earth2studio-discover
Find Earth2Studio models, data sources, and examples for a weather/climate use case. Do NOT use for writing inference code, downloading data, or installation.
Security Assessment
About earth2studio-discover
Earth2Studio Discover helps users find the right Earth2Studio models, data sources, and gallery examples for a weather or climate use case. It solves the problem of navigating a large, frequently updated catalogue by matching a user's task, region, temporal scale, variables, and hardware constraints to suitable components, verified against live documentation rather than a memorized list that would quickly go stale.
The skill's core principle is discovery from live docs: it always fetches the relevant model, data source, and examples pages before recommending, uses badge metadata (Class such as NWC/DS/MR/S2S/DA/CM, Region, recommended VRAM, and release year) to filter candidates, and verifies data-source-to-model compatibility through the lexicon system. Its interaction protocol first understands the user's problem (task type, region, temporal scale, variables, GPU/VRAM, deterministic versus ensemble) with well-phrased follow-ups capped at three questions, then fetches the appropriate prognostic, diagnostic, or data-assimilation model pages, extracts badge metadata, and presents a filtered short-list with cited doc URLs so users can explore further. It is deliberately scoped to discovery and does not write inference code, download data, or handle installation.
It targets developers, engineers, and researchers choosing components for weather and climate workflows. Typical use cases include comparing models by VRAM budget, choosing a forecast class for nowcast versus medium-range versus seasonal work, finding compatible data sources via lexicons, and locating gallery examples for downscaling, ensemble generation, or data assimilation.
FAQ
What does this skill help me do?
Identify suitable Earth2Studio models, data sources, and gallery examples for a weather/climate task, filtered by class, region, VRAM, and release, with citations to live documentation.
What are the prerequisites?
Internet access to fetch live documentation pages from nvidia.github.io and familiarity with the Earth2Studio badge system (Class, Region, VRAM, Release).
Why does it fetch live docs instead of using a built-in list?
Earth2Studio adds models, data sources, and examples every release and reorganizes examples, so any static list would rot; the skill always fetches current pages and uses badge metadata to filter.
How does it check that a data source works with a model?
It verifies data-source-to-model compatibility through the lexicon system, ensuring the model's input variables exist in the candidate data source's vocabulary.
What is it not intended for?
It is discovery-only: it does not write inference code, download data, or perform installation. Those tasks belong to the sibling forecast, data-fetch, and setup skills.
Install earth2studio-discover
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
nvidia/skills