Is there an existing issue for this?

• I have searched the existing issues

Problem statement

Developers frequently work in their local IDEs to build code that will run in Databricks, and need a way to give Agents context about their Unity Catalog estate.

A Genie tool allows for querying data, but we also need a way to get table schematics (e.g. information schema) as this would improve correctness when Agents are writing mature spark code.

For example, when developing a Pyspark job or pipeline that reads from a table, does transformations, and writes to another table, the UC metadata about the source & sink tables, their columns, comments, and properties would be very helpful to generate the correct code in .select(), .map, etc.

Proposed Solution

Add these new tools under the MCP server's unity_catalog tools, essentially translating the databricks.sdk.service.catalog.TablesAPI and a couple other functions from the SDK:

1. list_tables
2. get_table
3. list_table_summaries
4. list_catalogs
5. list_schemas

Additional Context

I'm willing to contribute a PR.