Database Testing in Ephemeral Agent Sandboxes

Give cloud coding agents disposable databases seeded with synthetic or scrubbed fixtures, never an implicit route to production data.

This guide focuses on the engineering decision behind database testing agent sandbox: what to standardize, what to constrain, and what evidence a reviewer should expect before accepting the result.

The core decision

Package schema creation and migrations as part of setup. Seed only the cases needed for the task and assign a unique database or schema per run. The test user should have exactly the permissions exercised by the application, not administrative access to the service.

A workflow that holds up in review

Validate forward migration, application behavior, and cleanup. For destructive migration work, include a representative copy in an isolated environment and require a rollback or restore rehearsal. Capture query plans only when data scale materially affects the issue.

The failure mode to design around

A tiny fixture can make a slow or unsafe query look correct. Balance reproducibility with realism by maintaining approved scale fixtures that contain no customer data and can be generated on demand.

Implementation checklist

Provision per-run database state.
Use synthetic or approved fixtures.
Test migrations and cleanup.
Keep production networks unreachable.

Turn the checklist into operating controls

Provision per-run database state: name the owner, the evidence that proves it happened, and the condition that should stop the run.
Use synthetic or approved fixtures: name the owner, the evidence that proves it happened, and the condition that should stop the run.
Test migrations and cleanup: name the owner, the evidence that proves it happened, and the condition that should stop the run.
Keep production networks unreachable: name the owner, the evidence that proves it happened, and the condition that should stop the run.

The list becomes useful when every item produces visible evidence. Store that evidence with the task or pull request rather than in a private chat. A future reviewer should be able to tell which repository revision was used, which permission profile applied, what stopped or failed, and who accepted the remaining risk. For database testing agent sandbox, a short, complete record is more valuable than a long narrative that cannot be reproduced.

Move from one run to a repeatable practice

Promote environment changes like build-system changes. Review the script and image inputs, test a clean build, test a warm build, and record the cache key. Roll the change out to a small task set before making it the shared default. Keep the previous environment definition available long enough to reproduce an active incident. If developers cannot run the core setup locally or in another sandbox, document why and provide an equivalent diagnostic path.

Before expanding the workflow, ask three review questions:

What evidence shows that provision per-run database state was satisfied, and would that evidence survive a rerun from the recorded commit?
What evidence shows that use synthetic or approved fixtures was satisfied, and would that evidence survive a rerun from the recorded commit?
What evidence shows that test migrations and cleanup was satisfied, and would that evidence survive a rerun from the recorded commit?

Write the answers in the same place as the code review. That creates a compact decision record and lets the team compare later runs without relying on memory.

A practical acceptance test

Run the workflow from a clean checkout at a recorded commit. Give the agent only the documented task packet and the intended permission profile. Then ask a reviewer who did not launch the run to reproduce the important checks, explain the changed behavior, and identify the rollback path. The task passes only when the artifact, evidence, and repository state agree. Keep the failed examples as regression cases; they are more useful than a polished demo because they reveal where instructions, environment, permissions, or tests need improvement.

Primary references

Vendor features, limits, preview labels, and pricing can change. Recheck the linked first-party documentation for the current state before making a purchase or rollout decision.