v2

README — Prompt Library v2

Overview

The v2/ prompt directory contains the production prompt templates for Riksdagsmonitor agentic workflows. All prompts integrate pre-computed analysis from analysis/daily/YYYY-MM-DD/ into article generation.

Files

File	Purpose
political-analysis.md	Core framework with pre-computed analysis integration
per-file-intelligence-analysis.md	Per-file AI analysis protocol with 6 analytical lenses
political-classification-prompt.md	Political classification (Temperature Index)
political-risk-prompt.md	Risk assessment methodology
political-threat-prompt.md	Political Threat Taxonomy (Attack Trees, Kill Chain, Diamond Model)
quality-criteria.md	Extended quality rubric with classification validation
stakeholder-perspectives.md	Multi-perspective analysis (6 perspectives)
swot-generation.md	SWOT generation with pre-computed data preference

Usage

All agentic workflows reference v2 prompts. Key integration points:

1. analysis/daily/YYYY-MM-DD/ files provide pre-computed analysis
2. Classification badges required for HIGH/CRITICAL articles
3. Confidence labels mandatory for all analytical claims
4. Forward indicators mandatory for all article types
5. Risk indicators (⚠️ inline tags) when riskLevel high/elevated

Analysis Reader API

See scripts/analysis-reader.ts for the TypeScript API:

import { readDailyAnalysis, readLatestAnalysis, deriveArticleClassificationMeta } from '../../analysis-reader.js';

// Read analysis for a specific date
const analysis = await readDailyAnalysis('2026-03-26');

// Read most recent available analysis (up to 7 days back)
const latest = await readLatestAnalysis();

// Derive metadata for ArticleData
const meta = deriveArticleClassificationMeta(analysis);
// meta: { classificationLevel, riskLevel, confidenceLabel, significanceScore, urgency }

Style Guide

All articles must comply with: analysis/methodologies/political-style-guide.md