isaaccorley/geospatial-skills

Python

75

20 commits

updated Aug 7, 2026

See the code

README

geospatial-skills

A collection of separately installable geospatial SKILL.md packages for AI coding agents.

Site: https://isaac.earth/geospatial-skills/

Install

The recommended path is skills.sh — a single-line install that works across most agents:

# install everything
npx skills add isaaccorley/geospatial-skills

# or install one skill at a time
npx skills add isaaccorley/geospatial-skills/<skill>   # e.g. gdal, geozarr, pmtiles-pipeline

Other paths:

# Claude Code plugin
/plugin install <skill>@geospatial-skills

# Or copy a skill directly into your shared skills folder
cp -R skills/<skill> ~/.agent/skills/<skill>

Skills

SkillWhat it does
gdalGDAL command-line workflows for raster and vector data
geoparquet-validationgpio-focused GeoParquet inspection, validation, and distribution
geozarrGeoZarr metadata conventions for georeferenced Zarr stores
tesseraTESSERA embedding downloads via the geotessera CLI / Python / R libraries
geospatial-viewersviewtif / viewgeom / viewinline quick-look CLIs (run via uvx)
geospatial-frontendMap-centric demo webapps with MapLibre globe + DuckDB-WASM
parallel-geospatialParallelize continental-scale vector/raster ops on one fat node
pmtiles-pipelineRaster predictions to a global PMTiles archive via sharded tippecanoe
zarr-multiscalesBuild multiscale pyramids into existing Zarr stores without losing shards
planet-bulk-downloadBulk PlanetScope downloads via COG range reads, retries, and resume
portolan-cliPublish and manage cloud-native STAC catalogs with the Portolan CLI
portolan-bootstrapBootstrap a full Portolan catalog from WFS, ArcGIS, or local files
reading-portolanExplore, query, and visualize Portolan STAC catalogs end to end
portolan-thumbnailsServer-side collection thumbnails with chiitiler + MapLibre GL Native
register-catalogList a published catalog in the Portolan registry via pull request
sourcecoopPublish geospatial data to Source Cooperative with Portolan
search-stacSearch and download Planetary Computer imagery via STAC
overture-dataDownload Overture Maps buildings, places, roads, and more
detect-objectsRun pre-trained detection models on satellite and aerial imagery

Provenance and licenses

Skills come from three sources; imported skills carry their upstream license file in their own directory and keep upstream content verbatim:

  • authored here (Apache-2.0): gdal, geoparquet-validation, geozarr, tessera, geospatial-viewers, geospatial-frontend, parallel-geospatial, pmtiles-pipeline, zarr-multiscales, planet-bulk-download
  • imported from portolan-sdi/portolan-skills (Apache-2.0, Portolan SDI): portolan-cli, portolan-bootstrap, reading-portolan, portolan-thumbnails, register-catalog, sourcecoop
  • imported from opengeos/geoai-skills (MIT, Qiusheng Wu): search-stac, overture-data, detect-objects — cross-references were repointed from the upstream plugin namespace to this catalog's skills, and error paths now give inline pip install instructions

Layout

skills/
  <skill>/                 SKILL.md  [references/]  [scripts/]  [LICENSE]
plugins/
  <skill>/                 .claude-plugin/plugin.json  skills/<skill>/  (mirror)
docs/
  skills/<skill>.md        one docs page per skill
overrides/
  home.html  skill.html

Docs

uv sync --group dev
uv run mkdocs serve   # local preview
uv run mkdocs build   # build to ./site

Contributor workflow, CI, and docs deploy notes live in .github/CONTRIBUTING.md.

Contributors

isaaccorley

18 commits

avsm

1 commits

jakenotjay

1 commits

isaaccorley/geospatial-skills

Python

75

20 commits

updated Aug 7, 2026

See the code

README

geospatial-skills

A collection of separately installable geospatial SKILL.md packages for AI coding agents.

Site: https://isaac.earth/geospatial-skills/

Install

The recommended path is skills.sh — a single-line install that works across most agents:

# install everything
npx skills add isaaccorley/geospatial-skills

# or install one skill at a time
npx skills add isaaccorley/geospatial-skills/<skill>   # e.g. gdal, geozarr, pmtiles-pipeline

Other paths:

# Claude Code plugin
/plugin install <skill>@geospatial-skills

# Or copy a skill directly into your shared skills folder
cp -R skills/<skill> ~/.agent/skills/<skill>

Skills

SkillWhat it does
gdalGDAL command-line workflows for raster and vector data
geoparquet-validationgpio-focused GeoParquet inspection, validation, and distribution
geozarrGeoZarr metadata conventions for georeferenced Zarr stores
tesseraTESSERA embedding downloads via the geotessera CLI / Python / R libraries
geospatial-viewersviewtif / viewgeom / viewinline quick-look CLIs (run via uvx)
geospatial-frontendMap-centric demo webapps with MapLibre globe + DuckDB-WASM
parallel-geospatialParallelize continental-scale vector/raster ops on one fat node
pmtiles-pipelineRaster predictions to a global PMTiles archive via sharded tippecanoe
zarr-multiscalesBuild multiscale pyramids into existing Zarr stores without losing shards
planet-bulk-downloadBulk PlanetScope downloads via COG range reads, retries, and resume
portolan-cliPublish and manage cloud-native STAC catalogs with the Portolan CLI
portolan-bootstrapBootstrap a full Portolan catalog from WFS, ArcGIS, or local files
reading-portolanExplore, query, and visualize Portolan STAC catalogs end to end
portolan-thumbnailsServer-side collection thumbnails with chiitiler + MapLibre GL Native
register-catalogList a published catalog in the Portolan registry via pull request
sourcecoopPublish geospatial data to Source Cooperative with Portolan
search-stacSearch and download Planetary Computer imagery via STAC
overture-dataDownload Overture Maps buildings, places, roads, and more
detect-objectsRun pre-trained detection models on satellite and aerial imagery

Provenance and licenses

Skills come from three sources; imported skills carry their upstream license file in their own directory and keep upstream content verbatim:

  • authored here (Apache-2.0): gdal, geoparquet-validation, geozarr, tessera, geospatial-viewers, geospatial-frontend, parallel-geospatial, pmtiles-pipeline, zarr-multiscales, planet-bulk-download
  • imported from portolan-sdi/portolan-skills (Apache-2.0, Portolan SDI): portolan-cli, portolan-bootstrap, reading-portolan, portolan-thumbnails, register-catalog, sourcecoop
  • imported from opengeos/geoai-skills (MIT, Qiusheng Wu): search-stac, overture-data, detect-objects — cross-references were repointed from the upstream plugin namespace to this catalog's skills, and error paths now give inline pip install instructions

Layout

skills/
  <skill>/                 SKILL.md  [references/]  [scripts/]  [LICENSE]
plugins/
  <skill>/                 .claude-plugin/plugin.json  skills/<skill>/  (mirror)
docs/
  skills/<skill>.md        one docs page per skill
overrides/
  home.html  skill.html

Docs

uv sync --group dev
uv run mkdocs serve   # local preview
uv run mkdocs build   # build to ./site

Contributor workflow, CI, and docs deploy notes live in .github/CONTRIBUTING.md.

Contributors

isaaccorley

18 commits

avsm

1 commits

jakenotjay

1 commits

Languages

Python

85.8%

HTML

14.2%