Riksdagsmonitor Intelligence Platform โ€” API Documentation - v1.0.30
    Preparing search index...
    AGGREGATION_ORDER: readonly string[] = ...

    Canonical narrative order. Each file is emitted as an <h2>-prefixed section in the aggregated article. Unknown artifacts (e.g. supplementary PESTLE / black-swan studies) are appended after the core sections in the order they appear on disk.

    The order follows a journalist-optimal narrative arc for political- intelligence reporting:

    Phase A โ€” Lead & headline judgments (BLUF) Phase B โ€” Primary evidence (documents/* expanded after significance-scoring.md; see aggregate.ts step 3) Phase C โ€” Actors & political arithmetic (stakeholders, coalition math, voter segmentation) Phase D โ€” Forward trajectory (indicators, scenarios, election, cycle, parliamentary season) Phase E โ€” Risk, threat & strategic posture (risk, SWOT, threat, STRIDE, wildcards, PESTLE) Phase F โ€” Context & narrative environment (history, comparative, feasibility, media framing) Phase G โ€” Critique (devil's advocate) Phase H โ€” Audit appendix (classification, cross-refs, methodology, manifest)

    Rationale: readers form their own view of the substance first (Phases Aโ†’D), then weigh risk and context (Phases Eโ†’F), and finally see narrative-environment / influence-operations and devil's-advocate critique (late F + G) before the appendix. Anchoring per-document evidence right after the so-what ranking ("show your work") is the standard intelligence-product structure (cf. ICD 203, NIC NIE).