# CSV Quality Auditor

> Audit CSV and delimiter-separated text files for blank or duplicate headers, inconsistent row widths, and fully empty data rows. Use when Codex needs to validate a CSV before import, diagnose structural data-quality failures, or produce a machine-readable audit report without third-party packages.

- Skill: `qwqqwqis666/csv-quality-auditor-3` (Agent Skill, multi-file: 5 files)
- Install (CLI): `npx skillmds@latest add qwqqwqis666/csv-quality-auditor-3`
- Raw SKILL.md: https://api.skillmd.com/api/skills/qwqqwqis666/csv-quality-auditor-3/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: qwqqwqis666 (https://skillmd.com/u/qwqqwqis666)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/qwqqwqis666/csv-quality-auditor-3

---


# CSV Quality Auditor

Run the bundled auditor before importing, transforming, or handing off a CSV file.

## Workflow

1. Identify the input file and its delimiter.
2. Run the auditor from this skill directory:

```bash
python scripts/audit_csv.py INPUT [--delimiter DELIM] [--json]
```

3. Omit `--delimiter` for comma-separated input. Pass `semicolon`, `tab`, `\t`, or one custom character when needed.
4. Add `--json` for automation. Parse stdout only when the exit status is `0` or `1`; failures are written to stderr.
5. Report the issue codes, logical record numbers, and messages. Do not modify the source file.
6. Read [references/rules.md](references/rules.md) when exact counting, normalization, or output semantics matter.

## Exit Status

- `0`: The file was read successfully and no quality issues were found.
- `1`: The file was read successfully and one or more quality issues were found.
- `2`: Command usage, file reading, text decoding, or CSV parsing failed.