portaljs-new-portal

$npx mdskill add datopian/portaljs/portaljs-new-portal

Scaffold a new PortalJS data portal from a brief.

  • Starts a brand-new data portal project from scratch.
  • Depends on Node.js, npm, and network access for template fetch.
  • Interviews user if brief is thin, then confirms before scaffolding.
  • Copies template, substitutes tokens, installs deps, and verifies with type check.

SKILL.md

.github/skills/portaljs-new-portalView on GitHub ↗
---
name: portaljs-new-portal
description: Scaffold a new PortalJS data portal from a brief. Copies the canonical template from examples/portaljs-catalog and substitutes project tokens. Use when starting a brand-new data portal project from scratch.
allowed-tools: Read, Write, Edit, Bash(npm:*), Bash(npx:*), Bash(cp:*), Bash(mkdir:*), Bash(git:*)
version: 1.0.0
author: Datopian <[email protected]>
license: MIT
compatibility: Claude Code with PortalJS portals (Next.js 14, React 18, Node 18+). Runs from any project via the plugin, a personal ~/.claude/commands install, or a portaljs clone.
tags:
  - portaljs
  - data-portal
  - scaffold
  - nextjs
  - template
  - catalog
---

# PortalJS — New Portal

## Overview

Scaffold a production-ready PortalJS data portal from a brief. The skill is
**interactive**: if the brief is thin it interviews the user in three short rounds
(mapped to the template's three surfaces — Home, Catalog, Showcase), echoes a brief
back for confirmation, then copies `examples/portaljs-catalog` (locally or via a
remote `tiged` fetch), substitutes placeholder tokens, sets the namespace mode, seeds
any datasets named in the interview, installs dependencies, and verifies the scaffold
with a type check.

## Prerequisites

- Node.js >=22 and npm available on `PATH`.
- Network access, unless a current local checkout of the portaljs repo is available
  (the resolver defaults to a remote fetch of the template).
- A destination directory name that does not already contain files (or user consent
  to overwrite one that does).

## Instructions

The canonical, full step-by-step workflow is
[`.claude/commands/portaljs-new-portal.md`](https://github.com/datopian/portaljs/blob/main/.claude/commands/portaljs-new-portal.md) —
the single source of truth. Read and follow it when executing. Summary:

1. Interview the user in up to three rounds — Home/basics, Catalog & discovery
   (datasets, namespace mode `theme` vs `owner`), Showcase/views — skipping any round
   already answered by the input brief. Accept "use defaults" at any point.
2. Confirm a short brief (name, slug, description, namespace, datasets, views) before
   building.
3. Resolve the template source: prefer a remote `tiged` fetch of
   `examples/portaljs-catalog` at `main` (or `PORTALJS_TEMPLATE_REF`); use a local
   checkout only when it is current (has `pages/[owner]/[slug].tsx`) and the
   destination is outside that repo.
4. Materialize the template into `./PROJECT_SLUG`, asking first if the destination
   already exists and is non-empty.
5. Substitute `__PROJECT_NAME__`, `__PROJECT_SLUG__`, `__DESCRIPTION__` tokens
   across all files with `perl -pi`, escaping `/`, `\`, and `&` in the values.
6. Set `NAMESPACE_TYPE` (`'theme'` or `'owner'`) in `lib/datasets.ts` per the
   interview.
7. Seed datasets captured in Round 2 (via `/portaljs-add-dataset` or by hand), or
   clear `datasets.json` to `[]` if none were named.
8. Run `npm install` inside the scaffolded portal.
9. Verify with `npx tsc --noEmit` (never `next build` here — it would corrupt a
   running dev server's `.next/` directory).
10. Report the scaffolded routes, namespace mode, and next steps.

## Output

- **Created:** a new directory `./PROJECT_SLUG/` containing the full
  `examples/portaljs-catalog` template with tokens substituted.
- **Modified:** `lib/datasets.ts` (`NAMESPACE_TYPE`); `datasets.json` (seeded
  datasets or cleared to `[]`).
- **Verified:** `npx tsc --noEmit` passes inside the scaffolded portal.
- **Result:** a runnable portal at `./PROJECT_SLUG` with Home (`/`), Catalog
  (`/search`), and Showcase (`/@<namespace>/<slug>`) surfaces wired up.

## Error Handling

| Symptom | Cause | Fix |
| --- | --- | --- |
| `DIR_EXISTS` | `./PROJECT_SLUG` already exists and is non-empty | Ask the user for a different name or consent to remove it; then proceed. |
| Remote fetch fails | Bad `PORTALJS_TEMPLATE_REF`, network outage, or `tiged` unavailable | Tell the user plainly and ask to retry, use a different ref, or check network. |
| Stale local scaffold (`pages/datasets/[slug].tsx`) | Old local clone missing `pages/[owner]/[slug].tsx` | Resolver already falls back to remote in this case — do not force local mode. |
| `npm install` fails | Node <22 or no network | Report the error and ask the user to check Node version and connectivity. |
| `tsc --noEmit` fails | Token substitution or manifest error | Print the log and fix before reporting success — never report success with a failing type check. |

## Examples

### Example 1 — Full brief up front

```
/portaljs-new-portal Auckland Open Data Portal — datasets published by several council
departments (multiple publishers). Start with ./data/parks.csv and ./data/budget.csv.
```
Infers name + description, picks `NAMESPACE_TYPE = 'owner'`, asks for namespace
values (e.g. `parks-dept`, `finance`), confirms the brief, scaffolds the template, and
seeds both datasets at `/@parks-dept/parks` and `/@finance/budget`.

### Example 2 — No arguments, full interview

```
/portaljs-new-portal
```
Runs all three interview rounds from scratch, accepting "use defaults" for any round,
then confirms the brief before scaffolding.

### Example 3 — Single-publisher portal with no datasets yet

```
/portaljs-new-portal Reference Data Hub — a single-team reference catalog, no data yet.
```
Picks `NAMESPACE_TYPE = 'theme'` with namespace `reference`, clears `datasets.json` to
`[]`, and reports `/portaljs-add-dataset` as the next step.

## Resources

- Full workflow: [`.claude/commands/portaljs-new-portal.md`](https://github.com/datopian/portaljs/blob/main/.claude/commands/portaljs-new-portal.md)
- Template and token reference: [`references/reference.md`](references/reference.md)
- Related skills: `portaljs-add-dataset`, `portaljs-add-chart`, `portaljs-add-map`, `portaljs-connect-ckan`
- PortalJS documentation: <https://portaljs.org>

More from datopian/portaljs

SkillDescription
arcgis-to-portaljsMigrate a whole ArcGIS Hub site into a PortalJS Arc portal end-to-end. Harvests the Hub /data.json (DCAT-US) inventory, exports every FeatureService layer through the ArcGIS REST query API with resultOffset paging, converts each to the serverless dual tier (PMTiles render + GeoParquet query) with tabular items to Parquet, pushes everything to Cloudflare R2 via Git LFS, appends dual-tier datasets.json entries, and writes a source-vs-derived parity report. Use to move a City or sector ArcGIS Hub open-data portal onto PortalJS with no server-side compute.
portaljs-add-chartAdd a chart (line, bar, area, pie, or scatter) to a dataset's showcase in a PortalJS portal. Installs recharts, writes a reusable Chart component, and renders it in the showcase Views section. Use when visualizing a dataset already registered in datasets.json.
portaljs-add-datasetAdd a dataset (CSV, TSV, JSON, or GeoJSON) to an existing PortalJS portal. Appends an entry to datasets.json so the catalog and showcase render it automatically; routes the data by source (local file vs remote URL) — R2 via Git LFS by default, remote URLs by passthrough. Use when registering a new dataset in a scaffolded portal.
portaljs-add-dcatMake a PortalJS portal harvestable by national/EU/US open-data portals — emit standards-compliant DCAT catalog feeds (DCAT 2/3, DCAT-AP, DCAT-US, national profiles) in JSON-LD, Turtle, and RDF/XML at build, with autodiscovery and per-profile conformance checking. Use when a portal needs to be harvested by data.europa.eu, data.gov, or a national open-data catalog.
portaljs-add-geoAuto-ingest a geospatial file (GeoJSON, Shapefile, GeoPackage, KML/KMZ, FlatGeobuf, CSV-with-geometry) into a PortalJS portal on the user's own machine, with no server. Normalizes CRS to EPSG:4326, derives a PMTiles render tier and a GeoParquet query tier, pushes all three artifacts to Cloudflare R2 via Git LFS, and appends one dual-tier datasets.json entry the showcase auto-renders. Use when the source is a vector geo format that needs a map or spatial-query view.
portaljs-add-mapRender a GeoJSON dataset on an interactive Leaflet map in the Views section of a dataset's showcase. Installs react-leaflet and a Map component, then renders the map for the chosen dataset. Use when a dataset's data is geographic and a map view is needed alongside the showcase's default metadata and download.
portaljs-add-resourceAdd another file (resource) to an EXISTING dataset in a PortalJS portal — a data dictionary, methodology, or an additional data file. Turns a single-file dataset into a multi-resource one; the showcase renders a section per resource. Use when a dataset needs a second file, such as a data dictionary, methodology doc, or an additional period's data.
portaljs-architectRecommend a data-portal architecture (storage, compute, catalog, access, hosting, metadata) from stated needs, then hand off to the build skills. The advisory entry point. Use when starting a new data-portal project and the underlying architecture has not yet been decided.
portaljs-check-data-qualityAudit a local or remote tabular file (CSV/TSV) for common data quality issues — schema, nulls, types, duplicates. Read-only. Use when a dataset needs a quality check before publishing, or a showcase renders wrong (blank cells, garbled numbers, an unsortable date column) and the cause needs isolating.
portaljs-connect-ckanWire a scaffolded PortalJS portal to a CKAN backend over its API. Generates a tiny server-side fetch client (no runtime dependency) and feeds the /search catalog and /@namespace/slug showcases from CKAN instead of datasets.json. Use when connecting an existing portal to a live CKAN instance instead of a static manifest.