Skip to main content

Dev Container

The scaffold ships a ready-to-use Dev Container, and we recommend developing inside it rather than on your host:

  • Reproducible, pinned toolchain. Python, Node, the RevOS CLI, Google Cloud CLI, gh, jq, dbt-bigquery, and the connector test runtime are baked into the image at known versions — every contributor gets an identical environment, with no "works on my machine."
  • Zero host credentials. The container shares nothing from your host. revos init provisions a project-scoped GCP service-account key into the container's ~/.revos/ only; credentials persist in named Docker volumes (not bind mounts), so they survive Rebuild Container without ever touching your host filesystem.
  • Setup is automatic. postStartCommand re-activates the service account on every start, so bq, dbt, and gcloud work with no manual steps — and the connector test loop (revos connectors check/discover/read) runs out of the box.
  • AI skills stay current. The container re-installs/refreshes the RevOS skills on every start (self-healing if a prior install failed), so your AI assistant always has the latest authoring guidance.
  • Guardrails on by default. .claude/settings.json deny-rules stop AI assistants from reading credential paths (~/.revos/, ~/.config/gcloud/) or editing their own restrictions.

You can develop outside the container, but then you own installing and version-matching all of the above yourself.

Getting in​

After revos init, open the project in VS Code and click Reopen in Container. Inside the container, finish setup with a single revos init — it signs you in via the browser (if needed) and materializes the GCP service-account key into the container's ~/.revos/. See Project Scaffolding for the full onboarding flow.