agentsclimarketplace

Open meteo weather forecast

Skill runxhq/runx/skills/open-meteo-weather-forecast

Resolve a place and fetch global forecast and air-quality evidence through governed, keyless Open-Meteo HTTP calls.From its SKILL.md

Install
npx -y skills add runxhq/runx --skill open-meteo-weather-forecast

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

3.2 KB, 700 tokens by cl100k_base, as published. Nobody here has run it

Open-Meteo Weather Forecast

Fetch global place, forecast, and air-quality evidence through runx's governed HTTP front. The skill never performs an ad-hoc fetch. Each stage is a bounded, read-only HTTP runner whose endpoint, inputs, status, authority scope, and output packet are recorded in the sealed receipt. Open-Meteo requires no API key and covers the whole globe, so this is the worldwide counterpart to the US-only nws-weather-forecast provider.

Procedure

  1. Run locate with a city or place name.
  2. Select the intended result and preserve its latitude, longitude, timezone, and country as a typed provider-evidence packet.
  3. Run the default forecast runner with the selected coordinates.
  4. Require a 2xx response and preserve the provider URL, timezone, generation time, current conditions, daily forecast, and receipt references.
  5. Treat the result as context only. Any notification, scheduling change, or operational mutation requires a separate skill and authority gate.

Inputs

  • name (locate, required): city or place to resolve.
  • count (locate, optional): maximum candidate results, default 5.
  • latitude (forecast, required): decimal latitude.
  • longitude (forecast, required): decimal longitude.
  • forecast_days (forecast, optional): days requested, default 3.

Output packets

Two runners are exposed. The default main runner runs locate then forecast. The air_quality runner runs locate then air_quality. Each graph step produces a typed sealed packet:

  • locate produces Open-Meteo geocoding provider evidence.
  • forecast produces Open-Meteo forecast provider evidence.
  • air_quality produces Open-Meteo air-quality provider evidence (PM2.5, PM10, carbon monoxide, nitrogen dioxide, sulphur dioxide, ozone, and the European and US air-quality indices).

Preserve provider metadata and receipt references when passing either packet to another skill. Do not convert provider evidence into high-stakes advice.

Failure, retry, and stop conditions

  • Ambiguous or missing place: return needs_input; do not silently choose a similarly named location.
  • Invalid coordinates or forecast-days value: return needs_input.
  • Non-2xx response, timeout, or rate limit: return needs_more_evidence, preserve the failure in the receipt, and retry only under the caller's retry policy.
  • Stale or incomplete provider response: return needs_more_evidence.
  • Life-safety, medical, aviation, maritime, or emergency use: refuse and direct the caller to official channels.
  • Mutation requested from forecast evidence: stop after the sealed read and require a separate authority-scoped action skill.

Worked example

  1. Run locate with name: Shanghai.
  2. Select the result for Shanghai, China.
  3. Run forecast using the sealed latitude and longitude with forecast_days: "3".
  4. Preserve both sealed receipts with the returned forecast evidence.

Credits

Combined from two community contributions: the forecast skill (#63, @why123-boop) and the air-quality snapshot (#64, @yanziwei).

What ships with it: 3 files

5.1 KB alongside SKILL.md

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.