How to Configure Windsurf Bigquery Credentials

Published
Length
3 min read

Luis Felipe Dias

All posts

You set up windsurf bigquery credentials through your local shell or a Model Context Protocol server. The coding agent then uses your cloud login or key file to run SQL queries. Many teams start by giving the agent a broad cloud role or personal login. This fast setup helps builders test ideas right away. Yet broad permissions let an automated tool read or change every table in the project. How windsurf bigquery credentials work in your local setup On June 2, 2026, Cognition renamed the Winds

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 windsurf bigquery credentials through your local shell or a Model Context Protocol server. The coding agent then uses your cloud login or key file to run SQL queries. Many teams start by giving the agent a broad cloud role or personal login. This fast setup helps builders test ideas right away. Yet broad permissions let an automated tool read or change every table in the project.

How windsurf bigquery credentials work in your local setup

On June 2, 2026, Cognition renamed the Windsurf editor to Devin Desktop. The app keeps the same features and uses the Cascade agent to assist developers. This agent can run terminal commands on its own after the user gives permission. It opens a dedicated terminal session using zsh on macOS or bash on Linux.

That terminal loads the user's standard shell profile right away. Because of this, all local aliases and exported variables are ready for the agent. When you log in to Google Cloud on your machine, the shell stores your session. Any command the agent runs in that terminal inherits your active cloud permissions.

Teams also connect the Cascade agent to external data through Model Context Protocol servers. You define these servers in a local config file on your workstation. That file often holds a path to a service account key or cloud project details. When the agent needs data, it calls the server, which logs in to BigQuery.

Why static roles leave a gap during live queries

Cloud roles and service account keys do one specific job well. They check who is allowed to log in to the data warehouse. You can scope a role down to specific datasets or read-only views. This native setup stops unauthorized users from reaching your cloud project from the outside.

Still, static cloud permissions cannot inspect what an individual SQL statement does in flight. Once the agent logs in, it holds the full power of that role for every query. If the role can read a customer table, the agent can pull every column by default. That means raw emails, phone numbers, card numbers, and national IDs flow into the prompt context.

Ephemeral Credentials + BigQuery IAM: architecture patterns and best practices

Continue reading? Get the full guide.

Free. No spam. Unsubscribe anytime.

Local config files also create key management chores for engineering teams. Service account files sit on laptops until someone rotates or deletes them. If a developer shares a config file, those static keys can spread across machines. Also, cloud audit logs only show the shared service account or the developer's name. They do not record which agent prompt triggered a costly full-table scan or a bad write.

Scoping permissions and checking queries in flight

A safer setup separates login keys from the coding agent on the workstation. First, teams replace broad project roles with dataset-level permissions in Google Cloud. The agent only gets read rights on the specific tables needed for the task. Drop and alter rights stay off the role used for daily coding work.

Second, teams move away from storing long-lived service account files in plain text. Keeping raw keys off the laptop stops the shell or agent from reading them directly. Third, teams add query checks and field masking to the database path. Masking rewrites matched fields like emails or card numbers in the query results. This step keeps sensitive values out of the agent context while preserving table structure.

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 inline in the protocol beside the service as an open-source proxy, a sidecar for your database or API. On every production query an agent or human runs, hoop.dev decides whether to pass, mask, block, or escalate it for review.

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 approval, and hoop.dev data masking rewrites sensitive fields in responses in memory. Explore the MIT-licensed project on GitHub.

Frequently asked questions

Does the rename to Devin Desktop change how the agent reaches BigQuery?

No, the editor keeps the same core features under the Devin Desktop name. The agent still runs commands in a local shell and talks to MCP servers.

Can the agent read my local shell variables when calling Google Cloud?

Yes, the dedicated terminal loads your zsh or bash profile when it starts. Any cloud login or key variable in your shell is open to the agent.

Why should teams avoid putting an admin service account in the MCP config?

An admin key lets the agent read all sensitive tables or delete production datasets. Using a scoped role with query checks keeps mistakes and data exposure contained.

The query runs. The emails stay hidden.