You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
📋 Document Owner: CEO | 📄 Version: 2.0 | 📅 Last Updated: 2026-05-01 (UTC)
🏢 Owner: Hack23 AB (Org.nr 5595347807) | 🏷️ Classification: Public
📌 Template instructions: Produce for every workflow that declares electionCycleAnchor ∈ {current, next, both} in analysis/article-types.json. Save as analysis/daily/${ARTICLE_DATE}/${DOC_TYPE}/election-2026-analysis.md (stable filename) or election-cycle-analysis.md (post-rollover alias). Integrates with voter-segmentation.md, coalition-mathematics.md, and synthesis-summary.md.
✨ What to produce: A focused electoral assessment of the day's documents against the active election cycle (resolved at runtime), classified CRITICAL → NEGLIGIBLE, with coalition-scenario consequences and a dated watchlist.
📐 Template Contract — every fill of this template MUST satisfy this row.
e.g. Tidö 2022–2026 (current) or Post-2026 Mandate (next)
governingCoalition
electionCycles[cycleAnchor].governingCoalition
party array or null (next)
daysSinceCycleStart
articleDate − cycleStart (days)
computed at runtime
daysToCycleEnd
cycleEnd − articleDate (days)
computed at runtime
daysToElection
daysToElection(articleDate)
signed integer (negative = past)
Cycle-anchor footnote: The Swedish general elections anchoring this template are 2026-09-13 (next election from the current cycle) and 2030-09-08 (projected next election from the post-2026 cycle). See analysis/article-types.json → electionCycles for the authoritative dates.
Electoral outcome scenarios use WEP (likely/very likely for coalition continuity; roughly even for bloc shift); confidence MODERATE (polling inherent uncertainty)
Source Diversity Floor
P0 (election outcome claims): ≥4 sources (≥3 polls + ≥1 historical model); single poll labeled [unconfirmed — single source]
SAT(s) Applied
Morphological Analysis (bloc combinations), Indicators and Signposts (campaign milestones)
Purpose: AI-FIRST principle requires a Pass-2 read-back-and-improve. After producing this artifact in Pass 1, re-read it end-to-end and verify each item below. Document any remediation in methodology-reflection.md §"Pass-2 audit log". Any unchecked ❌ box at the end of Pass 2 forces a Pass-3 rewrite of the affected section.
Tradecraft anchors honoured — F3EAD stage matches the artifact's role; PIRs declared in the §Tradecraft Context block are actually addressed in the body; Admiralty grades attached to every external source; WEP band + ODNI confidence on every probabilistic judgement.
Source diversity floor met — at least the minimum number of independent MCP sources required by the artifact's tradecraft block are cited; single-source claims are explicitly labelled [SINGLE-SOURCE — corroboration pending].
Evidence specificity — every quantified claim cites a dok_id (Riksdag), an SCB / IMF dataflow code, or a named external source with date; no "according to data" / "studies show" hand-waves.
Named-actor discipline — every political claim names ≥ 1 person (party + role + dated act/quote) or labels the absence ([diffuse — no named actor]).
Counter-narrative present — at least one explicit competing hypothesis, dissent quote, or framed objection appears in the body; "no opposition recorded" is itself a finding to label, not silence.
Election cycle lens applied — the §"Cycle Watchlist" subsection addresses electoral salience, coalition pressure, and forward indicators for the active ${cycleAnchor} cycle; not boilerplate. If cycleAnchor=current, §"Mandate-Fulfilment Scorecard" is present and populated. If cycleAnchor=next, §"Coalition-Formation Forecast" is present and populated.
No illustrative content shipped as fact — every [REQUIRED] placeholder is filled OR removed; every Example: block is clearly fenced or removed; no fabricated dok_id, vote count, or quote leaks into the final artifact.
Cross-references resolve — every [link](file.md) in this artifact points to a file that exists in the run folder (analysis/daily/$ARTICLE_DATE/$SUBFOLDER/) or to a methodology / template under analysis/.
Mermaid renders — every fenced ```mermaid block parses (no missing class definitions, no orphan nodes, no >40-node graphs that overflow viewport on mobile).
Line-floor check — artifact length ≥ the per-artifact floor in reference-quality-thresholds.json; shorter artifacts trigger Pass-2 rewrite, never a [truncated] note.