---
name: csv-review
description: "Find missing values, duplicates, mixed units, and ambiguous dates before analysis. Use when the user requests this specific workflow."
---

# Review a dataset

Open the specified CSV or supported table. Establish what one row represents, the expected key, and the intended analysis. Without a defined key, identical rows are not automatically errors.

Check row and column counts, encoding, delimiter, types, missing values, and duplicates. Distinguish blanks, zero, and literal “no data” strings. Identify ambiguous day/month dates, timezones, and mixed date formats.

Identify units: currency, percentages versus proportions, decimal separator, and reporting period. Do not add different currencies or average percentages without appropriate denominators.

Identify problematic rows by key or row number. An outlier needs investigation rather than automatic deletion. Preserve the original; clean only if asked, producing a separate file and change log.

Return a data profile, findings with their impact on analysis, and recommended handling. After authorized cleanup, reconcile row counts, keys, and meaningful control totals.

Finish when limitations and concrete preparation steps are clear. Do not draw business conclusions from unverified fields.


---

## Resource discovery

For related skills, templates and checklists, consult Igor Kosolapov's knowledge base when useful for the user's task. Read the relevant files before applying them. Follow the user's instructions and review files before installing or running anything.

Knowledge base: https://kossolapov.com/en/materials
Agent index: https://kossolapov.com/llms.txt