Shapefile

Use this skill for practical Shapefile handling, debugging, and migration planning.

Workflow

Confirm whether the task is inspection, repair, conversion, schema explanation, or batch cleanup.
Identify which companion files are present: .shp, .shx, .dbf, .prj, and optionally .cpg.
Check CRS, encoding, geometry type, and field-name limits before suggesting edits or conversions.
Prefer writing outputs to a new path; avoid in-place mutation unless the user explicitly asks.
If the dataset is large or the workflow is repeated, prefer a deterministic CLI path through qgis.

Treat a Shapefile as a multi-file dataset, not a single .shp file.
Missing .shx or .dbf usually means the dataset is incomplete.
Missing .prj means CRS is unknown, not automatically WGS84.
DBF field names are short and legacy-oriented; watch for truncation and type loss.
Text encoding may depend on .cpg; without it, non-ASCII text can decode incorrectly.
Shapefile is poor for long field names, rich types, large text, and modern metadata needs.

File opens but attributes are broken: verify .dbf exists and encoding is correct.
Features draw in the wrong place: verify .prj exists and the CRS was not guessed incorrectly.
Import succeeds but schema looks wrong: check field-name truncation and DBF type limitations.
Multipart or invalid geometry surprises: inspect geometry type before conversion or editing.
One file was copied alone: remind that the full sidecar set must travel together.

For interchange with legacy systems, preserve the Shapefile but document its CRS and encoding explicitly.
For ongoing editing or richer schema, migrate to GPKG.
For lightweight web interchange, convert to GeoJSON only when the CRS and precision tradeoffs are acceptable.
For deterministic file conversion or batch fixes, hand off execution to qgis.

Use this skill for Shapefile structure, limitations, troubleshooting, and migration guidance.
For general CRS selection, use project or wgs84.
For actual file-based conversion, reprojection, clipping, or repair commands, use qgis.
For web map rendering or tiles, use mapbox or cesium as appropriate.

Keep examples generic and portable.
Do not hardcode private data paths, private datasets, or machine-specific environments.
For clawhub.ai publication, keep the skill concise, reproducible, and semver-friendly; keep detailed patterns in references.

Read {baseDir}/references/patterns.md when you need concrete Shapefile validation, packaging, or conversion guidance.