · German · French · Spanish · Dutch · Arabic (RTL) · Hebrew (RTL) · Japanese · Korean · Chinese.

🤖 AI-Generated Political Intelligence News

"While traditional newsrooms debate whether AI will replace journalists, Riksdagsmonitor already runs a fully autonomous political-intelligence newsroom — 14 agentic workflows, 14 languages, zero human editors, and a publication schedule that would bankrupt any legacy outlet trying to keep up."

What makes it different

Traditional AI-generated news is shallow — it rewrites press releases. Riksdagsmonitor's pipeline performs deep political-intelligence analysis before a single sentence of an article is written:

🔎 Source verification — every claim traced to official Riksdag/Regering data via the riksdag-regering MCP server (32+ tools)
⚖️ Multi-party balance — algorithmic fairness across all 8 parliamentary parties, no editorial bias possible
📊 Quantitative rigor — voting cohesion metrics, attendance scoring, legislative productivity indices, ICD-203 Key Judgments, 45-rule transparency scoring
🌐 14-language reach — culturally adapted political analysis with RTL support (Arabic / Hebrew), not machine translation
🛡️ GDPR-compliant OSINT — only public parliamentary data; political opinions are processed under GDPR Art. 9(2)(e) (manifestly made public) / 9(2)(g) (substantial public interest)
🔭 Forward-look coverage — week / month / quarter / year / election-cycle horizons with registry-driven horizon stratification (see analysis/article-types.json)
🚦 Hard analysis gate — every news workflow MUST produce 9 (single-type) or 14 (Tier-C) analysis artifacts on disk before article generation; the gate at .github/prompts/05-analysis-gate.md is non-negotiable
🧪 Quality gates — HTMLHint validation, link checking, accessibility (WCAG 2.1 AA) and translation parity in CI before publication

End-to-end pipeline

flowchart LR
  A[📥 MCP + CIA + SCB + IMF<br/>data download] --> B[📐 Apply methodology<br/>analysis/methodologies]
  B --> C[📋 Populate templates<br/>analysis/templates]
  C --> D[📂 Write 9 or 14 artifacts<br/>analysis/daily/$DATE/$SUBFOLDER]
  D --> E{🚦 Analysis Gate<br/>prompts/05}
  E -- pass --> F[📰 Aggregate → article.md<br/>scripts/aggregate-analysis.ts]
  E -- fail --> C
  F --> G[🌐 Sanitised HTML<br/>scripts/render-articles.ts]
  G --> H[🌍 Translate into<br/>12 other languages]
  H --> I[🔀 One PR per article type<br/>S3 + CloudFront deploy]
  style A fill:#0a0e27,stroke:#00d9ff,color:#e0e0e0
  style E fill:#dc3545,stroke:#b02a37,color:#fff
  style I fill:#1a1e3d,stroke:#ffbe0b,color:#e0e0e0

A new .md artifact written anywhere under analysis/daily/$DATE/$SUB/ is enough to publish an English + Swedish HTML article on the next CI build — there is no manual scaffolding, no template fill-in, and no per-type generator class. The dedicated news-translate workflow then propagates to the remaining 12 languages out-of-band. See Article-Generation.md for the full agentic-workflow contract.

Autonomous publication schedule

Time (UTC)	Workflow	Coverage	Frequency
🌅 04:00	Committee Reports	Utskottsbetänkanden analysis, voting breakdowns	Mon–Fri
🌅 05:00	Propositions	Government bills, legislative impact assessment	Mon–Fri
☀️ 06:00	Motions	Opposition proposals, party-strategy decoding	Mon–Fri
❓ 07:00	Interpellations	Ministerial accountability, evasion detection	Mon–Fri
🔮 07:00	Week Ahead	Parliamentary calendar preview, agenda intelligence	Friday
📅 08:00	Month Ahead	Strategic outlook, coalition forecasting	1st of month
🔍 10:00 & 14:00 (Mon–Fri); 12:00 (Sat/Sun)	Real-Time Monitor	Breaking political developments, flash analysis	Mon–Fri (×2) + weekends
🌍 11:00 & 17:00 (Mon–Fri); 14:00 (Sat/Sun)	Translate	12 additional languages from EN/SV cores	Daily
🌆 18:00 (16:00 Sat)	Evening Analysis	Deep-dive intelligence synthesis	Mon–Sat
📊 09:00	Weekly Review	Week-in-review scorecard, party performance	Saturday
📈 10:00	Monthly Review	Comprehensive monthly intelligence assessment	28th of month

Authoritative schedules defined in .github/workflows/news-*.lock.yml — see .github/workflows/README.md for the complete inventory.

Result: dozens of articles per week across 14 languages — hundreds of localised intelligence products each month, generated autonomously with zero editorial intervention, every one of them auditable down to its source dok_id.

🕵️ Political Intelligence Methodology

Every analysis on the platform is governed by an explicit, version-controlled methodology. The full library is browseable at https://riksdagsmonitor.com/political-intelligence.html; the canonical sources live in analysis/.

Layer	Document	What you'll find
🧭 Framework	`analysis/README.md`	Artifact taxonomy, 9-artifact / 14-artifact contract, on-disk daily layout
📐 Methodology library	`analysis/methodologies/README.md`	11 methodology documents — AI-Driven Analysis Guide · OSINT Tradecraft Standards · Political Risk Methodology · Political SWOT Framework · Political Threat Framework · Political Classification Guide · Electoral Domain Methodology · Synthesis Methodology · Per-Document & Per-Artifact Methodologies · Strategic Extensions (scenario / wildcard / long-horizon) · Structural Metadata · Political Style Guide
📋 Template library	`analysis/templates/README.md`	23 templates — 8 core single-type (T1–T8) plus 15 extended / Tier-C (executive-brief, scenario-analysis, coalition-mathematics, election-2026, historical-parallels, comparative-international, devil's-advocate, intel-assessment / ICD-203 Key Judgments, …)
🚦 News-generation contract	`.github/prompts/README.md`	8 bounded-context prompt modules + Tier-C extension; the single blocking analysis gate
⚙️ Workflow orchestration	`.github/workflows/README.md` + `WORKFLOWS.md`	How each `news-*.md` source compiles to a hardened `.lock.yml` with SHA-pinned actions, egress firewall (Squid + iptables) and five-layer safe-outputs
🔍 IMF integration	`analysis/imf/README.md` + `.github/aw/ECONOMIC_DATA_CONTRACT.md`	Macro/fiscal/monetary/external context with T+5 projections; canonical pattern for every economic claim
🏛️ Statskontoret integration	`analysis/statskontoret/README.md`	Swedish agency structure and central-government budget execution (årsutfall / månadsutfall / tidsserier)
🤖 Specialist personas	`.github/agents/README.md`	14 persona agents incl. `intelligence-operative`, `news-journalist`, `content-generator` + 9 workflow-specialist agents + shared `developer.instructions.md`
🧠 Skills library	`.github/skills/README.md`	91 skills across 12 functional categories — 11 political-intelligence, 5 journalism, 14 ISMS/security, 13 GitHub Agentic Workflows, …

Evidence standard

Every claim must tie to: a dok_id citation, a named actor, a vote count, or a primary-source URL. Generic statements without evidence are rejected by the analysis gate. AI is an accelerator — never an excuse for shallow output.

AI-FIRST quality principle

Minimum 2 complete iterations for every analysis. Pass 1 produces the artifact; Pass 2 reads everything back and improves every section (stronger evidence, deeper analytic rigor, broader stakeholder coverage, quantified risk). Workflows that complete in under 75 % of their allocated time are treated as failed. See AGENTS.md §AI-FIRST.

📊 Interactive Intelligence Dashboards

Five flagship Chart.js / D3.js dashboards on https://riksdagsmonitor.com/ (and the consolidated CIA Intelligence Dashboard):

#	Dashboard	Coverage	Visualisations	Data
1	🌡️ Seasonal Activity Patterns	2002–2025 quarterly (23+ years)	Heat maps, time series, Z-score anomaly detection (\|Z\| ≥ 2.0)	`cia-data/seasonal/`
2	👤 Politician Dashboard	349 current MPs	Top-10 rankings, 45-rule risk profiles, influence metrics	`cia-data/politician/`
3	🗳️ Pre-Election Monitoring	Q4 2023 → 2025	Historical comparisons, election-year vs non-election Q4 patterns, early-warning indicator matrix	`cia-data/pre-election/`
4	🏛️ Party Performance & Effectiveness	1990–2026 (37 years, 8 parties)	Effectiveness trends, comparative bars, coalition-alignment, momentum with P50/P90 percentiles	`cia-data/party/`
5	🚨 Anomaly Detection & Early Warning	2002–2026 (41 quarters)	Timeline · Z-score distribution · type breakdown · severity heat map · recent-anomaly feed	`cia-data/seasonal/`

Dashboard properties — local-first data loading with 1-hour cache; WCAG 2.1 AA accessible (keyboard, screen reader, 4.5:1 contrast); 14-language; responsive 320 px → 1440 px+; CSP-compliant with SRI hashes (sha384) on all CDN resources.

🗳️ Transparency Statistics

Live numbers (updated daily at 03:00 CET via update-cia-csv-data.yml):

Metric	Value	Note
👥 Current MPs	349	All active Members of Parliament
📜 Historical politicians	2,494	1971–2024 (50+ years)
🗳️ Votes analysed	3.5+ million	Comprehensive voting-record corpus
📄 Documents processed	109,000+	Parliamentary documents (motions, propositions, interpellations, …)
🏛️ Committee documents	8,740	Committee work tracked
⚠️ Rule violations identified	2,308	Across 45 transparency rules
🇸🇪 Political parties	8	All Riksdag-represented parties
⏱️ CIA subsystems	15	anomaly · coalition · committee · distribution · election · election-cycle · ministry · parties · party · percentile · politician · pre-election · risk · seasonal · voting
📰 News articles published	2,669+ files under `news/`	14 languages

Data source: extraction_summary_report.csv · cached in cia-data/production-stats.json (24 h freshness).

🔗 Authoritative Data Sources

Riksdagsmonitor uses a provider-tiered data architecture, with each provider chosen for its area of strength.

Tier	Provider	Scope	Access
🏛️ Parliamentary primary	Riksdagen Open Data	Documents, motions, votes, MPs, speeches, committees	`riksdag-regering` MCP server (32+ tools)
🏢 Government primary	Regeringskansliet	Propositions, SOU, Ds, directives, press releases	`riksdag-regering` MCP server
📈 Primary economic	IMF (Datamapper REST + SDMX 3.0)	GDP, growth, unemployment, inflation, fiscal balance, debt, current account, bilateral trade, commodity prices, exchange rates, government spending by COFOG function — with T+5 projections	Pure-TypeScript client `scripts/imf-client.ts` (intentionally non-MCP)
🇸🇪 Swedish ground truth	SCB (PxWeb v2)	Swedish monthly labour (AKU), monthly inflation (KPI), regional/municipal, budget execution	`scb` MCP server (`@jarib/pxweb-mcp@2.0.0`, 1,200+ tables)
🏛️ Statskontoret	Statskontoret	Authority count, dept grouping, leadership form, FTE / headcount, central-government budget outturns	`scripts/statskontoret-client.ts`
🌍 Non-economic residue	World Bank	Governance (WGI, `source=75`), environment, social/education residue, defence historicals	`world-bank` MCP server (`worldbank-mcp@1.0.1`)
🗳️ Election authority	Valmyndigheten	Election results, voter turnout, electoral statistics	Public datasets
💰 ESV	Ekonomistyrningsverket	Government budget and spending data	Public datasets
🕵️ Citizen Intelligence Agency	Hack23/cia	15 CIA subsystems consumed nightly via `update-cia-csv-data.yml`	JSON / CSV exports

Why this split — IMF uses uniform SNA 2008 / GFSM 2014 / BPM6 methodology across countries (essential for cross-country comparison), publishes T+5 projections (essential for look-ahead workflows), and has fresher data than World Bank's economic indicators. World Bank remains canonical for the classes IMF does not publish (WGI governance, environment). SCB is the Swedish-specific ground-truth layer. Banned phrases (e.g. "the World Bank reports Swedish GDP growth of …") and vintage discipline (data > 6 months old → annotation required) are enforced by CI per .github/aw/ECONOMIC_DATA_CONTRACT.md v2.1.

🏗️ Technical Architecture

Full system architecture (C4 Context / Container / Component / Dynamic views) lives in ARCHITECTURE.md (v2.2).

Stack

Frontend — Static HTML5 / CSS3 with TypeScript-built Chart.js / D3.js dashboards (no SPA framework, mobile-first, cyberpunk theme)
Language — TypeScript 6.0.2 compatibility package (typescript) plus TypeScript 7.0.1-rc (typescript-7) running side-by-side while the migration is underway
Build — Vite 8.1.3 (ES modules, code splitting, SRI via vite-plugin-sri-gen)
Visualisation — Chart.js 4.5.1 + D3.js 7.9.0, hosted locally on CloudFront
Testing — Vitest 4.1.10 (2,890 unit tests, 100 % pass rate, 70 % line coverage) + Cypress 15.18.1 (E2E)
Hosting — AWS CloudFront + S3 dual-region (us-east-1 primary, eu-west-1 replica) via OIDC; GitHub Pages as DR fallback
CI/CD — 50 GitHub Actions workflow files (22 standard .yml + 14 agentic .md sources + 14 compiled .lock.yml); SHA-pinned, step-security/harden-runner everywhere
Data Platform — Citizen Intelligence Agency (CIA) Java/Spring Boot backend + 15 CIA subsystems
Runtime — Node.js 26.x

Architecture documentation portfolio

Current State	Future State
🏗️ Architecture	🚀 Future Architecture
📊 Data Model	📊 Future Data Model
🔄 Flowcharts	🔄 Future Flowcharts
🔄 State Diagrams	🔄 Future State Diagrams
🗺️ Mindmap	🗺️ Future Mindmap
💼 SWOT	💼 Future SWOT

🔐 Security, Privacy & ISMS Compliance

Full controls in SECURITY_ARCHITECTURE.md · threat model in THREAT_MODEL.md · CRA conformity in CRA-ASSESSMENT.md.

Classification (per Hack23 Classification Framework)

Dimension	Level	Note
🔒 Confidentiality	🟢 Public	All data intentionally disclosed (Swedish open data + website content)
✅ Integrity	🟠 High	Automated validation, GPG-signed commits, SLSA build provenance
🟢 Availability	🟠 High	99.998 % design target (CloudFront 99.9 % SLA + multi-region S3 + GitHub Pages DR)
🏷️ Privacy	🟠 Personal (public officials only)	GDPR Art. 6(1)(e/f); Art. 9(2)(e/g) for political opinions; no end-user PII, no accounts, no ads, no tracking
⏱️ RTO / RPO	1–4 h / 4–24 h	Automated multi-region failover, daily data refresh
💰 Business impact	🟢 Low (financial) · 🟡 Moderate (reputational)	Open-source project, no revenue dependency

Compliance frameworks

ISO 27001:2022 — 7 Annex A controls implemented · NIST CSF 2.0 — 6 functions aligned · CIS Controls v8.1 — 6 controls implemented
GDPR — public-interest / legitimate-interest grounds for public-official data; political opinions under Art. 9(2)(e)/(g)
EU CRA — self-assessment in CRA-ASSESSMENT.md
OpenSSF Best Practices — Project #12069

Defence-in-depth highlights

🔒 HTTPS-only — TLS 1.3, HSTS, CSP, X-Frame-Options, X-Content-Type-Options
🧱 Five-layer agentic-workflow security — read-only agent tokens, zero secrets in agent context, containerised + Squid/iptables egress firewall, safe-outputs validation, AI threat-detection scan
🔐 SHA-pinned actions + step-security/harden-runner + Dependabot + CodeQL + Secret Scanning + dependency-review + OIDC-only AWS deploy
📜 Aligned ISMS policies — Information Security Policy · Secure Development Policy · Threat Modeling · AI Policy · Vulnerability Management · Change Management · Incident Response Plan · Open Source Policy

Metric	Status
Risk level	🟢 LOW (5.52 / 10.0 — 99.7 % risk reduction)
HTML validation	✅ 0 errors (HTMLHint)
Dependencies	✅ Dependabot clean
Secrets	✅ Secret Scanning enabled
Code scanning	✅ CodeQL active

📦 npm Package

The platform's reusable shared utilities are published as riksdagsmonitor with SLSA build provenance.

npm install riksdagsmonitor

Includes — Theme System (cyberpunk, WCAG AA) · Chart Factory (Chart.js with responsive breakpoints + keyboard nav) · Resilient Data Loader (retry, cache, CSV/JSON) · DOM utilities · Full TypeScript types · 12 dashboard modules · CIA intelligence modules.

import { getActiveThemeColors, BREAKPOINTS, getPartyColor } from 'riksdagsmonitor';
import { loadJSON, loadCSV, createDataSource } from 'riksdagsmonitor';
import { showLoadingState, formatNumber, debounce } from 'riksdagsmonitor';
import { createChart, initDashboardSection } from 'riksdagsmonitor/shared/chart-factory';
import { CIADataLoader } from 'riksdagsmonitor/cia/data-loader';
import { CIADashboardRenderer } from 'riksdagsmonitor/cia/visualizations';

Peer dependencies — chart.js d3 papaparse (required for dashboards); chartjs-plugin-annotation (optional, loaded conditionally).

🚀 Development

Prerequisites

Node.js ≥ 26 · npm ≥ 10 · Git with GPG signing · GitHub MFA + SSH keys.

Quick start

git clone git@github.com:Hack23/riksdagsmonitor.git
cd riksdagsmonitor
npm install
npm run dev          # Vite dev server with hot reload → http://localhost:8080

# Tests
npm test                 # Vitest unit (2,890 tests)
npm run test:coverage    # with coverage
npm run cypress:open     # E2E interactive
npm run e2e              # full E2E (build + preview + Cypress)

# Quality
npm run htmlhint         # HTML5 validation
npm run linkcheck        # link integrity (linkinator)

# Production
npm run build            # Vite production build → dist/
npm run preview          # http://localhost:4173

CI/CD

HTMLHint validation · Linkinator · Vitest (2,890 tests) · Cypress · Vite build · Dependency review · CodeQL · Secret scanning · Translation validation
Releases — workflow_dispatch or v*.*.* tag → SBOM (SPDX) + SHA-256 + SLSA Build Provenance attestations + dual deployment (S3/CloudFront primary, GitHub Pages DR)
gh attestation verify riksdagsmonitor-vX.Y.Z.zip -R Hack23/riksdagsmonitor

See WORKFLOWS.md and RELEASE_PROCESS.md for the canonical reference.

🤖 GitHub Copilot — Agents, Skills & Agentic Workflows

Riksdagsmonitor uses GitHub Copilot personas, skills and agentic workflows as first-class automation. Directory READMEs are the single source of truth; AGENTS.md and SKILLS.md are the long-form catalogs.

Surface	Catalog	Count
🤖 Custom agents	`.github/agents/README.md`	24 files (14 personas + 9 workflow-specialists + 1 shared developer-instructions)
🧠 Skills	`.github/skills/README.md`	91 skills across 12 functional categories
📜 Prompt modules	`.github/prompts/README.md`	8 bounded-context modules + Tier-C extension
⚙️ Workflows	`.github/workflows/README.md`	50 files (22 standard + 14 agentic sources + 14 compiled)
🔌 MCP servers	`.github/copilot-mcp.json`	8 — `riksdag-regering`, `scb`, `world-bank`, `github` (insiders), `filesystem`, `memory`, `sequential-thinking`, `playwright`

14 persona agents (assignable via assign_copilot_to_issue): security-architect · documentation-architect · quality-engineer · frontend-specialist · isms-compliance-manager · deployment-specialist · devops-engineer · intelligence-operative · news-journalist · content-generator · data-pipeline-specialist · data-visualization-specialist · task-agent · ui-enhancement-specialist.

9 workflow-specialist agents (.agent.md, invoked from workflows): agentic-workflows · ci-cleaner · contribution-checker · create-safe-output-type · custom-engine-implementation · grumpy-reviewer · interactive-agent-designer · technical-doc-writer · w3c-specification-writer.

📖 Documentation Index

Project documentation

Article-Generation.md — Canonical agentic-newsroom contract (data → analysis → article → HTML → S3)
AGENTS.md · SKILLS.md — Copilot persona & skill catalogs
WORKFLOWS.md — End-to-end CI/CD reference (v7.2)
RELEASE_PROCESS.md — Release with attestations & docs-as-code
TRANSLATION_GUIDE.md — 14-language standards & glossary
SECURITY.md · CONTRIBUTING.md · CODE_OF_CONDUCT.md · LICENSE

Future planning

FUTURE_WORKFLOWS.md · FUTURE_MINDMAP.md · FUTURE_ARCHITECTURE.md · FUTURE_SECURITY_ARCHITECTURE.md · FUTURE_SWOT.md

External

📚 API Documentation (JSDoc) · Test Coverage · Cypress E2E Reports
🕵️ CIA Platform · CIA JSON Export Specs
🛡️ Hack23 ISMS-PUBLIC · 📰 Hack23 Blog

🔮 Future Roadmap (2026 → 2037)

From agentic news generation to AGI-powered democratic intelligence — an 11-year evolution.

Detailed planning: FUTURE_WORKFLOWS.md · FUTURE_MINDMAP.md · FUTURE_ARCHITECTURE.md

%%{init: {"theme": "dark", "themeVariables": {"primaryColor": "#00d9ff", "primaryTextColor": "#e0e0e0", "primaryBorderColor": "#ff006e", "lineColor": "#ffbe0b", "secondaryColor": "#1a1e3d", "tertiaryColor": "#0a0e27"}}}%%
timeline
    title Riksdagsmonitor Evolution — 2026 → 2037
    section Phase 3 — Foundation (2026)
        Q1-Q2 : ✅ TypeScript migration · 14 agentic news workflows
              : ✅ 14 languages · S3 + GitHub Pages dual deploy
        Q3-Q4 : 🔄 CIA pipeline integration · predictive dashboards
    section Phase 4 — Predictive Intelligence (2027–2029)
        2027 : Real-time streams · ML election models · Nordic expansion (DK · NO · FI)
        2028 : Multi-modal (audio/video) · AI editorial board · self-healing pipelines
        2029 : Fully autonomous · expert-level analysis · API platform launch
    section Phase 5 — Scale (2030–2033)
        2030-2031 : EU-wide coverage (27 + EP) · sovereign-AI evaluation
        2032-2033 : Global parliamentary watch (100+ democracies) · every UN language · pre-AGI
    section Phase 6 — AGI Era (2034–2037)
        2034-2035 : AGI-enhanced intelligence · 195 parliaments · predictive governance
        2036-2037 : Transformative democracy platform · democratic safeguards architecture

Year	Automation assets	AI model	Key capability
2026	44 → 50	Opus 4.8–4.9	🤖 Agentic news generation (current)
2027	50–55	Opus 5.x	🔮 Predictive analytics & Nordic expansion
2028	55–65	Opus 6.x	🎙️ Multi-modal content
2029	65–75	Opus 7.x	🚀 Fully autonomous pipeline
2030–2033	75–100	Opus 8–10.x / pre-AGI	🌍 EU → Global coverage
2034–2037	100–120+	AGI / post-AGI	⚡ Transformative democracy platform

Version numbers are illustrative — actual products and paradigm shifts (quantum AI, neuromorphic computing) will vary. Architecture is designed for graceful adaptation while preserving democratic safeguards: human oversight maintained regardless of AI capability, anti-weaponisation by design, constitutional alignment encoded in platform architecture.

🤝 Contributing

Contributions welcome under Hack23's secure-development standards.

Fork the repository and create a descriptive feature branch
GPG-sign every commit · enable MFA on your GitHub account
Run quality checks locally (npm run htmlhint && npm test && npm run build)
Submit a pull request with comprehensive description; address review feedback
Never introduce security vulnerabilities; follow CONTRIBUTING.md and CODE_OF_CONDUCT.md

🏢 About Hack23

Hack23 AB (Org.nr 559534-7807) — Swedish cybersecurity and open-source intelligence consultancy.

🌐 www.hack23.com · 📰 Blog
💼 LinkedIn — Hack23
👨‍💻 Founder: James Pether Sörling, CISSP, CISM
🛡️ Public ISMS · 🕵️ CIA platform

📜 License

🗳️ Empower citizens · 🔍 Strengthen democratic accountability · 🕵️ Illuminate the political process