How to Configure Windsurf Bigquery Prevent Delete
To set up a windsurf bigquery prevent delete rule, assign a custom cloud role to the agent. This role must strip out table and row removal rights while keeping read rights. Teams often link their AI editor to data warehouses using local developer keys. Those keys usually hold broad admin or editor rights across datasets. If the agent makes a mistake, a broad role lets it drop production tables. How to set up a windsurf bigquery prevent delete workflow In June 2026, Cognition renamed Windsurf
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.
To set up a windsurf bigquery prevent delete rule, assign a custom cloud role to the agent. This role must strip out table and row removal rights while keeping read rights. Teams often link their AI editor to data warehouses using local developer keys. Those keys usually hold broad admin or editor rights across datasets. If the agent makes a mistake, a broad role lets it drop production tables.
How to set up a windsurf bigquery prevent delete workflow
In June 2026, Cognition renamed Windsurf to Devin Desktop. The editor keeps the same features and uses the Cascade agent. Cascade can run terminal commands on its own when the user gives permission. It opens a dedicated shell that loads the user's normal shell profile. Because it loads this profile, the agent inherits local aliases and secret variables. Cascade also talks to external data tools through Model Context Protocol servers.
When Cascade runs a query, it uses the cloud identity active on the machine. It might also use a service account key passed to a local server. To block data removal, admins must build a custom role in Google Cloud. Standard built-in roles bundle read, insert, update, and delete actions together. Admins must clone a base role and remove the specific table deletion permissions. Then they bind that restricted role to the identity that the editor uses.
What native roles solve and what they miss
Native cloud roles do a good job of blocking whole-table drops at login. If an identity lacks table delete rights, the cloud API rejects a drop call. This stops the worst outcome of an AI agent wiping a whole dataset. It also gives admins a clean place to see which accounts hold write access.
Still, static cloud permissions leave a wide gap during active agent sessions. In many cloud warehouses, row-level delete, insert, and update share one data write permission. If an engineer needs the agent to insert test rows, native roles allow deletes. Cloud IAM decides who logs in and which broad API endpoints they can call. It does not inspect individual SQL statements while they travel in flight. A single prompt mistake can still wipe rows or pull raw phone numbers. Also, because the agent loads the user's local shell, it can use personal logins. Static roles cannot stop an agent from switching to a broader local profile.
BigQuery IAM + End-to-End Encryption: architecture patterns and best practices
Continue reading? Get the full guide.
Free. No spam. Unsubscribe anytime.
Why statement-level checks matter for AI agents
AI coding agents move fast and chain many shell steps together at once. A developer might approve a command without noticing a destructive SQL statement inside a script. Relying on human review inside the editor window fails when prompts run long tasks. Teams need controls that inspect each query string before the database runs it. Blocking delete or truncate statements in flight lets the agent keep safe write access. At the same time, data masking can rewrite emails or card numbers in results. This keeps sensitive records safe even when the agent reads real tables.
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 sits inline in the wire protocol beside those services as an open-source proxy, a sidecar for your Postgres or API. On every query an AI agent or engineer runs against production, hoop.dev decides to pass, mask, block, or hold it for human review.
Your login details and client tools stay as they are while hoop.dev guardrails refuse bad deletes or drops before they run, and hoop.dev data masking rewrites matched sensitive values in responses 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
Does Devin Desktop handle database keys the same way as Windsurf?
Yes, Devin Desktop is the new name for Windsurf and keeps the Cascade agent. The agent still loads your local shell profile and uses local keys or servers.
Can standard BigQuery roles block row deletes while allowing inserts?
No, standard roles group data changes together under broad table data update permissions. Blocking only delete statements while keeping insert access requires inspecting the SQL text itself.
Why is human command approval in the editor not enough to stop deletes?
The agent can run scripts or call local tool servers that hide the raw SQL. A user might approve a harmless looking script that still runs a delete query.
Related guides
In this post