# Data Quality Auditor

> Validate CSV, TSV, JSON, or JSON Lines data against an explicit JSON rule set and emit a redacted, machine-readable pass/fail report. Use when checking required columns, nullability, uniqueness, types, ranges, allowed values, or row-count expectations before analysis or ingestion. Do not use to repair data or infer unstated business rules.

- Skill: `knuckles-team/data-quality-auditor` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add knuckles-team/data-quality-auditor`
- Raw SKILL.md: https://api.skillmd.com/api/skills/knuckles-team/data-quality-auditor/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- License: MIT
- Author: Knuckles-Team (https://skillmd.com/u/knuckles-team)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/knuckles-team/data-quality-auditor

---


# Data Quality Auditor

Run deterministic checks from a reviewed rule file:

```bash
python scripts/validate_dataset.py data.csv --rules quality-rules.json
```

Supported rule keys are `required_columns`, `non_null`, `unique`, `types`,
`ranges`, `allowed_values`, and `row_count`. The report contains counts and row
numbers but never raw failing values. The command exits nonzero on violations;
use `--warn-only` only when a non-blocking audit is explicitly intended.

## Rule example

```json
{
  "required_columns": ["id", "created_at"],
  "non_null": ["id"],
  "unique": ["id"],
  "types": {"id": "integer", "created_at": "datetime"},
  "ranges": {"score": {"min": 0, "max": 1}},
  "allowed_values": {"status": ["active", "inactive"]},
  "row_count": {"min": 1}
}
```

Review the rules and dataset authorization before running. A passing report proves
only the declared checks; it does not establish correctness, fairness, fitness for
purpose, or permission to use the data.