v1.3.0
This plugin provides a specialized suite of skills for data engineers and database practitioners working on Google Cloud. It acts as an expert assistant, allowing you to use natural language prompts in your preferred coding agent to architect complex data pipelines, transform data with dbt, write Spark and BigQuery SQL notebooks, and orchestrate end-to-end workflows across GCP's data ecosystem.
Changelog
1.3.0 (2026-11-11)
Features
- gcp-managed-spark-upgrades: Add Scala/Java upgrade support to the Spark upgrade skill. (f1b64ce)
- gcp-spark-troubleshooting: add platform mechanics reference & methodology (55f5597)
- gcp-spark-troubleshooting: add read-only Spark environment inspector (eff0c60)
- mcp: add Cloud Composer remote MCP server to Data Cloud Agent Kit (95ba689)
- mcp: add Dataform remote MCP server to Data Cloud Agent Kit (f1e2a0c)
- skills: add discover-assets skill for Data Agent Kit (5422f53)
- skills: add discover-assets skill to DAK telemetry and EvalBench (dbc4600)
- skills: add Knowledge Catalog curate-context skill to Data Agent Kit (3729647)
- skills: add MERGE target partition pruning rules to bigquery-optimization (bf0a1af)
- skills: delegate resolving-mcp-region-configs to dak-setup (8b7b086)
- support AirflowTaskAction (airflowTask) in the gcp_pipeline_orchestration skill (4e39bd6)
- support retryPolicy and deprecate executionConfig in the gcp_pipeline_orchestration skill (b3035b9)
Bug Fixes
- correct issue tracker URLs and broken documentation links (b030611), refs #460 #461
- dak-setup: rewrite MCP config when server entries change (70e0714)
- gcp-spark: generate Spark Connect init code for google-cloud-spark-connect 2.x (42ce776)
- gcp-spark: refine skill routing, cluster creation, Spark SQL, and missing-input rules (4124772)
- mcp-proxy: Support new mcp server version 2026-07-28 (142a775)
- skills: harden curate-context MCP, IAM, and portability guidance (e7ff374)
Miscellaneous Chores
- force release 1.3.0 (5f11699)
1.2.0 (2026-10-29)
Features
- dak: add cloud_databases_onboarding skill and evalbench suite to Data Agent Kit. (05d8b85)
- dak: proxy MCP resources in the data-agent-kit CLI MCP server. (a29f288)
- dak: register data-agent-kit MCP server in plugin configs and update bundle. (36cc29f)
- gcp-managed-spark-upgrades: Added PySpark Upgrade Skill for Serverless Spark (4a63d89)
- mcp: add Firestore remote MCP server (dak-firestore) (e5ab3f3)
- mcp: consolidate visualization
render_charttool intodata-agent-kitMCP server (7d7bbb3) - skills: add bigquery-observability, bigquery-optimization, and bigquery-troubleshooting skills to Data Agent Kit (e2f6401)
- skills: add gcp-spark-observability skill and evalbench suite (751cbb6)
- skills: fold BigQuery workload monitoring deltas into bigquery-observability and bigquery-troubleshooting (e5bd77e)
- support notifications.onPipelineSuccess in the gcp_pipeline_orchestration skill (62b5941)
- support Vertex AI createAndRunCustomJob in the gcp_pipeline_orchestration skill (3c25afb)
- Update BigQuery AI skills/references in data_agent_common (439c66e)
Bug Fixes
- bigtable-basics: update broken Bigtable SQL reference link. (191e5b3)
- gcp-spark: add Amazon S3 and Azure Blob Storage / ADLS Gen2 support (44b58bf)
- gcp-spark: add Cloud SQL and AlloyDB Socket Factory and IAM auth support (ce9f362)
- gcp-spark: add parallel JDBC read, custom driver JAR, and Secret Manager patterns (f1ade85)
- gcp-spark: Ask the user when no data source is named (79377a4)
- gcp-spark: clarify Remote Kernel Spark session initialization (2c71b4a)
- gcp-spark: refine skill description for reliable routing in Claude harnesses (40358fd)
- internal: Add automatic sanitization / cleaning to external release pipeline and fix existing leak (ddb8a2e)
- skills: Remove the google-cloud-auth-verification skill. (323c760)
Miscellaneous Chores
- force release 1.2.0 (904dc9b)
1.1.0 (2026-10-13)
Features
- dak: check in the notebook/visualization MCP server bundle (mcp/dist/index.js) into source of truth. (f39934f)
- skill: Make gcp-spark trigger for Spark notebook authoring (4db7383)
- skills: add publisher/license frontmatter to 3 BigQuery skills (3c285ad)
- spark: read Spark stage data through the Dataproc sparkApplications API (aa96190)
Bug Fixes
- dak-setup: keep ${region} in MCP server URLs when the region is empty (db7da5b)
- gcp-spark: standardize BigQuery and BigLake Iceberg write patterns (05d1609)
- gcp-spark: update BigQuery SQL query and Spanner read configurations (b66bf71)
- skills: Minimize user prompting for external shell and SQL probing during notebook tasks (89c4fe5)
- skills: scope Spark production logging to batch scripts and add notebook rubric (f24fc90)
Miscellaneous Chores
- force release 1.1.0 (cf9f39b)
1.0.0 (2026-09-29)
Features
- add DAK setup script (6003384)
- Add onboarding support for Codex and Agy CLI. (36720b7)
- Cleanup Claude and Codex plugin configs. (d5357d1)
- dak-setup: add DAK plugin setup skill and CLI bundle (ca53915)
- Define BigFrames as the default Python library for BigQuery data (59173ba)
- mcp: Add BILLING_PROJECT_ID quota header support to CLI MCP proxy (81806f3)
- Remove plugin setup config prompts across agents (1a7a684)
- Replace
bigquery_graphskill withbigquery_graph_authorandbigquery_graph_query(graph v2) (fa97c18) - skill: Add cell execution failure handling protocol to notebook guidance (97892b3)
- skill: Add guidance for PySpark ML vector slicing and invalid label handling. (2febf2c)
- skill: guidance for batch submit command (b53d990)
- skill: Make sure libraries are not downloaded during notebook generation (59ff724)
- skills: add gcp-spark-troubleshooting skill for Dataproc and Spark diagnostics (1a044de)
- skills: add Spark optimization guidance as gcp-spark references (c3cb553)
- skills: Cap data asset discovery retries and update GCS listing in gcp_spark skill (5faa2af)
- skills: Enforce driver memory protection in gcp-spark optimizations (bccafd5)
- skills: GCP Managed Spark upgrade SKILL for PySpark job type on Cluster (6dc2c29)
- skills: harden gcp-spark code guidance for production logging, parameterized config, and broadcast joins (af143c8)
- skills: Update gcloud Dataproc skill guidance for multi-package batch submission and Spanner connector (4823a3c)
- skills: Update gcp_spark skill for ManagedSparkSession with better instructions (2e04bb5)
- skills: update gcp-spark skill to forbid whole-file reads of executed notebooks (66f0d37)
- skill: Update Iceberg guidelines to favor session templates for notebooks (764829f)
Bug Fixes
- clarify SQL action guidelines for BigQuery and Dataproc engines (c057fb5)
- clarify that the environment field is optional for Python actions (aedd28d)
- dak: restore default
npxresolution in package.json bin (5f4abac) - Document serving signature requirements for Vertex AI batch prediction with BigQuery sources. (3039bea)
- gcp-spark: rename spark_env virtualenv to venv in gcloud_dataproc.md (a6a3073)
- hooks: Make dak-setup SessionStart hook work in agy CLI (d045730)
- skills: Point resolving_mcp_region_configs at the Codex mcp.json that is actually loaded (117bbd0)
- skills: Sanitize internal references in Data Agent Kit skills (f1ef383)
- Update skill how to query Vertex AI batch prediction outputs in downstream SQL (8ef4e4b)
- Update SQL action guidelines for BigQuery and Dataproc engines in orchestration schema and documentation. (c057fb5)
Miscellaneous Chores
- force release 1.0.0 (a26cbd6)
0.11.0 (2026-09-11)
Features
- mcp: Add MCP support for Antigravity (agy) CLI. (856d2f2)
- mcp: replace npx git dependency with relative bundle paths for our plugin MCP servers. (e502daa)
- Switched managed spark skills to use spark connect to execute cells by default (fdd2b2a)
- Update mcp proxy and telemetry hook script as part of release (b5484f3)
Bug Fixes
- Add BigQuery dataset location discovery to gcp_pipeline_orchestration skill. (43a6f04)
- mcp: inline cross-spawn into standalone bundle (#316) (22c27d5)
Miscellaneous Chores
0.10.1 (2026-09-04)
Features
- add Bigtable basics skill and EvalBench test cases to DAK (e06c822)
- Add resolving-mcp-region-configs skill to Data Agent Common (0878ed8)
- Add Spark 4.0 BigQuery connector coordinates to gcp-spark skill (e86233d)
- mcp: route Google Cloud MCP services through bundled Node stdio proxy (d17797d)
- Simplify claude user config by removing BigQuery location and ProjectId string. (44d991f)
- skills: Update managing_python_dependencies with pre-flight environment check bundling (ad94694)
Bug Fixes
- Update BigFrames skills to reduce agent confusions. (deabd80)
Miscellaneous Chores
0.10.0 (2026-08-31)
Features
- Add semantic schema mapping skill for ontology driven data engineering (f015039)
- Add Spark Connect session lifecycle guide with DataprocSparkSession to gcp_spark skill (b977284)
- Avoid redundant cluster creation for serverless batch jobs in gcp_spark skill (15d9e27)
- Enforce argparse parameterization for Spark scripts in gcp_spark skill (d0cf7eb)
- Enforce Python standard logging over print() in gcp_spark skill (29f17a0)
- orchestration: update orchestration pipeline schema (eb5f3e3)
- orchestration: update orchestration pipelines schema to support AI actions (fc13f54)
- Remove hardcoded secrets from Cloud SQL examples in gcp_spark skill (682aee5)
- skills: Update google-cloud-auth-verification with bundled probe and IAM handling (a891ebc)
- Use @toolbox-sdk/server@latest and remove Windows-specific quoting (4e51d1c)
Bug Fixes
- Preserve host attribution values when tagging gcloud commands. (e830e87)
- skills: Generalize GCP auth verification pre-flight hierarchy (fd3ac55)
- skills: Remove unsupported aliases field from google-cloud-auth-verification frontmatter (3bc2e21)
Miscellaneous Chores
- force release 0.10.0 (3dd3f00)
0.8.0 (2026-08-07)
Features
Miscellaneous Chores
- force release 0.8.0 (35a2768)
0.7.2 (2026-07-28)
Features
- skills: Clarify BQ label enforcement rules for resource attribution. (c16b546)
- skills: Update bq label flag syntax in resource attribution skills. (1ff44be)
Miscellaneous Chores
- force release 0.7.2 (62d2f9c)
0.7.1 (2026-07-27)
Bug Fixes
- remove default hooks (1171475)
0.7.0 (2026-07-24)
Features
- Add AI.AGG function to Data Agent Kit (9431ff2)
- Bigtable MCP support (26f175e), refs #146
- Bigtable MCP support (c58f5e9), refs #146
- Modularize BigQuery skills in data_agent_common (b29c63a)
- Remove DB skills, which are just wrappers on MCP toolbox from data agent common. Will add helpful skills once we come up with them. (a9beba2)
- skills: Add more IDE/environment values for resource attribution environment labels. (3941694)
- Update BQ ai_function_best_practices and constraints. (f4394e1)
Bug Fixes
- Fix telemetry hook configurations for Codex, Claude Code, and Gemini CLI. (b93ca2e)
- Update skill references and bump version in dataform and dbt. (879f134)
Miscellaneous Chores
- force release 0.7.0 (3e89ba0)
0.6.1 (2026-07-10)
Features
- Duplicate graph-schema component to bigquery skill (b5bc330)
- gcp-managed-airflow-migration - a new skill for Airflow code migration (de2d876)
- rename plugin to dak and remove toolbox from mcp server names (#99) (cb3a6e8)
- skills: Add AlloyDB Omni skills (94b8582)
- skills: Add AlloyDB skills (fb5bd8b)
- skills: Add Cloud SQL MySQL skills (7f2442a)
- skills: Add Cloud SQL Postgres skills (420e3ac)
- skills: Add Cloud SQL SQL Server skills (7e887d0)
- skills: Add Firestore Native skills (f6fa0df)
- skills: add github-release-skill and update Copybara reviewers (ca05037)
- skills: Add Spanner skills (c1dc599)
- Update DAK Skill to support Gemini CLI, Claude Code, and Codex attribution tags (e1922d8)