GoldenCheck

GoldenCheck

Data-quality scanning for the Golden Suite. Profiles a dataset, surfaces issues (encoding problems, format violations, anomalies, suspicious placeholders), and feeds findings to GoldenFlow for automatic remediation.

Source: packages/python/goldencheck · PyPI: goldencheck

Install

pip install goldencheck
goldencheck scan customers.csv

What it does

Capability	Notes
Profile column types, nulls, cardinality	Polars-native, fast
Detect encoding issues	Mojibake, mis-decoded UTF-8
Format validation	Email, phone, ZIP, IDs (CUSIP, NPI, EIN, SSN, etc.)
Anomaly detection	Placeholder data ("john.doe@example.com", "123-45-6789")
Severity grading	info / warning / critical per finding
Suite integration	Findings drive goldenflow.engine.selector.select_from_findings()

Programmatic API

import goldencheck

result = goldencheck.scan_file("customers.csv")
print(f"{len(result.findings)} findings")

for f in result.findings:
    if f.severity == "critical":
        print(f"  ⚠ {f.check}: {f.column} ({f.rate:.1%} affected)")

result.to_dict() returns a JSON-serializable form with findings, summary, source, etc. — that's what the Airflow golden_suite_quality_gate.py DAG keys off.

Suite role

   raw rows ──► [GoldenCheck profile/scan] ──► findings
                                                ↓
                                      drives GoldenFlow transforms
                                                ↓
                              cleaned rows ──► GoldenMatch

GoldenCheck never modifies data. It only reports. Pair it with GoldenFlow when you want auto-remediation, or treat its findings as a hard gate (see golden_suite_quality_gate DAG).

CLI

goldencheck scan FILE                    # full scan
goldencheck scan FILE --severity critical # filter to critical findings only
goldencheck scan FILE --format json      # machine-readable
goldencheck mcp-serve --transport http --port 8100  # MCP server

MCP

Hosted at ghcr.io/benzsevern/goldencheck-mcp:latest. Tools include scan_file, profile_columns, detect_anomalies, format_check. Surfaced under the goldensuite-mcp aggregator alongside the rest of the Suite.

dbt integration

packages/dbt/goldencheck/ ships a dbt package that runs GoldenCheck against warehouse models as dbt tests. Drop into packages.yml and call from dbt test.

See also

• Quality gate Airflow DAG
• Containers
• Master MCP Aggregator
• Full README in repo: packages/python/goldencheck/README.md