Databricks Apps

databricks/databricks-agent-skillsOfficial

Scaffolds and deploys apps to the Databricks Apps platform, handling everything from warehouse-backed dashboards to operational apps with Lakebase databases. The workflow forces you to pick your data access pattern upfront: analytics queries hit warehouses in a few seconds and work for dashboards and aggregations, while Lakebase synced tables give you sub-second Postgres lookups for typeahead and real-time search. It won't let you scaffold until you've made that call, which saves you from rebuilding later. Includes AppKit SDK hooks, tRPC routing, Genie chat integration, and smoke test scaffolding. The strict validation catches TypeScript errors and double-assertions before deploy. Useful when you need to ship internal tools or data apps without leaving the Databricks ecosystem.

skills.sh installs: 1.1K
GitHub stars: 393
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add databricks/databricks-agent-skills --skill databricks-apps --agent claude-code

Installs into .claude/skills of the current project.

Files

Databricks Apps Development

FIRST: Use the parent databricks-core skill for CLI basics, authentication, and profile selection.

Package manager — resolve this before any install or command below:

Databricks Apps officially supports npm and pnpm only.

Detect the project's manager, in priority order: project instructions → package.json's packageManager field → lockfile (pnpm-lock.yaml → pnpm, package-lock.json → npm).
If signals conflict, or if an unsupported manager (yarn, bun, etc.) is detected, stop and ask the user which manager to use before proceeding.
Only if none is established, default to pnpm.

Throughout this skill and its guides, substitute what you resolved: