lev-narrative-rollup

name	lev-narrative-rollup
description	Use when revising the public current LEV story, homepage State of LEV story, what to watch next, where better evidence is needed, or story revision triggers. Compares the current story against public outlooks, state-of-the-field editions, public updates, and review due dates without introducing new unreviewed evidence.

LEV Narrative Rollup

Use this after reviewed public records have changed, or when the curator asks whether the homepage story still matches the evidence map.

If the task involves deciding which public updates to include, defer, omit, or escalate before revising public copy, use lev-reconciliation-orchestrator first and carry over its recorded recommendations and approvals.

Read First

• data/content/current-lev-story/current.json
• schemas/current-lev-story.schema.json
• latest data/content/state-of-the-field/*.json
• data/outlooks/overall-lev-outlook.json
• public outlooks named in the narrative related_outlook_ids
• data/publication-events/ newer than revision.last_reviewed
• docs/editorial-rollup.md
• docs/editorial-quality-system.md
• npm run state-of-field:packet output when State of the Field reconciliation affects the story
• homepage rendering in src/components/homepage.tsx
• scripts/current-lev-story.mjs

Staleness Checks

The narrative needs review when any of these are true:

1. A newer public update record has affected_outlook_ids.
2. Overall, hallmark, or track evidence stage, momentum, read firmness, evidence gap, strongest evidence, or LEV 2036 outlook changed.
3. A new or revised state-of-the-field edition changes the monthly public interpretation.
4. revision.review_due has passed.
5. The curator asks for a story, watchlist, focus-priority, or homepage framing review.

Workflow

1. Compare the current story against the latest public outlooks and public updates.
   • Start with npm run story:current -- status.
2. Decide whether the existing story is current, stale, or only needs metadata refresh.
3. If stale, revise only from already reviewed public records:
   • Generate a deterministic plain-language draft with npm run story:current -- draft --write --style plain and use it as review input, not automatic truth.
   • title and summary for the main homepage story
   • current_evidence_picture and what_changed for the hero interpretation
   • recent_developments for temporal progress that affected public interpretation
   • what_to_watch_next for signals that could change the outlook next
   • where_better_evidence_is_needed for clear next steps, underbuilt evidence, early promise, or blocking gaps
   • before_now_next, what_would_change_the_outlook, and track_examples_to_inspect when the homepage needs a clearer reader path
   • revision metadata and triggers
4. Keep evidence, interpretation, and outlook distinct. Do not turn activity, funding, recruiting status, or source coverage into proof of LEV progress.
5. If the revision would require new sources, findings, stages, read-firmness/confidence changes, or outlook upgrades, stop and use the field-change-check/evidence-review/staged-update workflow instead.

Writing Rules

• Lead with the aggregate field story, not isolated data points.
• Say what changed and what did not change.
• Use conservative language for speculative areas.
• Prefer plain reader language over internal research language: "early human signals" instead of "human-facing signals"; "missing-context update" instead of "coverage repair"; "reason to be more optimistic" instead of "forecast upgrade".
• Do not use "confidence" as public prose. For outlook stability use "read firmness", "tentative", "provisional", or "firm"; for individual findings use "evidence weight".
• Tie what-to-watch-next items to concrete signals that would change interpretation.
• Tie better-evidence needs to a reason: clear_next_step, underbuilt_evidence, early_promise, or blocking_gap.
• Update revision.last_reviewed, revision.review_reason, revision.review_due, and relevant related_*_ids.
• After accepting a reviewed story state, run npm run story:current -- snapshot --write-current so future checks can diff outlook evidence stages, field momentum, read firmness, evidence gaps, strongest evidence, and LEV 2036 outlook.

Checks

Run after edits:

npm run audit:editorial -- --write
npm run validate:records
npm run audit:data
npm run typecheck
npm run build

If visual layout changed, start the dev server and inspect / on desktop and mobile widths.