Skills for AI coding assistants (Claude Code, Cursor, etc.) that provide Databricks-specific guidance.
For Claude Code:
databricks experimental aitools skills installThis installs skills to ~/.claude/skills/ for use with Claude Code.
For Cursor:
Run this command in chat:
/add-plugin databricks-skills
- databricks-apps - Build full-stack TypeScript apps on Databricks using AppKit
Each skill follows the Agent Skills Specification:
skill-name/
├── SKILL.md # Main skill file with frontmatter + instructions
└── references/ # Additional documentation loaded on demand
When experimenting with new skill variations, create a "subskill" that references the main skill and adds specific guidance:
---
name: "ai-databricks-apps"
description: "Databricks apps with AI features"
---
# AI powered Databricks Apps
First, load the base databricks-apps skill for foundational guidance.
Then apply these additional patterns:
- Custom pattern 1
- Custom pattern 2This approach:
- Keeps the main skill stable and focused
- Allows experimentation without modifying core skills
- Makes it easy to follow the changes in the main skill
Generate manifest after adding/updating skills:
python3 scripts/generate_manifest.pyValidate that manifest is up to date (for CI):
python3 scripts/generate_manifest.py validateThe manifest is used by the CLI to discover available skills.
Please see SECURITY for vulnerability reporting guidelines.
All future release tags will be GPG-signed and verifiable via git tag -v <tag>.
- All changes require approval from a code owner (see CODEOWNERS).
- Documentation examples must follow least-privilege defaults — avoid suggesting elevated permissions or broad scopes unless explicitly necessary.