How to Configure JetBrains AI Snowflake Prevent Delete

Published
Length
3 min read

Luis Felipe Dias

All posts

You set up a jetbrains ai snowflake prevent delete rule by limiting database role permissions. You remove delete grants from the role that your code editor uses to log in. Most teams start by sharing one broad database role across all local tools. That shared role often lets the editor drop rows or wipe tables by mistake. A single bad prompt in the chat window can trigger a big data loss. How jetbrains ai snowflake prevent delete roles work Your code editor uses a local database tool to run q

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 a jetbrains ai snowflake prevent delete rule by limiting database role permissions. You remove delete grants from the role that your code editor uses to log in. Most teams start by sharing one broad database role across all local tools. That shared role often lets the editor drop rows or wipe tables by mistake. A single bad prompt in the chat window can trigger a big data loss.

How jetbrains ai snowflake prevent delete roles work

Your code editor uses a local database tool to run queries for the assistant. When you ask a question, the assistant writes a query for your active session. That session runs under the database role you set up in your editor profile. To block deletes, admins create a custom role inside the Snowflake permission model. They grant read and insert rights to that custom role, but skip delete rights. Next, the developer updates the local editor profile to use this limited role. Now the database checks every query from the assistant against those role permissions. If the assistant tries to remove rows, the database engine blocks the statement.

What native database roles solve and what they miss

Native role permissions do a good job of stopping basic delete commands. They set a hard floor on what the editor session can do. Still, static role permissions only check who logs in to the database. They do not inspect what individual statements do while they run in flight. A developer often needs update rights in the editor to fix a bug. With update rights, a bad prompt from the assistant can still overwrite whole columns. Static roles also fail to stop large data dumps of sensitive customer records. When the assistant reads a table, raw emails and phone numbers enter the prompt. Card numbers and national IDs can also flow right into the model context. Managing dozens of micro roles for every developer task creates huge admin toil. Teams end up switching back to broad roles just to get daily work done.

Why query checks and field masking matter for editor assistants

Safe editor workflows need controls that look at each query before it runs. Instead of relying only on broad role grants, teams can check statements in flight. Statement rules can block drops, deletes, or unindexed reads no matter who logs in. You can allow safe updates while pausing risky changes for a quick human check. At the same time, data masking rewrites matched fields in query results right away. Emails, phone numbers, and card numbers get masked before the editor reads the output. The assistant gets the table structure it needs without exposing live customer records.

AI Data Exfiltration Prevention + Snowflake Access Control: 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 JetBrains AI run delete queries on its own?

The assistant generates queries and can run them if your editor session allows it. If your active Snowflake role has delete permissions, those queries will remove data.

Does a read-only Snowflake role stop all assistant risks?

A read-only role stops deletes, but it still lets the assistant read sensitive tables. Raw emails, card numbers, and national IDs can still pass into the model prompt.

How can developers run safe updates without enabling deletes?

Admins can grant select and update permissions to a custom role while leaving out delete. For finer control, teams inspect each statement in flight and approve risky queries first.

The query runs. The emails stay hidden.