review-data-pr

name	review-data-pr
description	Review an OWID ETL data update PR end-to-end — runs the pipeline, compares snapshot fields against the previous version, verifies links, audits indicator metadata coverage, and cross-checks workflow items from /update-dataset. Trigger when the user asks to "review this PR", "review the data PR", or invokes this on an open dataset-update branch.
metadata	{"internal":true}

Review Data PR

End-to-end review of a dataset-update PR. Goes deeper than /review: actually runs the steps, compares to the previous version, audits metadata coverage against a fixed checklist, and reports on /update-dataset workflow status (Slack draft, Codex review, indicator upgrade, downstream deps).

Paired skill — keep in sync. /update-dataset is the author-side counterpart of this skill: the steps it defines are the outcomes verified here. Whenever you add, remove, or change a check in this file, check whether update-dataset/SKILL.md needs a matching author-side step (and add it in the same commit if so). The reverse also holds — see the mirror note there. The creation-side skills /create-dataset and /create-snapshot belong to the same family: the checks here (§5 snapshot fields, §6 links, §7 code clarity, §9 metadata coverage, §10 quality) also gate PRs produced by /create-dataset, so when one of them changes, check whether the create skills need a matching edit in the same commit too.

Inputs

• Optional PR number. If omitted, derive it from the current branch via gh pr list --head <branch>.

Workflow

1. PR metadata

gh pr view <num> --json title,body,isDraft,mergeable,statusCheckRollup,comments,reviews

Flag if PR description is empty (per user's standing rule: keep PR body in sync with substantial changes).

Flag 🟡 if the Summary doesn't open with a tracking-issue link (Tracks: owid/owid-issues#NNNN) — /update-dataset requires it as the first line; most data updates have a corresponding owid-issues ticket.

2. Diff and changed files

gh pr view <num> --json files --jq '.files[] | "\(.additions)+ \(.deletions)- \(.path)"'

For very large diffs (>1MB) skip gh pr diff and read the changed files directly with Read.

3. Locate the new dataset

From the changed files, identify:

• New snapshot path: snapshots/<namespace>/<new_version>/<short_name>.<ext>.dvc (a .py upload script is optional — .dvc + url_download or a local file path is enough)
• New step files: etl/steps/data/{meadow,garden,grapher}/<namespace>/<new_version>/<short_name>.{py,meta.yml}
• Old version (from dag/archive/*.yml or by grepping for the same <short_name>)

3b. Update shape — version bump vs restructure

Before running the pipeline, classify the PR. If any of the following are true, you're reviewing a restructure, not a version bump, and several downstream checks apply differently:

• The short_name changed (old version uses one name, new version uses another).
• The schema changed (wide ↔ long, different file format with a different column set, new dimensions).
• The set of policies/indicators changed substantially (splits, dropped composites, newly added areas).
• Score semantics changed (e.g. binary → continuous 0–1, units/scale changed).

When it's a restructure:

• Don't expect the auto-Indicator-Upgrader to have remapped charts. When short_names differ entirely, the upgrader has nothing to match on. Look for a hand-curated v1 title → v2 title mapping table in the PR description (or a follow-up PR thread). 🟡 if charts on the old chain are still published but no mapping plan exists.
• Don't expect a .py step copy from the old version. Step files should be authored from scratch, not produced by etl update rename. If the new step files look mechanically renamed (same logic, just version-bumped strings), flag 🟡 — the author may have skipped restructure-specific decisions.
• A chart remapped onto a successor indicator needs a config-vs-shape check. Verify its pinned selectedEntityNames exist in the successor's data (v1 regional aggregates often don't — expect the garden step to rebuild them, mirroring the retired step's method), that pinned yAxis bounds don't clip the new range, and that the subtitle doesn't still describe the old construction. Any of the three broken: 🔴 (the default view renders empty, clipped, or mislabeled).
• Slack + /latest drafts are not expected in the PR body at all. /update-dataset keeps them in the author's workbench/ (steps 9 / 9b, owned by /data-updates-comms and /data-update-announcement), so their absence from the PR is correct — don't flag it.

4. Run the full pipeline end-to-end

.venv/bin/etlr data://grapher/<namespace>/<new_version>/<short_name>
.venv/bin/etlr grapher://grapher/<namespace>/<new_version>/<short_name> --grapher --force --only

The --grapher upload is required to verify MySQL ingestion and to enable later checks (chart count, indicator upgrade verification). Confirm:

• All four steps run cleanly (snapshot pulled from S3 if .dvc is committed, otherwise re-fetched)
• MySQL upload returns a dataset id and shows variable upserts
• No errors / no empty tables

Shortcut: read DB checks off the populated staging server. OWID provisions a staging-site-<branch> server (via Buildkite) that runs the ETL chain and uploads to its MySQL. Once it's built, you can read the DB-dependent checks (chart count, attributionShort, rendered titles/Jinja coverage, indicator-upgrade, ghost variables) straight off staging instead of re-running --grapher locally — which also avoids re-triggering step side-effects (e.g. a grapher step that exports to Google Sheets). Confirm the staging ETL build actually ran and finished before trusting it: query staging-site-<branch> for the new dataset's variables (they exist) and check the owidbot PR comment shows a chart-diff block (✅) — that comment is produced after the staging build. ⚠️ Do not use the GitHub build-and-deploy check as that signal — it's the docs Cloudflare Pages deploy (.github/workflows/deploy-docs-cf.yml: make docs.build → deploys site/), with no ETL chain or Grapher upload, so a green build-and-deploy says nothing about pipeline correctness or the data DB. Reserve a local build for what the staging DB can't answer — chiefly entity-level canonicalization (§8c #2) (data lives outside MySQL). If you can't confirm staging is populated, run the pipeline locally per the steps above, and say in the report whether correctness rests on the staging build or a local run.

Review the actual PR head, not a stale local checkout. The local branch can lag origin (or carry an in-progress merge). Before reading step files locally, git fetch and confirm your tree matches the PR head — git diff HEAD origin/<branch> --stat should be empty, and gh pr view <num> --json headRefOid should match git rev-parse HEAD. gh pr view --files / gh pr diff and the staging DB always reflect origin; local Reads do not. If they diverge, sync (or review via gh pr diff) before trusting local files.

5. Snapshot field comparison

Read both .dvc files (old and new) and produce a side-by-side table for these fields:

Field	Check
title	Reasonable update if scope changed
description	Updated to reflect new source / scope
date_published	Should normally differ from date_accessed — source from url_main or the file. Equality is legitimate only as the documented fallback when no producer release date is discoverable (e.g. a scraped page carries fresh rows but no updated stamp — see /update-dataset Guardrails, "Scraped chart embeds"); expect a .dvc comment explaining it, and flag 🟡 for the author to confirm rather than 🔴. Bare equality with no rationale: ask.
date_accessed	Updated to today (or run-date)
producer / attribution_short	Same source, same values (unless changed deliberately)
citation_full / attribution	Year bumped to the new release year — etl update copies both verbatim from the old .dvc, so a stale year ships silently. 🔴 if still the old version's year.
citation_full year vs date_published year	Warn (🟡) if they differ. The year inside citation_full (and attribution) should normally match date_published's year. A mismatch is sometimes legitimate — the producer labels the release by edition rather than publish date (e.g. UN IGME's "2025 report" published 2026-03-17, so citation_full (2025) ≠ date_published 2026) — but it's just as often a stale citation the author forgot to bump. Surface it for the author to confirm; don't silently pass it.
url_main	Status check — see step 6
url_download	Status check; OK to remove if data is now fetched via API
license.url	Status check
version_producer

• Freshness check for scraped snapshots. When the snapshot .py scrapes the producer's page or a chart platform's endpoint, re-fetch the producer's page and compare against the committed snapshot — the endpoint the script reads can lag the page (e.g. a Datawrapper chart CDN trailing the page's own <noscript> data tables by a full release, so the committed snapshot silently misses the newest wave). The committed data must match the page's current tables; a missing latest row/wave is a 🔴 (see /update-dataset Guardrails, "Scraped chart embeds").

6. Verify all links

Run the HEAD-check loop from /update-dataset § 6c on every URL in the new .dvc and .meta.yml files. A curl non-2xx is a signal, not proof — Cloudflare-fronted hosts return false 404s to curl. Apply the same escalation as /update-dataset § 6c: re-check with WebFetch, then the Wayback availability API — remembering that no automated signal is decisive (hosts like BLS block both curl and WebFetch while serving browsers fine, a Wayback capture is historical evidence only, and a missing capture is non-evidence). A URL that fails all automated checks is a 🟡 — needs a human browser check (report the evidence trail: statuses, capture date or absence); escalate to 🔴 only once a browser check confirms the link is dead or the producer's site documents its retirement. A curl-only failure that WebFetch resolves is 🟢 informational.

• docs.google.com 200 ≠ publicly viewable. Google Sheets/Docs links return HTTP 200 even when they're behind a permission wall (the 200 is the "request access"/sign-in page). When a user-facing description_key/description_processing links a Google Sheet, confirm real public access with WebFetch (ask whether the page shows data or a "you need access"/sign-in wall) — curl status alone will pass a private sheet.
• Cross-check the same sheet is cited consistently. If a dataset links a "source per data point" sheet from more than one field, verify they're the same sheet ID — divergent IDs (one current, one stale) is a 🟡.
• HTTP 200 ≠ anchor exists. For URLs carrying a #fragment, run the anchor pass from /update-dataset § 6c: the fragment must match an id/name attribute in the page HTML (skip non-DOM fragments: any fragment containing = or / — text fragments, gid=, page=, hash routes like FAOSTAT's #data/FBS — plus #! hashbangs). Rule out client-side rendering and Cloudflare challenge bodies (WebFetch de-slugged-heading check) before flagging. A confirmed missing anchor on an otherwise-working page is 🟡 — the page loads, the reader just lands at the top; escalate to 🔴 only if the linked section genuinely no longer exists and the link's claim depends on it.

7. Code clarity & docs

For each step file, check:

• Snapshot script (if present): docstring explains source choice; no hidden hardcoded year/date constants without --cli-flag parametrization (or at minimum a clear update comment). Note: a .py upload script is optional — many snapshots ship with only the .dvc and a url_download. Don't flag the absence of a script.
• Meadow / garden / grapher: clear top-level docstrings; no commented-out code; no silent exception handlers
• Garden: harmonization uses paths.regions.harmonize_names(tb, ...) (the new API), not the legacy geo.harmonize_countries
• Garden assertions: sanity checks present when the step does non-trivial logic (harmonization, renames, aggregations, derivations) and not overly brittle (e.g. avoid hard-coded "X must always exceed Y" if it's not a true invariant). Check value-bound coverage per indicator type (shares in [0,1], percentages-of-a-whole in [0,100], non-negativity for level indicators, mutually exclusive share categories summing to 100 within rounding tolerance, exception sets for documented outliers) — but verify any bound against the actual data before suggesting it: "% of GDP" indicators legitimately exceed 100 (see /update-dataset §5b-bis)
• Unit-branched aggregation — verify every count sums and every rate averages. When a regional-aggregation step routes rows by a unit string — e.g. counts (unit == "Number of deaths") get summed while everything else gets population-weighted-averaged — a count series carrying a different unit label silently falls into the averaging branch and produces a meaningless regional "total". (Real case: IGME summed "Number of deaths" but "Number of stillbirths" fell through to the rates path, so regional stillbirth totals became population-weighted averages — Africa showed ~60k instead of ~1M.) Enumerate the distinct units, confirm each is routed correctly (a region's count value should be ≫ any member country's, not a mid-range average), and prefer a robust predicate (unit.startswith("Number of")) over an exact match. Catches a class of bug a green pipeline + Jinja-renders-fine review will miss.
• Blanket title-based unit scaling — audit the raw range of every affected indicator. When garden scales values by pattern-matching indicator titles (e.g. "titles containing 'share' or 'percentage' get ×100"), verify the source actually stores every matched indicator in the assumed convention: compute each matched indicator's raw min/max and flag any whose range contradicts the rule (a "fraction" with raw max ≫ 1, or a "percent" with raw max ≈ 1). (Real case: WWBI's 134 "share"-titled indicators are fractions, but its 2 "percentage"-titled wage-bill ratios are already percent — sourced from IMF FAD, not WB surveys — so a blanket ×100 shipped them 100× too large across versions.) A quick output-side check: no %-unit column's max should exceed a grounded bound (~150 for genuine percentages-of-a-whole). — magnitude bugs are inherited; judge absolute plausibility (a wage bill is not 1,242% of GDP). 🔴 if a scaled indicator's convention is contradicted by its raw range.

8. Outdated practices

Run the /check-outdated-practices skill on every new step file (snapshot, meadow, garden, and any helper modules like *_omms.py). It reads vscode_extensions/detect-outdated-practices/src/extension.ts as the single source of truth and greps the full pattern set — don't hand-maintain a copy of the patterns here, and don't eyeball helper calls and decide they look current (the geo.add_* family looks fine but is flagged). Report every hit it returns as 🟡.

Separately, the metadata/origin-stripping patterns from CLAUDE.md (pd.concat→pr.concat, pd.to_numeric/pd.to_datetime→pr.*, np.where, index.map(...), pd.DataFrame(tb) re-wrap) are not part of the extension — they're covered by the §7 code-clarity pass. Flag them there even when copy_metadata/fillna appears to mitigate.

8b. Carried-over annotations & sanity_checks (review side)

/update-dataset steps 1c+6a (annotations) and 1d+5b (sanity_checks) define the catalog/resolve procedure. As reviewer, verify the outcome:

• Annotations: scan the diff for any # NOTE: / # TODO: / # FIXME: / # HACK: / # XXX: that are unchanged from the old version. For each, confirm the PR body mentions whether the workaround is still needed, or that it was deleted with its code. Unresolved + undocumented = 🟡.
• Sanity-check log flags: grep the diff for SHOW_SANITY_CHECK_LOGS, DEBUG, LONG_FORMAT set to True. If a debug flag was left enabled, that's a 🔴 — must be reverted.
• Silent deletes: in any sanity_checks function, scan for drop, filter, tb = tb[...] — row removals that the user might miss. Make sure the PR body lists them.
• Findings surfaced, not just flags reverted: if the step has any sanity-check logic (function or inline # Sanity check block), the PR body should carry a "Sanity-check findings" section reporting what the checks said on the new data. A green pipeline run is not proof the invariants held — checks that paths.log.warning(...)/.critical(...) instead of assert/raise pass silently. If the new garden chain has logging-style checks and the PR body has no findings section, re-run the garden step (--private --force --only) and scan stdout/stderr for warning, dropped, outlier, AssertionError. Undocumented findings = 🟡; a check that newly raises on the new data = 🔴 (must be triaged with the author per /update-dataset §5b).

8c. Country harmonization audit (review side)

/update-dataset §5c defines the full audit (validate .countries.json targets against the canonical regions + income-groups catalogs, audit .excluded_countries.json, scan the garden log for the three warnings, and confirm garden-output entities are canonical). As reviewer, verify the outcome — every entity reaching Grapher must be canonical, and any that isn't must be documented in the PR body.

Run after the §4 pipeline build. Three checks:

1. Garden log warnings. Re-run the garden step capturing output and scan for the three stable warning strings:

   .venv/bin/etlr data://garden/<namespace>/<new_version>/<short_name> --private --force --only \
       > /tmp/<short_name>_harmon.log 2>&1
   rg -n "missing values in mapping\.|unused values in mapping\.|Unknown country names in excluded countries file:" /tmp/<short_name>_harmon.log
   

   missing values in mapping (source countries not in .countries.json) is the actionable one — 🟡 unless the PR body documents the gap. unused values in mapping / Unknown … excluded are informational 🟢.

2. Garden-output entities are canonical. This is the check that catches inline tb["country"] = "…" assignments and post-harmonization mutations the .countries.json review can't see. This one needs a local build — entity lists aren't in MySQL (modern grapher stores indicator data outside the DB), so make query can't answer it; build the garden step and load it with owid.catalog.Dataset("data/garden/<ns>/<v>/<short>"). Build the canonical set (regions + latest income groups) and diff against the entities actually in the built garden tables — see the Python snippet in /update-dataset §5c (Python checks #3 + #5). Note geo.REGIONS already includes the four WB income groups, so an isin(REGIONS) filter de-dups them too. Any entity in the garden output that isn't in canonical regions or income groups is 🔴 unless it's a legitimately custom source aggregate (e.g. " (ILO)"/" (WB)"-suffixed regions, BRICS, G7) that the PR body explicitly notes lives outside the canonical system.

3. Over-exclusion. If .excluded_countries.json exists, flag any entry that is a canonical region/aggregate (/update-dataset §5c Python check #4) — dropping a real country/region silently is 🟡 unless the PR body says why (e.g. source double-counts "World").

If the garden step doesn't use the harmonizer at all (no .countries.json; country assigned inline), checks #2 and #3 still apply — #2 is the only thing that catches non-canonical inline values.

8c-bis. Did another dataset update merge while this PR was open?

Cheap and worth doing on any PR more than a few days old. A branch serves its own snapshot of every other dataset, so if another update merged to master meanwhile, the branch's staging is behind on that dataset — and any chart combining both differs from production on two axes. Approving it syncs the stale config back and reverts the other update on a published chart. Nothing flags this: CI is green and the chart renders.

Check git log HEAD..origin/master --oneline for 📊 dataset commits; for each, look for charts carrying indicators from both datasets and compare every dimension's dataset version, staging vs production — not just the dimension this PR touches. The fix is merging master in and remapping the affected charts' foreign dimensions; charts using only the other dataset are out of chart-diff scope and never sync, so they're correctly left alone. 🔴 if a shared chart would regress.

8d. Empty-entity audit (optional to run — always offer it)

The author-side audit is optional to run in /update-dataset (the check-empty-entities skill sweeps every chart/MDim/explorer/narrative/gdoc surface, which can consume many tokens) — so a missing audit is not a finding. If the author ran it, verify the outcome: a selection that had data on production but none on staging is a 🔴 regression from the update; a gap identical on production is 🟡 pre-existing — it still needs fixing (chart-config edit or content follow-up on the gdoc), just not necessarily in this PR, so confirm the PR body documents it and a fix is planned.

If the author didn't run it, you MUST offer it to the user as an optional add-on to this review (name the token cost) — surfacing this offer is mandatory, never silently skip it — and recommend accepting when the risk is real: many charts remapped, hand-curated (non-auto) mappings, a restructure, or indicators whose country coverage shrank. Run the full sweep on opt-in.

For a cheap version of the same question — which surfaces carry this dataset at all, without the per-view availability checks — run find-chart-references --dataset-id <id>. It's the surface list both step-7 audits are built on, and it answers "did the author miss a surface entirely" in one query.

Either way, do a cheap manual spot-check as part of the base review: open 2–3 of the most-viewed upgraded charts on staging (SVG render is enough) and confirm their pinned entity selections still draw lines — an empty published chart is a 🔴 however it's found, and a spot-check hit is itself a reason to recommend the full sweep.

8e. Hardcoded-time-bounds audit (standard)

/update-dataset step 7 runs the check-hardcoded-years skill after all remaps — it sweeps the same surfaces as 8d (charts, map tabs, MDim views, explorer views, narrative charts, article time= embeds/links) for numeric minTime/maxTime/timelineMinTime/timelineMaxTime/map.time pins and grades each against the new indicators' latest time. It's standard, so a missing audit is a finding (🟡). Verify the outcome either way with a cheap spot-check: query staging for the dataset's chart configs, filter numeric pins client-side ("latest"/"earliest"/absent are fine), and compare against the new data's latest time. When the release added a partial year (only some series reach it), also check the inverse: single-time discrete views on partially-published indicators should carry a deliberate timelineMaxTime clamp at the last complete year (see the chart-level incomplete-latest-year exception in check-hardcoded-years) — an unguarded one renders a tolerance-backfilled, mixed-vintage bar.

• A maxTime/map.time/timelineMaxTime pin below the new latest time = the update is invisible on that surface — 🟡 unless it's deliberate (pinned year in title/subtitle/slug, narrative charts, single-year comparisons); confirm the PR documents the fix or it's already applied on staging (charts ride Chart Diff; MDim/explorer fixes must be in the ETL YAML, not DB-only — a DB-side edit is overwritten at the next rebuild).
• A pin equal to the new latest time goes silently stale next cycle (the manual-bump treadmill) — worth flagging as a suggestion.
• Deliberate pins are coupled to FAUST text — flag any fix that bumped a pinned year without updating the words around it.

9. Indicator metadata coverage & dataset block

The mandatory-fields checklist, the dataset.update_period_days requirement, and the presentation.attribution_short non-inheritance gotcha all live in /update-dataset § 6c. As reviewer, build the indicator × field matrix from that checklist and flag any missing field as 🔴.

Quick verification that presentation.attribution_short actually landed on the produced indicators (origin's value does NOT propagate):

make query SQL="SELECT shortName, attributionShort FROM variables WHERE catalogPath LIKE '%<ns>/<v>/<short_name>%'"

Any NULL row is a 🔴.

Staging query mechanics. make query re-interprets % and single quotes via shell+make and breaks on the LIKE patterns / quoted strings these checks need. Connect directly instead and feed SQL via stdin or a .sql file: mysql -h staging-site-<normalized-branch> -u owid --port 3306 -D owid < /tmp/q.sql (host = branch lowercased, /._ → -, staging-site- prefix stripped, first 28 chars — see the query: target in the Makefile). Note datasets.catalogPath has no channel prefix — it's <ns>/<v>/<short> (e.g. un/2026-06-09/igme), not grapher/un/...; match catalogPath LIKE '<ns>/<v>/%'. Batch the metadata-gap checks in one file: counts of name='', name LIKE '%<\%%'/'%{definitions%'/'%<<%' (unrendered Jinja), attributionShort IS NULL, descriptionShort IS NULL, plus a double-space/leading-space scan on name and descriptionShort (Jinja whitespace artifacts).

Distinguish regressions from inherited gaps. Before flagging update_period_days / attributionShort / description_short gaps as 🔴, check the previous version's .meta.yml — if the gap was already there, it's pre-existing (carried over by etl update), not introduced by this PR. Still report it (the fix is cheap and the standing convention wants it), but say so: a pre-existing gap is a 🟡 "worth adding while you're here", not a regression that blocks the update. A field that was set and is now missing is the real 🔴.

Additional reviewer-side metadata checks:

• dataset.owners lists the PR author. /update-dataset step 1a-bis requires the author — the person who ran the update / opened the PR — to append their canonical OWID name (an entry in the schemas/dataset-schema.json enum) to the garden .meta.yml owners: list, preserving the existing order and any # review / # backport / # fasttrack markers. Running an update makes you a contributor, so the author belongs there alongside the original owners (don't drop the originals). Verify the PR author is present: missing = 🟡; existing owners reordered or dropped = 🔴. Note the actor running this review skill is usually a different person (the reviewer) — the reviewer does not add themselves; the check is only that the author is listed. The exception is a self-review (author reviewing their own PR), where reviewer and author are the same person.
• processing_level: major must come with description_processing. Grep the new garden meta.yml for processing_level: major. Each occurrence (whether on definitions.common or per-indicator) requires a description_processing field on that same scope. 🟡 mismatch.
• Per-indicator description_processing should describe the indicator's own derivation, not just point at a shared generic note. When every aggregate indicator's description_processing is the exact same string (e.g. all four region indicators just reference {definitions.description_regions_processing} with no per-indicator detail), 🟡 flag — author should compose per-indicator sentences.
• Long-format-with-dimensions Jinja coverage. When variables are keyed by a long-column name (e.g. proportion) with <% if <dim> == "X" %>...<% endif %> blocks for title, description_short, display.name, verify every active (dim1, dim2) cell renders a non-empty value. Easiest check: read every column from the grapher dataset and assert metadata.title is non-empty.
• paths.regions.add_population(tb) / paths.regions.add_aggregates(tb, regions=[...]) auto-resolve their DAG dependencies. If the garden step loads (or ) via but never passes the dataset to anything, that's dead code — 🟡. The DAG dependency still needs to be declared either way.

10. Metadata quality skills

Run /check-metadata-typos, /check-metadata-spacing, /check-metadata-style against the new garden + grapher .meta.yml files. See /update-dataset § 6b for the full procedure (typos / spacing / style + a manual clarity checklist for general-audience readability — apply that checklist here too). Report findings as 🟡 (or 🔴 if a violation breaks rendering or makes the text outright misleading).

Also grep the metadata prose for numbers carried over from the previous release (country counts, category counts, year ranges in description_key/descriptions): validated fields are covered by checks, prose numbers are not — a panel-composition change (dropped country, new category set) silently strands them (see /update-dataset Guardrails, "Grep metadata prose"). Stale prose count: 🟡.

Five further prose checks from /update-dataset § 6b that the skills don't automate:

• Redundant user-facing text. Read each indicator's description_key as the reader gets it, together with description_short and the chart title: a bullet that restates another bullet, or the short line, at the same level of detail is padding — 🟡, with a proposed merge into the bullet that already covers it. Unpacking description_short (full definition, how it's measured, what's included) is not redundancy and must not be flagged. Watch new bullets added by this PR against the ones already there, and Jinja variants that restate the shared bullet for only some dimension values.

• Dimensional text that only holds for one breakdown. For indicators templated over a dimension (or sharing a definitions: key across variants), read each rendered variant, not just one: a caveat that the data doesn't control for X is wrong on the variant grouped by X, a scope word like "all employees" overclaims on a variant filtered to a subgroup, and a sentence about a toggle is wrong on views that exist for only one choice of that dimension. Text that misdescribes the population on some variants: 🟡 (🔴 when it contradicts what the view actually shows). Author side: /update-dataset § 6b dimension sweep.

• Methodology-attribution claims ("following guidance from …"): open the cited link and confirm it actually says that — agencies revise methodology, and a stale claim survives every link check (real case: metadata cited BEA guidance for an office-PPI deflator after BEA had switched that category to a different composite). A claim the cited page doesn't support: 🔴 (it's factually wrong reader-facing text).

• Scope qualifiers in the origin title (private-only, adults-only, market-exchange-rates-only) must surface in description_short/description_key, not only in the citation. Missing: 🟡.

• Change-magnitude claims in the PR body (revision medians, % changes): recompute at least the headline number independently from the raw old/new snapshots rather than trusting the author's diff — derived pct columns from assign/sort chains can silently misalign (a real update shipped "median +8.2%" where the true value was +14.3%). A wrong magnitude that feeds an announcement: 🔴.

10b. Adversarial data review (optional to run — always offer it)

/update-dataset § 6c-bis offers an adversarial factual review via /adversarial-data-review — optional to run because it's token-heavy (~25–45 web calls), so a missing report is not a finding; don't flag its absence. If the author didn't run it, you MUST offer it to the user as an optional add-on to this review (name the token cost; in review context it runs in that skill's spot-check scope, not the full author scope) — surfacing this offer is mandatory, never silently skip it — and recommend accepting when the update carries red flags: large unexplained value churn, an in-place source revision, a producer new to us, or editorial claims riding on specific values. Run it on opt-in.

When the PR body (or workbench/<short_name>/update-context.yml) does reference an ai/adversarial-review-<short_name>-<date>.md report, verify the outcome: its 🔴 findings must be resolved — metadata edited, or a <short_name>.corrections.yml added with reason/provider/status filled in. Then spot-check independently — don't take the report's word for it: re-verify 2–3 of its findings and 2–3 anchor values (World total + one major country, latest year) against an independent source yourself, following that skill's independence rules (a different producer measuring the same quantity; never OWID republishers or mirrors of the same producer). A value you confirm wrong that the report missed or waved through is 🔴.

11. DAG checks

The remove-and-reorder procedure is in /update-dataset § "Removing the old version & reordering the DAG". As reviewer, verify the outcome (the archive dag is regenerated separately by etl archive-dag, so don't expect the old version under dag/archive/ in this PR):

rg "<namespace>/<old_version>/<short_name>" dag/ -g "*.yml" | grep -v "^dag/archive"   # should be empty (old version removed from active dag)
rg "<namespace>/<new_version>/<short_name>" dag/ -g "*.yml" | grep -v "^dag/archive"   # should be in old slot, not at bottom

Internal version consistency (the silent-stale-data bug). etl update occasionally leaves a new step depending on an old-version dep (e.g. new garden still pointing at old meadow or old snapshot), which silently loads stale data. Verify every dep inside the new chain's block is on the new version. Read the new chain's DAG block and confirm none of its dependency lines reference <old_version>:

# Print the new chain's block and eyeball the dependency lines — none should contain <old_version>.
rg -n -A8 "<namespace>/<new_version>/<short_name>" dag/ -g "*.yml" | rg "<old_version>"   # should be empty

Any hit here is a 🔴 — the new step is wired to a stale dependency.

Visual inspection of the diff for:

• Comment headers (# Source — dataset name.) preserved above the new entries
• Indentation consistent ( # vs # is a frequent typo)
• New entries placed in the old block's slot, not orphaned at the bottom (🟡 if at the bottom)
• Both the flat and nested (compact) DAG forms are valid (the loader accepts either). The nested form (chain declared inline, grapher → garden → meadow → snapshot) is the preferred style — don't flag a nested block as wrong, and don't flag the archive edit itself (archiving is the explicitly-requested workflow step; Codex's "don't edit archived DAG" warning is a known false-positive here).

12. Downstream dependency check

Procedure in /update-dataset § "Downstream dependency check". One-liner:

rg "<namespace>/<old_version>/<short_name>" dag/ -g "*.yml" | grep -v "^dag/archive"

After excluding the dataset's own chain, any remaining hits are downstream consumers — flag 🟡 unless the PR body already documents them under a "Downstream dependencies" section.

Silent-breakage check (when consumers were repointed in this PR). Mirrors /update-dataset § "Silent-breakage check". If the PR bumps downstream consumers to the new version (rather than deferring them), a consumer can still build green while quietly losing data — a region aggregate that goes NaN, a reclassified country that disappears, a join that stops matching. A green pipeline run does not prove coverage held. Verify with the two existing instruments:

1. Downstream builds: check the buildkite/etl-automated-staging-environment status on the PR — the staging bake runs etl run ... --modified --continue-on-failure, which re-raises the first failure at the end, so any consumer crash turns the check red. A red check is a 🔴 in itself, and it also means the data-diff report under-reports (dependents of the failed step are skipped, stay stale in the catalog, and diff as unchanged) — don't accept report verdicts until the check is green. .venv/bin/etlr --modified --private --dry-run lists the affected scope locally when you need the list.
2. Value changes: open owidbot's data-diff HTML report (https://catalog.ourworldindata.org/diffs/<sanitized_branch>/data-diff.html, easiest via the full report link in owidbot's PR comment — the path keeps the branch name's dots and underscores and replaces only characters outside [A-Za-z0-9._-] (e.g. /) with - — unlike the staging subdomain, which does replace ./_) — staging (new dep) vs production (old dep) — and read its verdicts rather than scanning the diff: every red "− lost N data point(s)" entry in the "Top changes" list and every dataset with a red coverage chip is a coverage loss to triage (legitimate churn vs. a silent drop); 🔴-tier datasets (median anomaly score ≥ 15%) get a review, 🟡 a skim, 🟢 is noise. The 📝 metadata-only filter separates pure metadata edits. Locally: .venv/bin/etl diff REMOTE data/ --changed --include garden --output-html data-diff.html. (Local build+diff only for small fan-outs; at foundational-dataset scale — hundreds of downstream steps — use owidbot's hosted report and treat any local rebuild as a one-time gate, ~35 min / ~7 GB.) Then run the full-report audit probes from /update-dataset § "Silent-breakage check" over all changed datasets: structural changes / "World" rows / raw-country rows / >30%-of-rows indicators / wipe-vs-edge for every coverage loss. An entity losing all its rows is the bug signature — classically a stale pinned-country requirement nulling a whole income-group aggregate after a reclassification (build-time ValueError since the geo guard, but verify) — while sparse-edge losses are membership churn.

Any downstream build failure, dropped table/column/entity, or all-NaN series is a 🔴 (the update silently dropped data downstream) unless the author has already triaged it in the PR body. If consumers were deferred to a follow-up PR, these checks belong to that PR — here just confirm the "Downstream dependencies" list is complete.

13. /update-dataset workflow status

Verify the author completed each post-step item from /update-dataset. The procedures live there — here we just confirm the outcomes:

Item	Verify by
Indicator upgrade ran (§7)	make query SQL="SELECT COUNT(*) FROM chart_dimensions cd JOIN variables v ON cd.variableId=v.id WHERE v.catalogPath LIKE '%<ns>/<new_v>/%'" — non-zero
Explorers / MDims re-exported (§7)	Only if the DAG has export://explorers/... or export://multidim/... steps for this dataset (rg -e "export://explorers/.*/<short_name>" -e "export://multidim/.*/<short_name>" dag/ -g "*.yml"). The indicator-upgrader never touches these, so run the two staging queries from /update-dataset §7 (old-version references in explorer_variables / multi_dim_x_chart_configs) — both must return empty. A hit = 🔴, the export step wasn't re-run.
Hardcoded-time-bounds audit ran (§7 / §8e)	Standard author-side step. Spot-check per §8e: numeric maxTime/map.time/timelineMaxTime pins among the dataset's staging chart configs vs the new indicators' latest time — non-deliberate pins below it must be fixed on staging or documented in the PR body. Skipped audit or an unfixed, undocumented pin = 🟡.
Chart-diff bot result	PR comments include <!--chart-diff-start--> block ✅. A diff that shifts every historical year/region by a tiny amount is usually upstream-dataset drift (the live data was built against an older population/regions/income_groups snapshot), not a regression — don't flag it as a 🔴; the real change should be isolable by rebuilding the old version on the current catalog (see /update-dataset §5).
Scheduled-issue workflow checked (§6d)	An update-*.yml in owid/owid-issues covers this dataset — exact name update-<namespace>-<short_name>.yml, fuzzy filename, or group workflow (check ~/owid-issues/.github/workflows/, cloning the repo first if absent — a filename-only gh api …/contents listing can't verify cron/body or find group workflows); its cron is consistent with update_period_days + the release cadence evident from the PR (and fires after the producer's release window, not before); the issue body tells the next updater to run /update-dataset <short_name> (space-separated short names for group workflows). Missing workflow for a ≥annual dataset, a cron that contradicts the observed cadence, a stale body, or a workflow file missing its extension (Actions never runs it) = 🟡.

A clean Codex review has a different shape: no inline comments and zero review threads — just a single top-level "no issues" comment from chatgpt-codex-connector[bot] in the issue comments. That counts as reviewed (the threads row passes vacuously); don't flag the absence of inline threads as "review missing". Codex review state is COMMENTED (not APPROVED); fetch its inline notes via gh api repos/owid/etl/pulls/<num>/comments --paginate --jq '.[] | select(.user.login|test("codex";"i"))'. After @codex review, it usually responds in ~2–5 min — poll with a backgrounded until loop rather than blocking.

GitHub API gotchas (when posting/fetching review comments):

• GraphQL via gh api graphql -f query='...' breaks on shell quoting (the inline query gets truncated → Expected NAME parse errors). Write the query/mutation to a file and use gh api graphql -F query=@/tmp/q.graphql with variables as separate -f name=val (string) / -F name=val (typed int/file). Read bodies from a file with -F body=@/tmp/comment.md.
• Editing/adding comments inside a pending review (the author's in-progress review) needs GraphQL, not REST: REST PATCH pulls/comments/{id} 404s on a draft, and POST pulls/{n}/comments errors one pending review per pull request. Add with addPullRequestReviewThread(input:{pullRequestReviewId, path, line, side:RIGHT, body}), edit with updatePullRequestReviewComment(input:{pullRequestReviewCommentId, body}). Get the pending review id via reviews(first:50, states:PENDING) and a draft comment's node id via that review's comments.
• Disclosure on review comments. Open any inline review comment you author with > _Written by Claude <model name> — @<handle> at the wheel._ (model name = the model actually generating the content, e.g. "Sonnet 5"; handle = the human directing the work), same as PR descriptions. Skip it only on genuinely trivial one-liners (an @codex review ping).

Out of scope for review: Slack announcement and the QA hand-off (Anomalist, Chart Diff, data-diff report) are author-side concerns, not reviewer checks.

Producer-docs vs. data consistency. If the PR description notes a discrepancy between the producer's documentation (codebook, methodology page, README, release notes — whatever is available) and the actual file shipped, that's a 🟢 informational item — the author has surfaced it for producer follow-up. Don't ask them to "fix" the data to match the docs; the PR should preserve what the source shipped and flag the discrepancy.

14. Final report

Structure the review with:

1. Overview — one-paragraph summary of what the PR does
2. Pipeline test result — ✅/❌ for each step + grapher upload
3. Snapshot comparison table — old vs new
4. Indicator metadata table — fields × indicators, ✓/❌ matrix
5. 🔴 Blockers — must-fix before merge
6. 🟡 Suggestions — nice-to-have
7. 🟢 Informational — observations, no action needed
8. Workflow gaps from /update-dataset — PR description, Codex review, indicator upgrade, downstream deps, etc. (The Slack + /latest drafts live in workbench/, not the PR — don't expect them here.)
9. What's still open — carried forward from the PR body, covering the categories in .claude/docs/open-items.md plus the update workflow's fourth one (deferred to a follow-up PR — downstream repoints, old-version archiving). Re-state the full list on every re-review, not just the delta, and mark what cleared since last time.

Check the PR body doesn't leave pending work unmentioned. A PR whose description lists only what was done, while the session left content edits pending, audits unrun, or a follow-up PR's scope undefined, is missing the one artifact that survives after the chat is gone — flag it 🟡. Judge it on whether a reader can tell what's outstanding, not on whether it uses any particular headings or wording. Work that was deliberately handed off needs a locator in the body too, not just a description: an item the next person can't act on without redoing the analysis isn't handed off.

Severity rubric

• 🔴 Blocker: missing mandatory metadata field, genuinely broken link (confirmed dead by a human browser check or documented retirement — automated all-three failure alone is 🟡), failing pipeline step, breaking change to chart data, missing update_period_days, missing presentation.attribution_short, stale year in citation_full/attribution, outdated __main__ block in snapshot, DAG reference to old version that should be archived, new step wired to a stale (old-version) DAG dependency, explorer/MDim still referencing old-version variables on staging, non-canonical garden-output entity that isn't a documented custom aggregate, sanity check that newly raises on the new data, downstream consumer that fails to build or silently lost coverage after a same-PR repoint (a "− lost N data point(s)" entry / red coverage chip in the data-diff report) without triage in the PR body, upgraded view (chart, MDim view, narrative chart, or article link) whose pinned entity selection has data on production but none on staging (renders empty), metadata claim contradicted by the producer's own documentation, data value confirmed wrong by two independent sources without a corrections.yml entry or PR-body triage
• 🟡 Suggestion: brittle assertion, hardcoded year that should be dynamic, duplicated grapher meta.yml that could be removed, non-blocking style issues, undocumented sanity-check findings, phantom sort: labels with no backing value, over-exclusion of a canonical region, undocumented missing values in mapping countries, count series mis-routed into a rate/average aggregation branch, citation_full year ≠ date_published year (verify), pre-existing inherited metadata gap (update_period_days/attributionShort/description_short already missing in the prior version), PR author missing from dataset.owners, missing tracking-issue link in the PR body, owid-issues scheduled workflow missing, mis-timed, or not pointing at /update-dataset, pre-existing empty-entity gap surfaced by the optional check-empty-entities audit or a spot-check (identical on production — needs a chart-config or gdoc fix via follow-up, just not necessarily in this PR), hardcoded-time-bounds audit (check-hardcoded-years) skipped by the author or a non-deliberate maxTime/map.time/timelineMaxTime pin below the new data's latest time left unfixed and undocumented (the update is invisible on that surface), "verify manually" item from an adversarial-review report left untriaged (only applies when the author ran the optional review)

Notes

• The /review skill is for general PR review — this skill is the dataset-specific superset.
• If the user explicitly asks to skip the pipeline run (e.g. "don't run it, just look"), still do steps 1–3 and 5–13, but skip step 4 and note that pipeline correctness is unverified.
• Always include the --grapher flag when running the grapher step end-to-end — without it, MySQL ingestion is not exercised and indicator metadata in the DB is not verified.