Data Agent Kit is now GA: Bring Google Data Cloud to any coding agent

Google Cloud today announced the general availability of the Data Agent Kit, a free set of Model Context Protocol tools and agent skills designed to let coding agents such as Antigravity, Claude Code, or Codex work directly with Google Cloud data products. The kit adds support for BigQuery Graph, Bigtable, and Managed Service for Apache Spark access to open Lakehouse environments, alongside dozens of usability improvements aimed at streamlining daily workflows. Agents can now inspect schemas, run queries, read job logs, and manage resources in live environments using connections to more than 15 Google Data Cloud services.

The kit includes open-source skills authored by Google Cloud engineers that teach agents data best practices, such as optimizing BigQuery SQL, designing Bigtable row keys, and building dbt pipelines. It integrates as an IDE extension for VS Code-compatible editors, plugins for Antigravity 2.0 and CLI, Claude Code, and Codex, or directly in Cloud Shell and Cloud Workstations where it is pre-installed. The IDE extension embeds a lightweight Google Cloud console view for browsing data, running queries, and reviewing agent outputs without leaving the editor.

New at general availability is support for building and querying BigQuery property graphs using natural-language descriptions, with automated schema mapping and GQL generation. Bigtable is now a fully supported database in the kit, with a new skill that designs row-key schemas to prevent hotspots and full table scans, and enables querying with GoogleSQL. Managed Service for Apache Spark can now query Apache Iceberg tables in the Lakehouse with serverless sessions, supporting branching, time travel, and schema evolution, while federated catalog skills connect to AWS Glue and Databricks Unity Catalog for in-place querying.

The kit also automates pipeline creation and troubleshooting, allowing agents to generate dbt or Dataform models and schedule them as Orchestration Pipelines on Managed Service for Apache Airflow. Skills guide agents toward cost-aware query patterns and integrate with Identity and Access Management, Model Armor, VPC Service Controls, and Principal Access Boundary policies for governance. Data Agent Kit is included at no additional cost, with users paying only standard pricing for the Google Cloud services their agents consume.