csv-wizard

What This Skill Does

CSV Wizard is a comprehensive, interactive command-line interface designed to streamline the data preparation workflow directly within the OpenClaw environment. Data cleaning is often the most time-consuming phase of any data science or administrative project; this skill automates the heavy lifting by performing intelligent type inference, identifying data quality issues, and applying standardized cleaning procedures. Whether you are dealing with messy spreadsheet exports, logs, or raw datasets, CSV Wizard acts as a virtual data steward. It detects data types automatically, handles missing values through statistical imputation, scrubs duplicate records, and ensures that column headers meet best practices for programmatic consumption. Its interactive mode transforms the cleaning process into a guided experience, allowing you to make granular decisions about your data without needing to write complex Python scripts or manual Excel formulas.

Installation

To add this skill to your OpenClaw agent, execute the following command in your terminal or command-line interface:

clawhub install openclaw/skills/skills/antonia-sz/csv-wizard

Once installed, you can immediately invoke the /clean-csv command to begin processing your data files located within your workspace.

Use Cases

• Data Science Preparation: Quickly sanitize training datasets by imputing missing values and removing redundant rows before passing the data to machine learning models.
• Administrative Data Consolidation: Standardize inconsistent naming conventions in CSV exports from multiple legacy systems, ensuring uniform column headers like snake_case.
• Report Preparation: Generate quick statistical summaries and previews to verify file integrity before importing data into visualization tools or dashboards.
• Data Migration: Ensure that incoming CSV files are properly formatted and encoded (UTF-8) before importing them into databases or cloud storage solutions.

Example Prompts

1. "/clean-csv customer_leads.csv --info" - This prompt will provide an immediate high-level summary of your customer data, highlighting column types and identifying where missing values are concentrated.
2. "/clean-csv raw_sales.csv --drop-duplicates --fill-missing mean --output processed_sales.csv" - This command cleans a sales file by stripping out redundant entries and filling in gaps in numerical data with the column mean, saving the clean result to a new file.
3. "/clean-csv legacy_data.csv --interactive" - This launches the recommended interactive mode, allowing the agent to prompt you for specific cleaning decisions as it parses each column.

Tips & Limitations

• Safety First: Always use the --output flag to save your cleaned data to a new file. Avoid overwriting your original source files until you have verified the results.
• Encoding: CSV Wizard automatically standardizes files to UTF-8. If your legacy data uses unconventional encodings, you may need to verify the output contents after processing.
• Interactive Mode: If your dataset is exceptionally large (multi-gigabyte), the interactive mode may be slower than batch processing. Use batch commands for high-volume automated pipelines.
• Imputation: Be mindful of the statistical method chosen for missing values; using 'mean' or 'median' on categorical data will result in errors, so ensure the column type is numeric when selecting these strategies.