GoldenCheck Community Types

Community-contributed semantic type definitions for GoldenCheck.

Domain packs teach GoldenCheck about industry-specific column types, improving detection accuracy and reducing false positives. The same packs also drive InferMap's detect_domain (which industry a dataset belongs to) and, through it, GoldenPipe's auto-config — a dataset that scores a domain confidently gets a schema-inference stage prepended.

Available Domains

15 industry packs (plus a generic fallback used when no domain matches):

| Domain | Types | Description | |--------|-------|-------------| | automotive | 4 | Vehicles: VIN, license plate, registration, odometer, make/model | | ecommerce | 9 | SKUs, order IDs, tracking numbers, product categories, shipping | | education | 5 | Students, enrollment, courses, grades/GPA, terms, programs | | energy | 4 | Meters, consumption (kWh), tariffs, utility accounts | | finance | 8 | Account numbers, routing numbers, CUSIP/ISIN, currency, transactions | | healthcare | 10 | NPI, ICD codes, insurance IDs, patient demographics, CPT, DRG | | hospitality | 4 | Reservations, bookings, stays, rooms, guests | | hr | 11 | Employee IDs, job titles, departments, compensation, hire dates, status, org hierarchy, performance | | insurance | 9 | Policies, claims, premiums, coverage, underwriting, beneficiaries | | legal | 4 | Cases, dockets, matters, parties, courts, jurisdictions | | logistics | 5 | Shipments, tracking, carriers, freight, warehouses | | manufacturing | 4 | Parts, work orders, batches/lots, BOM, quality | | marketing | 5 | Leads, campaigns, pipeline, funnel metrics, attribution | | real_estate | 4 | Listings, properties, features (beds/baths/sqft), sale prices | | telecom | 5 | Subscribers, device IDs (IMEI/IMSI/ICCID), usage, plans, network |

Usage

Bundled (built into GoldenCheck)

goldencheck scan data.csv --domain healthcare

Community domains (download and use)

curl -o goldencheck_domain.yaml https://raw.githubusercontent.com/benseverndev-oss/goldencheck-types/main/domains/telecom.yaml
goldencheck scan data.csv

Via MCP (Claude Desktop)

Use the install_domain tool to browse and install community domains.

Contributing

1. Fork this repo
2. Add a YAML file in domains/ following the format below
3. Open a PR — CI validates your YAML automatically

YAML Format

description: "Short description of the domain"

types:
  my_type:
    name_hints: ["column_name_hint", "another_hint"]
    value_signals:
      min_unique_pct: 0.90    # optional: minimum uniqueness
      max_unique: 20          # optional: maximum unique values
      format_match: "email"   # optional: regex format
      mixed_case: true        # optional: mixed case values
      avg_length_min: 15      # optional: minimum average string length
      short_strings: true     # optional: short string values
      numeric: true           # optional: numeric values
    suppress: ["pattern_consistency", "cardinality"]  # checks to suppress

Name Hints

• Plain string: substring match ("npi" matches npi_number, provider_npi)
• Ending with _: prefix-only match ("is_" matches is_active but NOT diagnosis)
• Starting with _: suffix-only match ("_id" matches patient_id)

Valid Suppress Values

uniqueness, nullability, format_detection, type_inference, range_distribution, cardinality, pattern_consistency, temporal_order, encoding_detection, sequence_detection, drift_detection

License

MIT