onboarding

Onboarding

Learn how to use the Taxonomy Creation Plugin: what it does, how to invoke each skill, the directory structure, and a worked example.

When to use

• "How do I use this plugin?"
• "Show me an example of creating and loading a taxonomy"
• "Where does the data go?"

Procedure

What this plugin does

The Taxonomy Creation Plugin generates and manages lookup tables and hierarchical taxonomies:

• Standard taxonomies: ISO 3166 countries, ISO 4217 currencies, ISO 639 languages, US states, Canadian provinces, IANA timezones, HTTP status codes, MIME types.
• Custom taxonomies: From a natural-language spec (e.g., "types of evaluations", product categories, content classifications).
• Hierarchical taxonomies: Parent/child trees (categories/subcategories, organizational structures).
• Load & export: Push to Postgres, SQLite, MySQL; or export to CSV, JSON, NDJSON, SQL, YAML.

Data directory structure

All data lives under:

${CLAUDE_USER_DATA:-${XDG_DATA_HOME:-$HOME/.local/share}/claude-plugins}/taxonomy-creation/
├── data/                    # Generated taxonomies, ready-to-load
│   ├── countries/
│   │   ├── countries-iso-3166-1.csv
│   │   ├── countries-iso-3166-1.json
│   │   ├── README.md        # Schema and source documentation
│   │   └── exports/         # Alternative formats (JSON, SQL, NDJSON, etc.)
│   └── eval-types/          # Custom taxonomy example
├── cache/                   # Cached standard datasets (30-day TTL)
├── state/                   # Operational logs
│   ├── loads/               # Load operation logs
│   ├── validations/         # Validation reports
│   └── exports/             # Export operation logs
└── config.json              # (Optional) settings: paths, dataset sources, pins

No default config.json is required; all operations are self-contained.

Typical workflow

1. Generate a taxonomy:

   • generate-standard — Pull ISO countries, currencies, languages, or other standard dataset.
   • generate-from-spec — Build a custom taxonomy from a description.
   • hierarchical-taxonomies — Create a tree structure (categories/subcategories).

2. Validate (optional but recommended):

   • validate-taxonomy — Check for duplicate codes, missing values, referential integrity, cycles.

3. Load or export:

   • load-to-postgres / load-to-sqlite / load-to-mysql — Push to a database.
   • export-formats — Convert to CSV, JSON, SQL, NDJSON, or YAML for use elsewhere.

Worked example: ISO countries to SQLite

1. Generate standard: "I want ISO 3166-1 countries with alpha-2 and alpha-3 codes."

   • Skill: generate-standard → Inputs: "countries", "ISO 3166-1 alpha-2 and alpha-3".
   • Output: data/countries/countries-iso-3166-1.csv and .json.

2. Validate: "Check the countries list for any issues."

   • Skill: validate-taxonomy → Input: data/countries/countries-iso-3166-1.csv.
   • Output: Validation report to stdout; exit 0 if clean.

3. Load to SQLite: "Load the countries into a local SQLite database."

   • Skill: load-to-sqlite → Inputs: ~/my-app.db, table name "countries", mode "replace".
   • Output: Table created in my-app.db; load log written to state/loads/.

Done. The countries table is ready in your SQLite database.

Skill reference

• generate-standard — Standard taxonomies (countries, currencies, languages, timezones, HTTP codes, MIME types).
• generate-from-spec — Custom taxonomies from a description.
• hierarchical-taxonomies — Parent/child trees.
• validate-taxonomy — Check for errors before load.
• load-to-postgres — Push to Postgres.
• load-to-sqlite — Push to SQLite.
• load-to-mysql — Push to MySQL/MariaDB.
• export-formats — Convert to CSV, TSV, JSON, NDJSON, SQL, YAML.
• onboarding — This skill.

No configuration required

The plugin requires no upfront setup. All data paths default to ${CLAUDE_USER_DATA}/taxonomy-creation/. If you want to override paths or pin standard-dataset sources (e.g., use a local mirror of ISO lists), edit ${CLAUDE_USER_DATA}/taxonomy-creation/config.json — but this is optional and not needed for basic use.

Questions?

Each skill's documentation includes detailed workflows, schema recommendations, and safety constraints. Start with generate-standard or generate-from-spec to create your first taxonomy.