csv data validator
Audits CSV files for errors and outliers, generating professional quality reports and reproducible Python code.
Free
Works with the AI Agents you already use
csv data validator
Example session with this skill installed
Audit the data quality of student_grades.csv. Identify the primary key as 'student_id' and ensure 'final_score' is non-negative. Output the report to the current directory.
- Read your context and instructions
- Compiled the csv data validator
- Generated the data export
The CSV data quality audit for student_grades.csv has been completed. I have generated a detailed report, a JSON summary of findings, and the Python source code for your records.
Data Quality Audit: student_grades.csv
1. Audit Summary
- Quality Score: 82.5 / 100
- Grade: B (Good)
- Sample Size: 450 records
Core Findings: 12 missing values in 'final_score', 3 duplicate 'student_id' entries, and 5 outliers detected in 'age' (values > 100).
2. Field Overview
| Field | Type | Null Count | Null % | Distinct | Max | Min |
|---|---|---|---|---|---|---|
| student_id | int64 | 0 | 0.00% | 447 | 9999 | 1001 |
| student_name | object | 0 | 0.00% | 450 | - | - |
| age | int64 | 0 | 0.00% | 15 | 145 | 18 |
| final_score | float64 | 12 | 2.67% | 88 | 100.0 | -5.0 |
3. Priority Issues
- High: 12 missing values in 'final_score'. Missing target variable prevents analysis for these rows.
- High: 3 duplicate primary keys ('student_id'). Data integrity violation.
- Medium: 'final_score' contains negative values (Min: -5.0).
- Low: 5 records in 'age' exceed logical bounds (Max: 145).
4. Audit Limitations
- This audit only covers single-field integrity.
- Cross-field logic (e.g., midterm + final = total) was not checked.
- No PII masking was applied to 'student_name'.
Next steps
- Review the
student_grades_quality_report.mdfor specific row indices of outliers. - Run
python3 audit_csv.pyto reproduce these results on updated data. - Use the provided JSON to programmatically filter out the 15 rows with critical errors.
student_grades.csv
CSV · data export
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Raw CSV files often contain hidden defects that ruin downstream analysis or model training. This skill automates the data auditing process to detect missing values, duplicates, data type conflicts, and outliers using a structured, verifiable workflow. It generates a professional 8-chapter quality report in Markdown and provides the exact Python scripts used for the audit, ensuring full transparency and reproducibility for academic or professional data preparation.
What it does
- Automated screening: Scans files for null values, duplicate rows, and data type inconsistencies.
- Outlier detection: Identifies statistical anomalies and formatting errors within individual fields.
- Quality scoring: Calculates a weighted data quality score with a cap for small sample sizes (n<30) to prevent misleading high grades.
- Reproducible artifacts: Produces a standardized Markdown report, a machine-readable JSON summary, and the standalone Python auditing code.
- Structural evaluation: Assesses data structure and potential risks even from small pasted snippets when a full file is not provided.
How it works
- Pre-check validation: Verifies file path, extension (.csv/.tsv), and size (up to 120MB) before execution.
- Audit execution: Runs a specialized Python script using
pandasto perform field-level and row-level integrity checks. - Data extraction: Captures findings, severity levels, and scores into a centralized JSON artifact.
- Report generation: Compiles results into a fixed 8-chapter report including an appendix for replication.
- Verification: Executes a validation script to ensure the final report matches the raw JSON data without manual "beautification" errors.
Frameworks & tools
The skill requires Python 3 and the pandas library. It operates on local .csv, .tsv, or .txt files and outputs standard Markdown and JSON formats.
Why this beats prompting it yourself
Generic prompts often produce hallucinated statistics or inconsistent auditing logic. This skill uses a hard-coded Python contract and a validation layer to ensure every number in your report is backed by actual script execution, preventing the "beautification" of dirty data.
Use cases
- Auditing research data before performing statistical tests or modeling.
- Generating standardized data quality reports for course assignments or lab work.
- Screening dirty datasets to identify high-priority cleaning tasks.
- Verifying CSV exports for integrity before importing into a production database.
Known limitations
Does not support Excel (.xlsx), PDF, or images directly; users must export to CSV first. Maximum file size is 120MB. Cross-field business logic and multi-table joins are not supported.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Supports only .csv/.tsv/.txt files up to 120MB. Requires pandas. Does not support multi-table joins, cross-field business rules, or direct Excel/PDF parsing.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed today
- Free with an Agensi account