SQL Changelog Reviewer

Before any migration reaches production, a thorough review of the changelog or migration file can catch errors that would otherwise cause outages, data loss, or silent corruption. This AI assistant acts as your expert SQL changelog reviewer, analyzing migration files with the same rigor a senior DBA would apply during a formal code review.

You can paste a Liquibase changelog (in XML, YAML, JSON, or SQL format), a Flyway migration script, a raw SQL file, or any migration artifact, and the assistant will perform a multi-dimensional review. It checks for syntactic correctness, semantic risks, missing rollback definitions, transaction handling, locking implications, and compliance with version control best practices.

The assistant identifies common anti-patterns such as missing WHERE clauses in UPDATE or DELETE statements, column drops without a deprecation period, index creation without CONCURRENTLY on PostgreSQL, implicit data type coercions that could fail on existing data, and changesets without preconditions in Liquibase. It also checks for stylistic issues that make changelogs harder to maintain: missing descriptions, ambiguous changeset IDs, and overly large changesets that combine unrelated changes.

For each finding, the assistant provides a severity rating (blocking, warning, informational), a clear explanation of why the issue matters, and a concrete suggestion for how to fix it. Where a corrected version of the SQL or changeset is appropriate, it generates one.

Ideal users include developers submitting database changes for peer review, DBAs conducting formal migration reviews before production deployments, and CI/CD engineers building automated migration linting into their pipelines. The assistant can also be used to establish a review checklist for your team's migration review process.