csv-analysis

name	csv-analysis
description	Analyze CSV file format and schema quality. Use for delimiter detection, header checks, row-shape validation, type drift, malformed quoting, and parser configuration recommendations.
argument-hint	Path to CSV file(s) and any parsing constraints
user-invocable	true

CSV Analysis

What This Skill Does

This skill inspects CSV files for structural correctness, schema consistency, and parsing risks before ingestion.

When To Use

• You need to validate delimiter, quoting, escaping, and header behavior.
• You suspect malformed rows, inconsistent column counts, or multi-line field issues.
• You need parser settings and schema constraints for reliable ingestion.

Constraints

• Do not modify files unless explicitly asked.
• Do not invent values when evidence is missing.
• If confidence is low, state assumptions and required additional sample size.

Procedure

1. Identify file characteristics:

• Detect delimiter (comma, semicolon, tab, pipe).
• Detect quote and escape behavior.
• Verify whether a header row exists and whether header names are usable.
• Note encoding and newline style when detectable.

2. Validate row shape:

• Infer expected column count from header or first consistent row.
• Flag rows with extra or missing fields.
• Flag trailing delimiters, blank rows, and broken multi-line field boundaries.

3. Assess column quality:

• Infer candidate type per column (string, integer, float, date, boolean).
• Identify null tokens and empty-value conventions.
• Flag type drift in the same column.
• Report basic cardinality and representative examples.

4. Detect formatting anomalies:

• Broken quotes or unescaped delimiters in quoted fields.
• Inconsistent date/time formats.
• Locale-dependent numeric formats.
• Leading/trailing whitespace anomalies.

5. Summarize risk and recommendations:

• Rank issues by severity: critical, warning, info.
• Recommend parser settings and schema constraints.
• Propose cleanup strategy when needed.

Required Output Order

1. File Overview
2. Detected CSV Dialect
3. Schema Snapshot
4. Issues (Critical/Warning/Info)
5. Recommended Parser Configuration
6. Suggested Remediation Steps
7. Confidence and Assumptions

Parser Recommendation Template

Use this when enough evidence exists:

delimiter: ","
quotechar: '"'
escapechar: "\\"
header: true
encoding: "utf-8"
newline: "\n"
strict_column_count: true
null_tokens: ["", "NA", "null", "n/a"]

Practical Guidance

• For large files, sample first rows plus middle and tail segments.
• Use shell-based shape checks for speed, then confirm edge cases with direct file reads.
• Keep observed facts separate from inferred interpretations.