Data warehouse · 1 skill

BigQuery MCP and Claude skills

Run read-only SQL on BigQuery tables and INFORMATION_SCHEMA, and report bytes billed and cost by user and query, slow jobs and the numbers your tables hold.

Connect BigQuery to your agent (MCP)

BigQuery MCP server, Google's remote server at https://bigquery.googleapis.com/mcp, available in any project where the BigQuery API is enabled. Add it to Claude Code, Codex or Cursor as a remote HTTP server and sign in with Google OAuth using the https://www.googleapis.com/auth/bigquery scope; Google's client setup guide covers each client. Its tools list datasets and tables, read schemas and run SQL; execute_sql_readonly rejects DML and DDL, and queries stop after three minutes or 3,000 rows (docs).
The bq CLI or the API with a service account. bq query --use_legacy_sql=false runs the same SQL with your gcloud login. For scheduled runs, point GOOGLE_APPLICATION_CREDENTIALS at a service account key and set BIGQUERY_PROJECT_ID.

Grant BigQuery Job User on the project and BigQuery Data Viewer on the datasets the report reads (plus roles/mcp.toolUser for the MCP server), and BigQuery Resource Viewer only if the report reads other people's jobs, which needs bigquery.jobs.listAll. An IAM deny policy on the writable execute_sql tool leaves the agent with the read-only one.

1 BigQuery skill

BigQuery cost by user and query

Every Monday: BigQuery bytes billed and estimated on-demand cost for the last 7 days, with the users and repeated queries behind the biggest share.

Data teamsMondays at 8:00Stat tile

What agents do with BigQuery

Agents read BigQuery with GoogleSQL through jobs.query or the bq CLI, and BigQuery describes its own use in INFORMATION_SCHEMA. Cost reports read region-us.INFORMATION_SCHEMA.JOBS_BY_PROJECT (or the JOBS alias), which records every job with user_email, creation_time, statement_type, total_bytes_billed, total_slot_ms, cache_hit and the query text; on-demand cost is bytes billed in TiB times your price per TiB. INFORMATION_SCHEMA.TABLE_STORAGE holds logical and physical bytes per table, JOBS_TIMELINE shows slot use over time, and TABLES and PARTITIONS show when data last landed.

Data teams schedule a daily or weekly cost report by user and query pattern, the ten most expensive queries with who ran them, tables that stopped receiving new partitions, storage growth by dataset, and business metrics read straight from a reporting table.

cloud.google.com/bigquery

Teams that use these

Your BigQuery numbers, on one board with everything else

Skills are free to use with any agent. Briefkin is where their results land: one board for every report, number and question your agents send, with a ping when something needs you.

Join the open beta Browse all skills

Free during the open beta. No card.