| name | climate-update |
| description | Run OWID's monthly climate data update. Bumps all updateable climate-namespace datasets to one common version in a single PR with one announcement, skipping the frozen sources. Use when the user wants to update climate data or run the monthly climate update. For a wildfires-only refresh during fire season, use the wildfires-update skill instead. |
| metadata | {"internal":true} |
Climate update (monthly umbrella update)
All OWID climate data is updated once a month as a single batch: every updateable
climate-namespace dataset is bumped to one common version date, in one PR, with one
Slack announcement. This skill owns which datasets, in what order, and the
single-PR / single-announcement discipline. It delegates the per-dataset mechanics
(snapshot → meadow → garden → grapher, diffs, metadata checks, chart upgrade) to
/update-dataset — run each chain through that flow on
the same branch.
The matching reminder lives in owid-issues/.github/workflows/update-climate.yml
(one monthly issue on the 10th). Wildfires is the one dataset with a reminder of its
own on top of that: update-climate-wildfires.yml fires weekly from May to the end of
September.
Modes
- Full monthly update (default): bump every updateable dataset below to a common
new version date.
- Wildfires-only subset: bump just
weekly_wildfires (and optionally
yearly_burned_area) to a new version, for the weekly in-season refresh. Follow
, which covers the source's weekly
grid, its provisional last data point, and the snapshot's silent per-country failure
mode. Come back here only if you are bumping the whole batch.
