BigQuery agent skills

BigQuery provides agent skills that can help your agent accomplish common tasks. Agents that use these skills can then accomplish tasks with higher fidelity than an agent without the skills.

Agent skills are a standardized way to give AI agents specific knowledge and instructions they can reuse. They contain resources like scripts, reference materials, and templates that help the agent complete tasks accurately and reliably. For more information about skills, see Agent Skills Overview.

The BigQuery basics skill

The BigQuery basics skill, bigquery-basics, contains instructions for completing basic tasks in BigQuery, such as creating a dataset or running a query.

Install the BigQuery basics skill

To install the bigquery-basics skill, run the following command in your agentic client or terminal:

npx skills add google/skills --skill bigquery-basics

BigQuery basics reference files

The BigQuery basics agent skill includes several reference files that provide technical depth. These references let the skill execute functions or commands accurately by providing specific context to the underlying model. By default, the BigQuery basics skill includes the following reference files:

The BigQuery AI & ML skill

The BigQuery AI & ML skill, bigquery-ai-ml, contains instructions for completing data science, machine learning, and AI tasks in BigQuery.

Install the BigQuery AI & ML skill

To install the bigquery-ai-ml skill, run the following command in your agentic client or terminal:

npx skills add google/skills --skill bigquery-ai-ml

BigQuery AI & ML reference files

The BigQuery AI & ML agent skill includes reference files that provide technical depth. By default, the BigQuery AI & ML skill includes the following reference files:

Ground your agent in official documentation

The Developer Knowledge MCP server is a remote server that searches current Google developer documentation, including the BigQuery documentation on docs.cloud.google.com, so your agent can verify SQL syntax, function signatures, and bq command-line tool flags against the published documentation rather than relying only on its training data. It provides the search_documents, get_documents, and answer_query tools and complements the BigQuery skills: the bigquery-basics skill instructs your agent to search the official documentation when it needs product information that isn't in its reference files, and the Developer Knowledge MCP server gives your agent that capability.

To add it to your agent, follow Configure your IDE or coding agent and select your client. To teach your agent when to search the documentation and which Developer Knowledge tool to use, also install the companion retrieving-developer-knowledge agent skill alongside the BigQuery skills.

With both installed, your agent can combine a documentation lookup with a BigQuery task. For example:

  • "Look up the current syntax for AI.FORECAST in the BigQuery documentation, then write a query that forecasts the next 30 days of COLUMN_NAME in PROJECT_ID.DATASET_ID.TABLE_ID."

What's next