Four new pieces on our blog if you want a walkthrough of what SQL-native cloud infrastructure actually looks like in practice: 1. Multi-cloud bucket audit, one SQL query across AWS, GCP, Azure to find exposed buckets stackql.io/blog/preview-buck… 2. Query before you mutate, how agents should read live cloud state before touching anything stackql.io/blog/query-before… 3. Continuous cloud audit in GitHub Actions, the audit pattern running in CI, catches drift between security reviews stackql.io/blog/continuous-c… 4. Declarative cloud stacks with stackql-deploy, manifest-driven infrastructure with idempotence baked in stackql.io/blog/declarative-… Cloud infrastructure and cloud APIs are already relational. Treating them that way lets you audit across three clouds in one query, gate agent mutations with policy checks written in SQL, run continuous drift detection in CI, and declare full stacks as manifest-plus-anchored-queries. All from a single query engine that ships as a signed binary. Star the repo if it lands for you: github.com/stackql/stackql

When agents write to your infrastructure, "plan, review, apply" starts to break down. Agents run continuously. Multiple actors touch the same resources. A state file that caches what one tool did last time can't reflect what's actually there right now. Drift stops being an exception and becomes the default. Tutorial 2 shows a different pattern: query before mutation. → Read live cloud state from the API → Apply a policy gate to bound what any single run can do → Mutate only what's out of policy → Verify Every run starts from reality instead of a cached view. Demonstrated end to end against Google Cloud Storage bucket encryption. Same pattern works from a shell or from an agent through stackql's MCP server in Claude Desktop. If the pattern makes sense to you, star the repo: github.com/stackql/stackql Full tutorial: stackql.io/blog/query-before… #MCP #DevOps #CloudSecurity #AgenticAI

Auditing storage buckets across AWS, GCP, and Azure usually means three CLIs, three auth setups, and glue code to reconcile the outputs. One stackql command replaces the whole flow. The setup is a Docker command against a compose file. The query is one SQL statement against `stackql_preview.audit.omni_storage_buckets`. Result: every bucket across three providers, side by side. Encryption class, public flag, HTTPS enforcement, all in one table. This is what the output looks like. Same columns for S3, GCS, and Azure Storage. No custom reconciliation needed. That's the point. One data model across providers. The same audit drops into GitHub Actions as a scheduled cross-cloud check. Ready-made workflows for all-clouds, single-cloud, deep-audit, and OIDC-authenticated variants are available in the `stackql-audit-action` repo. Full walkthrough on the StackQL blog, from Docker setup to the CI pattern: stackql.io/blog/auditing-thr… Star the `stackql` repo if useful: github.com/stackql/stackql

Edgepilot Demo youtu.be/btCUiq1v29c?si=E6gu… via @YouTube

Great meetup at the #brisbanegdg bit.ly/42eHGH5 #stackql #xops #GoogleCloud

Latest Google/GCP provider for StackQL available #GCP #GoogleCloud #sql #stackql

Kubernetes provider available, query k8s resources (nodes, pods, services, namespaces, etc) using StackQL