Add skill: mapbox-tiling-service-patterns

[mattpodwysocki]

Summary

Mapbox Tiling Service (MTS) is how developers publish custom vector tilesets at scale, but the recipe authoring, validation, and publishing workflow is complex and Mapbox-specific. AI assistants have essentially no useful knowledge here.

Topics to cover

• Recipe authoring - layer definitions, feature filters, zoom level configuration, simplification
• Tileset sources - uploading GeoJSON/CSV/Shapefile as a tileset source via the API
• Publishing and updating - creating, publishing, and updating tilesets; tracking job status
• Validation - common recipe errors, size limits, attribute type constraints
• Using tilesets in styles - referencing MTS tilesets in Mapbox Studio and GL JS
• Update pipelines - automating tileset updates when source data changes (CI/CD patterns)
• Cost and performance - understanding processing time, tile counts, and when MTS is the right choice vs. uploading GeoJSON directly

Key patterns a skill should teach

• Minimal viable recipe structure for common use cases (point data, polygon data, line data)
• How to check tileset processing status and handle failures
• The source-layer name convention and how it maps to recipe layer IDs
• When to use MTS vs. the Uploads API vs. serving your own tiles

Why this matters

MTS is the right solution for large or frequently-updated datasets, but developers often reach for GeoJSON uploads instead because they don't know the workflow. A skill here would steer developers toward the correct tool for their data scale.