How to Configure Devin Bigquery Read Only Access
You set up devin bigquery read only access with a cloud service account and viewer roles. This setup lets the AI agent run queries and read tables without changing data. Many teams start by giving the agent a broad project role to move fast. Broad roles let the agent read every dataset and run huge, expensive queries. A better plan limits the agent to specific datasets and tracks every query. How devin bigquery read only access works in cloud projects Google Cloud uses Identity and Access Man
Auditor Read-Only Access + Customer Support Access to Production: the complete guide
Architecture patterns, implementation strategies, and security best practices. Delivered to your inbox.
You set up devin bigquery read only access with a cloud service account and viewer roles. This setup lets the AI agent run queries and read tables without changing data. Many teams start by giving the agent a broad project role to move fast. Broad roles let the agent read every dataset and run huge, expensive queries. A better plan limits the agent to specific datasets and tracks every query.
How devin bigquery read only access works in cloud projects
Google Cloud uses Identity and Access Management to control who reads data and runs jobs. BigQuery splits permissions into two main parts for any user or automated agent. First, the agent needs permission to create and run a query job in the project. Second, the agent needs permission to read the table rows inside a target dataset. You create a dedicated service account just for the AI agent to use. You grant a job user role at the project level so queries can run. Next, you grant a data viewer role only on the datasets the agent needs. This stops the agent from seeing other tables in the same cloud project. Finally, you pass the service account key or workload identity to the agent workspace. The agent uses these credentials to log in and run select queries on your tables.
What native roles solve and where gaps remain
Cloud viewer roles block write actions like drop, delete, update, or insert statements. The database engine rejects those commands before they can change or harm your stored records. Dataset scoping also keeps the agent out of restricted folders like finance or payroll. Still, cloud roles only check who logs in and which tables they can open. A viewer role cannot tell safe rows apart from sensitive rows inside one table. If a shared table holds emails, phone numbers, or card numbers, the agent reads them. The agent pulls those raw values into its context window during a normal select query. Static service account keys also sit inside the agent workspace until someone rotates them. If a prompt causes a bad query, native roles will not mask the output.
Ways to tighten query controls for AI agents
Teams can add extra database rules to reduce the risk of broad data exposure. You can build authorized views that leave out columns with emails or national IDs. You give the agent viewer rights on the clean view instead of the base table. You can also set up column-level security tags inside the cloud data catalog. These tags block the service account from reading flagged columns during a select query. Setting custom cost quotas per user or service account also helps control spend. Quotas stop a runaway agent loop from scanning terabytes of data in one afternoon. You should also review cloud audit logs each week to see what queries ran.
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.
Check database and API queries on the wire
Identity tools, roles, and network rules decide who opens a session. Setup alone cannot check what each query does or mask sensitive values in a response once a session starts. When workflows touch Postgres, MySQL, SQL Server, MongoDB, ClickHouse, Kubernetes APIs, SSH, or internal HTTP APIs, hoop.dev runs beside those services as an open-source proxy, a sidecar for your Postgres or API. It inspects requests coming in and results going back on the wire.
Your login details and client tools stay as they are while hoop.dev guardrails refuse bad deletes or drops before they run. On responses, hoop.dev data masking rewrites matched sensitive values in memory. Furthermore, hoop.dev holds SOC 2 Type II, is MIT licensed, and writes every query and decision to an audit trail that generates evidence for HIPAA, GDPR, and PCI DSS. See the source on GitHub.
Frequently asked questions
Can a BigQuery viewer role still run up high cloud costs?
Yes, a viewer role with job permissions can scan huge tables many times. You should set daily byte limits to cap how much data the agent scans.
Why does the agent need two separate permissions in BigQuery?
BigQuery separates compute from storage in its security model. One role lets the agent run the query job, while the other reads the data.
How do we handle sensitive columns when an agent queries a table?
You can use authorized views, cloud policy tags, or data masking rules on results. Data masking rewrites matched fields like emails or card numbers before results return.
Related guides
In this post