How to Configure Continue Bigquery Read Only Access
You set up continue bigquery read only access by pairing a viewer role with local credentials. The coding assistant then reads table schemas and runs select queries without changing your data. Many teams start by handing out broad project editor keys to get the tool working fast. That shortcut lets an automated prompt drop tables or alter production records by mistake. Scoping the cloud identity down to read permissions stops destructive write commands at the source. How continue bigquery read
Your coding agents run with your access.
Claude Code reaches prod with a developer's credentials, and approval prompts get clicked through. hoop.dev checks every query, masks sensitive data, and holds risky changes for a human.
You set up continue bigquery read only access by pairing a viewer role with local credentials. The coding assistant then reads table schemas and runs select queries without changing your data. Many teams start by handing out broad project editor keys to get the tool working fast. That shortcut lets an automated prompt drop tables or alter production records by mistake. Scoping the cloud identity down to read permissions stops destructive write commands at the source.
How continue bigquery read only access works in practice
The assistant runs inside the code editor and talks to your data warehouse through a provider. You first create a dedicated service identity or user group inside your cloud console. Next, you attach two core permissions to that identity. The first permission lets the user start query jobs in the project. The second permission lets the identity read table metadata and rows inside specific datasets. You then link those credentials to the local editor setup on the developer laptop. When a developer asks a question, the assistant fetches the schema and writes a select statement. The cloud engine checks the attached role, sees only viewer rights, and runs the read query. If the assistant tries to insert, update, or delete rows, the database rejects the job right away.
Where static cloud roles fall short during live queries
Cloud viewer roles do a great job of blocking table drops and bad data writes. They decide who can log in and which datasets a user can read. Still, static roles cannot inspect the content of a select query while it runs. Once a developer has viewer rights on a table, every column in that table is open. The coding assistant can pull customer emails, phone numbers, card numbers, or national IDs into a prompt. A static role also cannot tell if a query will scan ten rows or ten billion rows. A single bad prompt from the editor can trigger a massive full table scan and spike cloud costs. Also, shared service keys often end up sitting as plain text files on developer laptops. When five engineers share one key, your audit logs only show the service account name. You lose track of which person ran which query from their editor.
Scoping datasets and tracking editor queries
To reduce risk with native tools, teams split sensitive tables into separate cloud datasets. You grant the coding assistant viewer rights only on clean datasets that hold no personal details. For tables that mix business metrics with emails or card numbers, you can build authorized views. An authorized view hides sensitive columns so the assistant only reads safe fields. Teams also set byte limits on query jobs to stop huge scans before they finish. Finally, ask each developer to log in with their own cloud user identity instead of a shared key. Individual logins keep your cloud audit logs clean and tie every query back to a real person.
Auditor Read-Only Access + Customer Support Access to Production: architecture patterns and best practices
Continue reading? Get the full guide.
Free. No spam. Unsubscribe anytime.
Add runtime guardrails and masking beside your databases and APIs
Login rules, role grants, and CI checks govern who connects to your stack. Yet any allowed script, engineer, or AI agent still runs commands with the full scope of its role. For workloads on Postgres, MySQL, SQL Server, MongoDB, ClickHouse, Kubernetes, SSH, or internal HTTP services, hoop.dev sits beside the service as an open-source proxy, a sidecar for your database or API.
Before a query runs, hoop.dev guardrails check the statement against your rules. An AI Analyzer you configure can score intent or hold risky calls for human review. Next, hoop.dev data masking rewrites sensitive fields in responses, and each verdict is logged per listener. Explore the MIT-licensed project on GitHub.
Frequently asked questions
What permissions does the coding assistant need to run a read query?
The identity needs permission to read dataset tables and permission to create query jobs in the project. Without job creation rights, the user can view table schemas but cannot run select statements.
Can a cloud viewer role hide sensitive columns from the coding assistant?
A standard dataset viewer role opens every column in the allowed tables by default. To block specific fields like emails or phone numbers, you must set up column rules or data masking.
Why should teams avoid putting service account keys on developer laptops?
Local key files are easy to copy, leak in code commits, or share across a team. Using personal logins with short sessions keeps audit logs accurate and removes static keys from laptops.
Related guides
In this post