BigQuery Essentials
Manage BigQuery data with CLI commands and run SQL analysis.
The documentation explicitly recommends least privilege and covers IAM, service accounts, VPC Service Controls, row and column security, audit logs, and authorized views; the MCP execute_sql tool is restricted to SELECT and labels its queries. Points are deducted because CLI examples include deletion, copying, loading, and writes without required user confirmation, a sensitive-data workflow, uniform data-flow disclosure, or rollback guidance, and queries can incur project charges.
The main document and reference directory are internally coherent, covering CLI, client libraries, IaC, MCP, and security topics, with plausible common paths on static inspection. Points are deducted for no test suite, dependency pinning, abnormal-input handling, or diagnosable failure feedback; authentication, project, location, and dependency prerequisites are also incomplete. Static calibration caps this dimension at 10.
The intended scenarios are reasonably clear: BigQuery queries, datasets, tables, jobs, ingestion, and analysis, with an explicit SELECT-only boundary for MCP. Points are deducted because trigger exclusions, input/output expectations, capability boundaries, and non-fit scenarios remain incomplete; Chinese-language support is not described, and the core service depends on Google Cloud/BigQuery reachability without addressing mainland-China network conditions.
The skill uses a main entry document with topic-specific reference files and provides setup examples, related skills, an Apache-2.0 license, contribution channels, and an active-development note. Points are deducted for no explicit maintenance owner, version policy, changelog, dependency versions, FAQ, or systematic troubleshooting guidance; several examples rely on placeholder projects and resources.
CLI, Python, Java, Node.js, Go, Terraform, and MCP examples cover core BigQuery operations and are potentially usable starting points. Points are deducted because the files provide no execution results or test evidence, and completing a real task still requires the user to supply authentication, project configuration, permissions, cost controls, and error handling. Static calibration caps this dimension at 7.
The files cite Google Cloud, Terraform, client-library, and MCP documentation, and the supplied revision enables basic source traceability. Points are deducted because there are no committed tests, CI coverage, or third-party execution records; key commands and product claims therefore receive only indirect verification. Static calibration caps this dimension at 5.
- Before running bq rm, bq cp, bq load, bq insert, or remote queries, confirm the target project, resources, data scope, permissions, and expected charges; the skill lacks a general confirmation or rollback workflow.
- Do not treat the examples as secure production configurations; verify authentication, least-privilege IAM, location consistency, sensitive-data controls, and BigQuery network reachability first.
- The MCP SELECT restriction does not guarantee harmless execution: the document states that remote functions or Python UDFs may have side effects.
What does this skill do, and when should you use it?
This skill covers basic BigQuery resource management and data analysis. It provides examples for enabling the BigQuery API, creating datasets and tables, and running SQL queries. Its reference directory also points to CLI usage, client libraries, MCP, Terraform, and IAM and security topics. It fits users who need to operate BigQuery resources directly; advanced BigQuery AI and machine learning capabilities are outside this skill's scope.
It guides users through enabling the BigQuery API with gcloud, creating a dataset and table with bq, and supplying a table schema through schema.json. It also shows how to run standard SQL with bq, including querying a public dataset with a row limit. The references cover analytics workflows, client libraries, a remote MCP server, Terraform, IAM, and data governance.
- A Google Cloud user needs to enable BigQuery and create a first dataset.
- A data engineer needs to create a BigQuery table from a JSON schema.
- A data analyst needs to run standard SQL through the bq CLI and inspect query results.
- A developer or platform engineer needs references for BigQuery client libraries, Terraform, or IAM.
What are this skill's strengths and limitations?
- Covers core operations for datasets, tables, and queries.
- Includes copyable gcloud and bq CLI examples.
- Provides references for CLI usage, client libraries, MCP, Terraform, and IAM security.
- Belongs to a Google skills collection described as under active development.
- Focused on basic BigQuery operations rather than BigQuery AI and machine learning capabilities.
- No test suite, version compatibility matrix, or platform validation is provided.
- Pricing, authentication setup, and common failure scenarios are not explained.
- The installation command targets the full google/skills repository, so the specific skill must be selected during installation.
How do you install this skill?
Run npx skills add google/skills and select BigQuery Basics during installation. The skill is located at skills/cloud/bigquery-basics. The repository is released under the Apache 2.0 license.
How do you use this skill?
In an Agent Skills-compatible client, request a BigQuery task such as “Use BigQuery Basics to create a US dataset and table, then run a query.” You can also follow the documented commands: gcloud services enable bigquery.googleapis.com --quiet, bq mk --dataset --location=US my_dataset, create schema.json, run bq mk --table my_dataset.mytable schema.json, and use bq query --use_legacy_sql=false for SQL. Authentication, detailed IAM configuration, and error-handling procedures are not documented in the supplied material.