Works with the AI Agents you already use
BigQuery Data Catalog & Auto-Documentation
A warehouse nobody can navigate is a cost centre; one that is documented becomes the foundation for self-service analytics and for every data agent you build later.
$12
BigQuery Data Catalog & Auto-Documentation
Example session with this skill installed
I exported last month's catalog to catalog.json (with descriptions people wrote by hand) and today's schema to snapshot.json. Update the catalog without losing anyone's work, tell me what changed and which columns might hold personal data, and write a plan in CATALOG_PLAN.md for keeping it documented, including AI-drafted descriptions.
- Read your context and instructions
- Compiled the bigquery data catalog
Merged catalog keeps the hand-written descriptions of order_id and amount. Changes in acme-dw.sales.orders: coupon removed, customer_email added, amount NUMERIC → FLOAT64. New table acme-dw.sales.stores added (not reported as changes). Likely personal data: orders.customer_email; confirm with a DLP sample scan and label the table confidential. Documentation scores: orders 40 (missing validation, example queries), stores 0. CATALOG_PLAN.md: nightly sync with change report to owners, AI drafts from names, types and stats only, owner approval, policy tags for PII columns.
Connects securely to your tools. The creator never sees your data.
About this skill
A warehouse nobody can navigate is a cost centre; one that is documented becomes the foundation for self-service analytics and for every data agent you build later. The hard parts are not the UI: they are keeping human documentation alive across syncs, catching schema changes, finding personal data, and doing all of it without expensive or unsafe queries.
What your agent does with it
- Sync without destroying documentation (
catalog_sync.py): merges a fresh schema snapshot into your catalog, keeps hand-written descriptions, reports columns added, removed and type-changed (never for brand-new tables), lists tables that vanished. - Flags likely personal data by column name, then explains how to confirm with a Cloud DLP sample scan, sensitivity labels and BigQuery policy tags.
- Generates safe profiling SQL (
profile_sql.py) with strict identifier validation: one-pass column stats,TABLESAMPLEpreviews, usage fromINFORMATION_SCHEMA.JOBSviareferenced_tables, nested column paths. - Scores documentation 0–100 per table and lists the gaps, so you work on the most-queried tables first.
- AI-drafted descriptions from metadata and statistics only (never sampled rows), approved by an owner before they count.
- Two shapes: an agent that writes descriptions back into BigQuery, or a full catalog app (data model, roles, encrypted secrets, notifications, deployment notes included).
Included: SKILL.md, catalog-app reference, catalog_sync.py and profile_sql.py (standard library, no credentials needed), and an eval case with fixtures.
Part of the Production AI Agents series. The catalog is the semantic context for Text-to-SQL Data Agent and the Looker Studio Dashboard Generator.
BigQuery-specific SQL. Name-based PII detection is a first pass, not a guarantee; content scanning needs Cloud DLP or similar. It does not include a ready-made catalog UI, only its design.
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 any payment.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 day ago
Requires
Permissions
File Scopes
The scripts only merge JSON and generate SQL; they never connect to BigQuery. You run the SQL with your own credentials.