Data Cleaning: Fix Messy Tables With a Full Change Log

    1

    Makes your AI agent clean messy CSV/Parquet/Excel tables (stray spaces, N/A text, numbers stored as text, spelling variants, duplicates) and log every change before → after. Never overwrites your raw file.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Data Cleaning: Fix Messy Tables With a Full Change Log

    Example session with this skill installed

    Clean titanic3.csv before I start modelling. Show me everything you changed.

    • Read your context and instructions
    • Compiled the data cleaning
    • Generated the data export

    Summary

    Cleaned titanic3.csv into data/clean.csv. 2 fixes changed 7 cells, nothing was deleted, and 9 things are flagged for you to decide. The raw file is untouched.

    Cleaning log

    Input: examples/titanic3.csv
    Rows: 1309 -> 1309 · Columns: 14 -> 14
    2 fixes (7 cells/rows changed) · 0 type conversions · 9 things to check

    Changes made

    StepColumnRowsWhat changedBeforeAfter
    strip_whitespacename4removed spaces at the start/end of text'Daly, Mr. Peter Denis ''Daly, Mr. Peter Denis'
    category_variantshome.dest3different spellings of the same category merged into the most common one'Sweden Worcester, MA''Sweden Worcester, MA'

    To check (nothing was changed)

    CheckColumnRowsNote
    missing_valuesage26320.1% missing. Not filled in here: impute inside the model pipeline, after the split
    outlierssibsp15values far outside the usual range (3 x IQR). Check if they are errors; do not delete real extreme values
    outliersfare84values far outside the usual range (3 x IQR). Check if they are errors
    missing_valuescabin101477.5% missing
    missing_valuesboat82362.9% missing
    missing_valuesbody118890.8% missing
    missing_valueshome.dest56443.1% missing
    missing_valuesembarked20.2% missing
    missing_valuesfare10.1% missing

    Also saved: cleaning_log.md and cleaning_log.json.

    Next steps

    • Decide on the flags (for example, keep real extreme fares).
    • Run leakage-check before training: boat and body are mostly empty for a reason.

    titanic3.csv

    CSV · data export

    Generated

    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

    Clean messy tables with your AI agent, and see every single change it made.

    It fixes, and logs before → after:

    • Stray spaces, and "N/A", "?", "null" or "-" turned into real missing values
    • Numbers and dates stored as text ("$1,200", "12%", "(150)", mixed date formats)
    • Category spelling variants ("male", "Male ", "MALE")
    • Exact duplicate rows

    It only flags what needs a human decision: missing values, placeholder codes like -999, outliers, negative ages or prices, empty columns.

    Safe by design: it never overwrites the raw file, has a dry-run mode, protects ID and zip-code columns, and does not impute or delete outliers (those belong in your model pipeline, after the split).

    • Works with Claude Code and any agent that reads SKILL.md
    • One tested Python script: pandas, numpy
    • No API keys, no network access
    • Open source (MIT): github.com/m-shahmeer-kz/dataskills

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Data & Databases.

    Does not perform data imputation or outlier removal. Requires a local Python environment. User must manually specify columns to protect, such as IDs or ZIP codes, to prevent auto-conversion.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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 2 days ago

    • Free with an Agensi account

    Software
    Python 3.9+pip

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Listed2 days ago

    What's inside

    Frequently Asked Questions

    The skill natively supports CSV, Parquet, and Excel (.xlsx) files, outputting your cleaned data in either CSV or Parquet format along with detailed Markdown and JSON logs.

    Unlike generic AI prompts that might silently delete data, this skill uses a "log-everything" protocol that documents every cell transformation, ensuring you have a full audit trail for reproducibility and debugging.

    It never deletes outliers or rows with missing values; it flags them for you to decide. The only rows it removes are exact duplicates, and each removal is logged. Use --keep-duplicates if repeated rows are real.

    The SKILL.md instructions, the tested clean_data.py script, requirements.txt and a short install note. Each run writes the cleaned file plus cleaning_log.md and cleaning_log.json. The skill is free and open source (MIT).

    The skill is built for any AI agent or workflow capable of executing Python scripts within a local environment or containerized data science stack.

    Yes. Pass --protect (for example --protect zip,phone) so those columns stay exactly as they are, keeping leading zeros and formatting.

    New in Data & Databases

    Free