midforecastlisted
Install: claude install-skill seokhoonj/data-go-kr
# data-go-kr — 기상청 중기예보
Fetch the 4-to-10-day outlook for a forecast region. Where the `weather` skill covers the
next ~3 days on a 5km grid, this covers days 4-10 for a coarser 예보구역 named by a
`region` code. The rows are **wide** -- one row per region, a column per forecast day:
| operation | 예보 | clean columns |
|---|---|---|
| `land` | 중기육상예보 | `region`, `precip_prob_4am`..`precip_prob_10` (강수확률 %, int), `sky_4am`..`sky_10` (날씨 문구) |
| `temperature` | 중기기온예보 | `region`, `temp_min_4`..`temp_min_10`, `temp_max_4`..`temp_max_10` (℃, int) |
Days 4-7 split into morning/afternoon (`_4am`/`_4pm` .. `_7am`/`_7pm`); days 8-10 are
single (`_8`..`_10`). A day the announcement does not cover is `None` (the 1800 announcement
starts at day 5, the 0600 one reaches day 4).
## Prerequisite
```
pipx install data-go-kr # or: pip install data-go-kr
```
A data.go.kr **decoding** key must be configured (env `DATA_GO_KR_API_KEY` or
`~/.config/data-go-kr/credentials.json`), and the 중기예보 dataset (service 1360000,
MidFcstInfoService) applied for (활용신청) on that account.
## Running
```
data-go-kr midforecast land --region REGID --base-time YYYYMMDDHHMM [--json]
data-go-kr midforecast temperature --region REGID --base-time YYYYMMDDHHMM [--json]
```
- `--region`: the 예보구역코드. 육상 uses a 광역 code (`11B00000` 서울/인천/경기, `11H20000`
부산/울산/경남, ...); 기온 uses a 도시 code (`11B10101` 서울, `11H20201` 부산, ...). Ask if
unsure; do not guess a code.
- `--base-time`: the 발표시각 as `YYYYMMDDHHMM`, iss