Unfragile

The request is for a data cleaning agent, not a named product, so the best-fitting tools are those whose pages show cleaning, validation, or tabular data work: DuckDB 5 for SQL-based data wrangling and SQLFluff 7 for SQL linting, while Cerberus 3 validates records against schemas and Zepworks 8 compares and fixes Python data. Airtable 1 offers record-cleaning actions in its workspace, and Polars 2 provides a fast DataFrame engine for cleaning at scale.

1
Airtableairtable.comverified and signed
Its Clean Record action updates a record and empties every field left blank 11, and its MCP server and SDK let agents update records programmatically.
Website Docs Pricing Get started
official:MCP server · @airtable/mcp-cli SDK · airtable repository · skills skill · airtableMCP server · airtable-mcp-cli+29 community tools built for it
for agents: airtable_clean_record · airtable_delete_record · airtable_update_record
2
Polarspola.rsverified and signed
A high-performance DataFrame library for Python and Rust built on Apache Arrow for filtering and transforming tabular data at scale 2.
Website Docs
+1 community tools built for it
for agents: dataframe · dataframe · arrow
3
DuckDBduckdb.orgverified and signed
An in-process SQL OLAP database whose Create and Query DB action builds a database from data and runs a query against it 15.
Website Docs
+27 community tools built for it
for agents: createAndQueryDB · SQL · OLAP
4
Cerberuspython-cerberus.orgverified and signed
A lightweight Python data validation library that checks dictionaries against a defined schema 3.
Website Docs
for agents: normalization · validation · dictionaries
5
SQLFluffsqlfluff.comverified and signed
An open source SQL linter that finds style and syntax issues and auto-fixes most of them 7.
Website Docs
+3 community tools built for it
for agents: linter · clickhouse · formatter
6
Zepworkszepworks.comwebsite read, tools not yet checked
Creator of DeepDiff, a library for deep difference, search, and hashing of Python objects, used to compare and fix data 8.
Website Docs

Sources

1Airtable: Build Enterprise-ready AI Workflows, Apps & Agents | Airtableairtable.com · read 2026-10-29
2Polars — DataFrames for the new erapola.rs · read 2026-10-29
3Cerberus — Data validation for Pythondocs.python-cerberus.org · read 2026-10-29
5An analytical SQL database management system – DuckDBduckdb.org · read 2026-10-29
7SQLFluffsqlfluff.com · read 2026-10-29
8Zepworkszepworks.com · read 2026-10-29
11Airtable (Activepieces): declares "airtable_clean_record" (the record we read and signed)unfragile.ai · read 2026-10-24
15DuckDB (Activepieces): declares "createAndQueryDB" (the record we read and signed)unfragile.ai · read 2026-10-24

For agents: the 10 actions behind this answer

airtable_clean_record · Airtable (Activepieces) verified and signed
Clean Record — Updates a record and empties every field you leave blank.
dataframe · polars verified and signed
normalization · Cerberus verified and signed
data · pydata-google-auth verified and signed
createAndQueryDB · DuckDB (Activepieces) verified and signed
Create and Query DB — Create DB from data and run query
· SciPy website read, tools not yet checked
Why SciPy? Fundamental algorithms. Broadly applicable. Foundational. Interoperable. Performant. Open source.
linter · sqlfluff verified and signed
· Zepworks website read, tools not yet checked
Sep Dehpour builds tools for comparing and fixing data. Creator of DeepDiff, founder of Qluster.
· NumPy website read, tools not yet checked
Why NumPy? Powerful n-dimensional arrays. Numerical computing tools. Interoperable. Performant. Open source.
· VisiData website read, tools not yet checked
Command-line interactive multitool for tabular data.

Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.