Skip to content

K-Dense-AI/scientific-agent-skills / skills/geomaster

geomaster

Verified

Supports geospatial research workflows for remote sensing, vector and raster GIS, spatial statistics, terrain and network analysis, and machine learning for Earth observation. Use when processing satellite imagery, aligning coordinate systems and raster grids, accessing STAC catalogs, analyzing geospatial time series, or implementing scientific GIS workflows in Python, R, Julia, JavaScript, C++, Java, Go, or Rust.

references/review.md

View on GitHub
# Review scope and sources

Reviewed 2026-10-01. This skill contains broad geospatial references, not one package.
Every preexisting reference was read. Invalid scientific shortcuts were corrected
or replaced with bounded workflows; useful API fragments were retained where verified.

## Evidence boundaries

- The isolated Python 3.13 suite passed 43 tests. Synthetic checks cover masked
  indices/calibration, terrain with known planes, raster output metadata, classification
  masks/labels/CRS, refusal to overwrite existing files, and corrected local recipes.
- Public STAC GET searches at Planetary Computer, Earth Search and CDSE each returned
  two Sentinel-2 L2A items plus next-page links. Separate PySTAC Client searches followed
  pagination and bounded results to three items per catalog, checking collection/cloud
  filters. These are metadata tests, not imagery or authenticated download tests.
- A local ODC-STAC 0.5.3 load preserved raw DN values despite asset scale/offset metadata;
  callers must calibrate explicitly once. Real local COG, GeoParquet and Zarr 2 round trips,
  Dask raster/vector, Fiona, PyKrige, ESDA, MovingPandas, OSMnx and watershed checks passed.
- Desktop QGIS/ArcPy/GRASS/SAGA, native GDAL, R/Julia/Java/Go/Rust, GPUs and real-scene
  ML/physical calibration were not executed. Their examples are illustrative and
  use referenced current interfaces; source review is not runtime validation.
- No authenticated Earth Engine/CDS/Google/Mapbox/OpenWeather calls, service mutations,
  cloud exports, training jobs, commercial image orders or pretrained weights were used.

## Upstream contracts checked

| Surface | Source and result |
|---|---|
| Core GIS | [GeoPandas](https://geopandas.org/en/stable/docs/reference.html), [PyProj](https://pyproj4.github.io/pyproj/stable/api/transformer.html), [Rasterio masks](https://rasterio.readthedocs.io/en/stable/topics/masks.html): axis order, CRS transforms, nodata and array dimensions |
| Raster/COG | [GDAL COG](https://gdal.org/en/stable/drivers/raster/cog.html), [viewshed](https://gdal.org/en/stable/programs/gdal_viewshed.html): actual driver layout, projected coordinates and nodata limitation |
| STAC | [PySTAC Client](https://pystac-client.readthedocs.io/en/latest/api.html), [CDSE](https://documentation.dataspace.copernicus.eu/APIs/STAC.html), [Earth Search](https://element84.com/earth-search/): bounded iterator, pages, collection and asset names |
| Planetary Computer | [SDK signing source](https://github.com/microsoft/planetary-computer-sdk-for-python/blob/main/planetary_computer/sas.py): account/container SAS endpoint, expiry and item modifier |
| Sentinel/Landsat | [S2 SR](https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED), [Landsat L2](https://developers.google.com/earth-engine/datasets/catalog/LANDSAT_LC08_C02_T1_L2), [normalizedDifference](https://developers.google.com/earth-engine/apidocs/ee-image-normalizeddifference), [SAR](https://developers.google.com/earth-engine/guides/sentinel1): QA, reflectance scales, thermal products, negative values and power/dB |
| Climate APIs | [CDS](https://cds.climate.copernicus.eu/how-to-api): PAT configuration, current API URL and data_format request |
| Other HTTP APIs | [Google](https://developers.google.com/maps/documentation/geocoding/guides-v3/requests-geocoding), [Mapbox](https://docs.mapbox.com/api/search/geocoding/), [OpenWeather](https://openweathermap.org/api/current), [Overpass](https://wiki.openstreetmap.org/wiki/Overpass_API): current routes, query keys, response status/shape and traversal limits |
| OSM services | [OSMnx](https://osmnx.readthedocs.io/en/stable/user-reference.html), [Nominatim policy](https://operations.osmfoundation.org/policies/nominatim/), [tile policy](https://operations.osmfoundation.org/policies/tiles/): 2.x names, free-flow times and service use boundaries |
| Cloud arrays | [ODC](https://odc-stac.readthedocs.io/en/latest/_api/odc.stac.load.html), [Rioxarray](https://corteva.github.io/rioxarray/stable/rioxarray.html), [Xarray Zarr](https://docs.xarray.dev/en/stable/generated/xarray.Dataset.to_zarr.html), [Dask-GeoPandas](https://dask-geopandas.readthedocs.io/en/stable/docs/reference/api/dask_geopandas.GeoDataFrame.spatial_shuffle.html): chunks, spatial shuffle, explicit store format |
| Specialist local APIs | [MovingPandas](https://movingpandas.readthedocs.io/en/main/api/trajectorysplitter.html), [PyKrige](https://geostat-framework.readthedocs.io/projects/pykrige/en/stable/generated/pykrige.ok.OrdinaryKriging.html), [SciPy MILP](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.milp.html): current constructors, variance and integer constraints |
| ML/GPU | [group splits](https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.GroupShuffleSplit.html), [SHAP](https://shap.readthedocs.io/en/latest/generated/shap.plots.beeswarm.html), [cuSpatial](https://docs.rapids.ai/api/cuspatial/stable/api_docs/spatial/), [PyTorch AMP](https://docs.pytorch.org/docs/stable/amp.html): output types, holdouts and current namespaces |
| Native GIS | [QGIS](https://docs.qgis.org/4.2/en/docs/pyqgis_developer_cookbook/plugins/plugins.html), [ArcGIS viewshed](https://doc.esri.com/en/arcgis-pro/latest/tool-reference/spatial-analyst/viewshed.html), [GRASS](https://grass.osgeo.org/grass-stable/manuals/r.watershed.html), [SAGA](https://saga-gis.sourceforge.io/saga_tool_doc/9.12.1/ta_morphometry_0.html): actual processing contracts |
| Languages | [sf](https://r-spatial.github.io/sf/reference/geos_unary.html), [terra](https://rspatial.github.io/terra/reference/extract.html), [ArchGDAL](https://yeesian.com/ArchGDAL.jl/latest/geometries/), [GeoStats](https://juliaearth.github.io/GeoStatsDocs/stable/interpolation/), [orb](https://pkg.go.dev/github.com/paulmach/orb/geo), [GeoRust](https://docs.rs/geo/latest/geo/), [Turf](https://turfjs.org/docs/api/interpolate): corrected method names, units and ownership boundaries |

The executed core stack included GeoPandas 1.2.0, Shapely 2.1.2, PyProj 3.8.0,
Rasterio 1.5.2, Rioxarray 0.23.0, Xarray 2026.9.0, Dask 2026.8.0,
Dask-GeoPandas 0.5.0, scikit-learn 1.9.1, OSMnx 2.1.1, PySTAC Client 0.9.0,
Planetary Computer 1.0.0, ODC-STAC 0.5.3, PyArrow 25.0.1, Zarr 3.4.0,
MovingPandas 0.23.0, PyKrige 1.7.3, ESDA 2.10.0 and rio-cogeo 7.0.3.
Upstream warnings concerned Rioxarray affine multiplication deprecation and the
absent optional Stone Soup smoother; the tested MovingPandas operations do not use it.

The core environment intentionally excludes heavy GPU/desktop/specialist stacks.
A package being importable or a synthetic check passing does not validate real-world
scientific accuracy, field transfer, calibration or service availability.
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall