This is a really neat approach! In API security tooling, we've been experimenting with analyzing OpenAPI specs semantically to detect risky endpoints before deployment. It’s interesting to see a system that auto-generates endpoints and docs while keeping sensitive connection info server-side. Curious how you handle complex query logic or multi-step operations in production? Could similar intent-aware checks help catch risky edge cases?