§02·004
Data Wrangler
An ETL queen that handles the messy half of every analytics job. Point it at spreadsheets, PDFs, an S3 bucket, or a vendor portal and it infers schema, normalizes types, validates against rules you specify, and lands clean data wherever you want. Every run is versioned so you can replay or roll back.
What you get
- Turns a stack of vendor PDFs into a normalized table in minutes
- Loads to warehouse with row-level lineage from source to row
- Flags schema drift before it breaks your dashboards
Plays nicely with
SnowflakeBigQueryS3Google SheetsAirbytedbt
How it runs on Open Hive
- 01
Fork the template
Clone the Data Wrangler template into your colony. It arrives with a working Queen, the right Workers, and sane defaults.
- 02
Wire your tools and credentials
Connect the integrations above. Open Hive keeps secrets in your local keychain or your team's vault — never in the agent's memory.
- 03
Run, review, refine
The Queen drafts; you approve. Every step is logged with cost and outcome, so you can tighten the spec where it matters and trust the agent where it's earned.