How to Configure Continue Snowflake Prevent Delete

Published
Length
3 min read

Luis Felipe Dias

All posts

You set up a continue snowflake prevent delete workflow by assigning read-only database roles. Then you give those restricted credentials to the Continue coding extension in your editor. Teams often share broad developer roles with their local coding assistants by mistake. When a role holds broad rights, one bad prompt can wipe out production tables. Limiting what the assistant can run keeps your warehouse safe from accidental row drops. How to set up continue snowflake prevent delete with rol

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 continue snowflake prevent delete workflow by assigning read-only database roles. Then you give those restricted credentials to the Continue coding extension in your editor. Teams often share broad developer roles with their local coding assistants by mistake. When a role holds broad rights, one bad prompt can wipe out production tables. Limiting what the assistant can run keeps your warehouse safe from accidental row drops.

How to set up continue snowflake prevent delete with roles

Continue is an open-source coding assistant that runs inside your code editor. It reads project context and can run queries against your warehouse to help write code. To stop delete commands, you must change how you configure Snowflake permissions first. Every query from the assistant uses the session role you assign to its user.

Start by creating a dedicated role in your warehouse just for the coding assistant. Grant this role usage rights on the specific warehouse, database, and schema it needs. Next, grant select permissions on the tables or views the assistant should read. Do not grant delete, drop, or truncate permissions to this specific role.

Create a service user and assign the new read-only role as its default. Put those user credentials into your local Continue settings file. Now the warehouse blocks any query from the editor that tries to remove rows. The database engine checks the role before it runs any statement from the session.

Why role rules leave gaps in daily work

Static database roles work well when an assistant only needs to read table data. Still, developers often use the exact same local credentials for their own daily tasks. Those developers sometimes need to insert test rows or update staging tables during work. If you strip all write rights, engineers must switch accounts all day long.

When you grant write permissions for staging work, delete rights often come bundled together. Standard role setups decide who can log in and which objects they can touch. They do not inspect the actual text of each query while it runs. A role cannot allow a safe three-row delete while blocking a full table wipe.

Snowflake Access Control + End-to-End Encryption: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Static permissions also fail to protect sensitive columns inside the tables the assistant reads. If the role has select rights, the assistant pulls back raw customer data. That means emails, phone numbers, and card numbers flow straight into the model prompt. Teams need a way to check individual statements and mask fields in real time.

Safe habits for AI coding assistants

Never let a local coding assistant share your personal admin account in the warehouse. Separate accounts make it easy to read query logs and see what the tool ran. You can spot bad query patterns fast when the assistant has its own user name.

Limit the assistant to small compute warehouses with strict timeout rules in place. A bad prompt can trigger a massive join that burns through your compute credits. Short query timeouts stop runaway reads before they drive up your monthly cloud bill. Pair these limits with strict role grants to keep your data safe every day.

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

Can Continue block delete queries on its own?

No, the editor extension only sends prompts and runs the resulting queries locally. You must enforce delete blocks inside the database role or through an external check.

What happens if the assistant tries to drop a table?

When you use a read-only role, the warehouse rejects the drop statement right away. It returns a permission error to the editor and leaves your table data untouched.

Should developers use their own login for Continue?

No, personal developer logins usually have too many rights across staging and production schemas. Always create a separate user with a locked-down role for any coding assistant.

The query runs. The emails stay hidden.