What does this skill do, and when should you use it?
This skill helps users discover Earth2Studio models, data sources, and examples for weather and climate workflows. It filters candidates by task type, region, forecast horizon, variables, GPU memory, and deterministic versus ensemble requirements. It requires consulting NVIDIA’s live Earth2Studio documentation and checking compatibility through model input variables and data-source lexicons. It does not install software, download data, write inference code, train models, or debug runtime failures.
It interprets the user’s requirements, fetches relevant model, data-source, and gallery pages from nvidia.github.io/earth2studio, reads Class, Region, VRAM, and Release badges, and produces a short model shortlist. It compares a model’s input_coords variables with a data source’s lexicon VOCAB to assess compatibility, then recommends one to three relevant examples with documentation links. It may ask up to three targeted follow-up questions when requirements are incomplete.
- A weather researcher needs to compare medium-range forecast models against available GPU memory.
- A climate application developer needs diagnostic models and examples for downscaling, super-resolution, or derived variables.
- A data engineer needs to verify whether a reanalysis, operational forecast, or observation source covers a model’s required variables.
- A meteorological modeling team needs to choose between deterministic and ensemble workflows and find matching Earth2Studio examples.
- A researcher needs to locate suitable models for data assimilation, seasonal, or subseasonal forecasting.
How do you install this skill?
- Core recommendations depend on live nvidia.github.io pages; page changes, outages, or network inaccessibility may make results incomplete or unavailable.
- Lexicon compatibility checks may be approximate when only documentation is accessible and should not be treated as runtime validation.
- The evaluation report does not expose CI workflows, test implementations, or raw per-task evidence, so its benchmark scores should be treated as limited support.
- The skill does not handle installation, downloading, inference code, or runtime debugging; additional skills may be required for follow-on work.
- Network access
Install the named skill with the skills CLI command documented in the repository README: npx skills add nvidia/skills --skill earth2studio-discover --yes. To target Codex explicitly, use: npx skills add nvidia/skills --skill earth2studio-discover --agent codex --yes. The source does not document other installation methods or runtime requirements.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/earth2studio-discover" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- I need a global 10-day ensemble forecast with a 48GB GPU; find suitable models, compatible data sources, and examples.
After installation, ask the agent for a concrete Earth2Studio discovery task, for example: “I need a global 10-day ensemble forecast with a 48GB GPU; find suitable models, compatible data sources, and examples.” The skill will consult live documentation using the task, region, horizon, variable, and hardware constraints, then return a linked shortlist. Do not use it for Earth2Studio installation, data downloads, or inference-code authoring.
What are this skill's strengths and limitations?
- Prioritizes live documentation, reducing the risk of relying on an outdated static catalog.
- Uses Class, Region, VRAM, and Release badges to filter model candidates.
- Checks data-source compatibility through lexicon VOCAB and model input_coords variables.
- Covers discovery for forecasting, downscaling, ensemble generation, data assimilation, seasonal work, and derived diagnostics.
- Clearly separates discovery from installation, inference coding, training, and debugging.
- Requires network access to NVIDIA’s Earth2Studio documentation.
- Results depend on documentation updates, missing badges, and stable page URLs.
- Full lexicon compatibility checks may require access to Earth2Studio source code.
- The skill does not install software, download data, or produce inference code.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Earth2Studio Discover this page ✓ NVIDIA · Official | 56 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Earth2Studio Forecast Wrapper Builder ✓ NVIDIA · Official | 50 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Earth2Studio Deterministic Forecast ✓ NVIDIA · Official | 53 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Earth2Studio Diagnostic Builder ✓ NVIDIA · Official | 56 · Use with care | ★ 3.8k | today | Apache-2.0 |
| Earth2Studio Installation Guide ✓ NVIDIA · Official | 56 · Use with care | ★ 3.8k | today | Apache-2.0 |
How did FollowSkills review this skill?
The skill is explicitly limited to discovery and recommendations; it does not perform installation, downloads, or inference code, and it relies mainly on reading public documentation with source citations. It does not explain user-data handling, network-request privacy, confirmation, rollback, or dependency security, so points are deducted.
The workflow covers input extraction, document selection, badge filtering, compatibility checks, and troubleshooting clearly. However, it depends on live external documentation and lacks executable reproduction, automated tests, and comprehensive abnormal-input feedback; static calibration keeps this at or below 10.
Audience, weather/climate scenarios, task categories, hardware constraints, output structure, and non-fit boundaries are reasonably clear. Chinese-language interaction is not specified, and core functionality depends on potentially mainland-China-inaccessible overseas documentation, so points are deducted.
The skill includes a version, Apache-2.0 license, author, purpose, limitations, troubleshooting, and ownership information. It lacks separate Instructions and Examples sections, a clear change log or update path, and complete installation/dependency notes, so points are deducted.
It can guide model, data-source, lexicon-compatibility, and example discovery, with a directly usable proposed output structure. However, recommendations depend on live pages and the skill provides no representative outputs or statically reproducible results; the score is conservatively bounded by static calibration.
It provides pinned documentation categories, a lexicon verification method, evaluation tasks, and a benchmark report. The benchmark results are primarily claims in repository files, without visible CI, committed key-path tests, or independent corroboration, so only limited credit is awarded.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →