Claude-code-plugins databricks-install-auth

install

source · Clone the upstream repo

git clone https://github.com/jeremylongshore/claude-code-plugins-plus-skills

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/jeremylongshore/claude-code-plugins-plus-skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/plugins/saas-packs/databricks-pack/skills/databricks-install-auth" ~/.claude/skills/jeremylongshore-claude-code-plugins-databricks-install-auth && rm -rf "$T"

manifest: plugins/saas-packs/databricks-pack/skills/databricks-install-auth/SKILL.md

Databricks Install & Auth

Overview

Set up Databricks CLI v2, Python SDK, and authentication. Covers Personal Access Tokens (legacy), OAuth U2M (interactive), and OAuth M2M (service principal for CI/CD). Databricks strongly recommends OAuth over PATs for production.

Prerequisites

Python 3.8+ with pip

Databricks workspace URL (e.g.,

https://adb-1234567890123456.7.azuredatabricks.net

)

For PAT: User Settings > Developer > Access Tokens in workspace UI
For OAuth M2M: Service principal with client ID and secret

Instructions

Step 1: Install Databricks CLI and Python SDK

set -euo pipefail

# Install CLI v2 (standalone binary — recommended)
curl -fsSL https://raw.githubusercontent.com/databricks/setup-cli/main/install.sh | sh

# Verify CLI
databricks --version

# Install Python SDK
pip install databricks-sdk

# Install Databricks Connect for local Spark development
pip install databricks-connect==14.3.*

Step 2: Configure Authentication

Option A: Personal Access Token (Quick Start)

Generate a PAT in workspace UI: User Settings > Developer > Access Tokens.

# Interactive setup — prompts for host and token
databricks configure --token

# Or set environment variables directly
export DATABRICKS_HOST="https://adb-1234567890123456.7.azuredatabricks.net"
export DATABRICKS_TOKEN="dapi_your_token_here"

Option B: OAuth U2M (User-to-Machine — Interactive)

Opens browser for OAuth consent. Token auto-refreshes (1-hour lifetime).

# Interactive OAuth login
databricks auth login --host https://adb-1234567890123456.7.azuredatabricks.net

# Verify — prints current user
databricks current-user me

Option C: OAuth M2M (Service Principal — CI/CD)

Uses client credentials flow. No browser required. Create a service principal in Account Console > Service Principals, then generate an OAuth secret.

export DATABRICKS_HOST="https://adb-1234567890123456.7.azuredatabricks.net"
export DATABRICKS_CLIENT_ID="00000000-0000-0000-0000-000000000000"
export DATABRICKS_CLIENT_SECRET="dose00000000000000000000000000000000"

# Verify
databricks current-user me

Step 3: Configure Profiles for Multi-Workspace

# ~/.databrickscfg — one section per workspace
[DEFAULT]
host  = https://adb-dev-workspace.7.azuredatabricks.net
token = dapi_dev_token_here

[staging]
host  = https://adb-staging-workspace.7.azuredatabricks.net
token = dapi_staging_token_here

[production]
host       = https://adb-prod-workspace.7.azuredatabricks.net
client_id  = 00000000-0000-0000-0000-000000000000
client_secret = dose_prod_secret_here

# Use a specific profile
databricks workspace list / --profile staging

Step 4: Verify SDK Connection

from databricks.sdk import WorkspaceClient

# Auto-detects from env vars or ~/.databrickscfg
w = WorkspaceClient()

me = w.current_user.me()
print(f"Authenticated as: {me.user_name}")
print(f"Workspace: {w.config.host}")
print(f"Auth type: {w.config.auth_type}")

# Quick smoke test — list clusters
clusters = list(w.clusters.list())
print(f"Clusters found: {len(clusters)}")

Step 5: Service Principal Authentication (Python SDK)

from databricks.sdk import WorkspaceClient
from databricks.sdk.config import Config

# Explicit M2M config for CI/CD scripts
config = Config(
    host="https://adb-1234567890123456.7.azuredatabricks.net",
    client_id="00000000-0000-0000-0000-000000000000",
    client_secret="dose00000000000000000000000000000000",
)
w = WorkspaceClient(config=config)

# Or use a named profile
w = WorkspaceClient(profile="production")

Output

Databricks CLI v2 installed and on PATH
Python SDK (
```
databricks-sdk
```
) installed
Authentication credentials stored in env vars or
```
~/.databrickscfg
```
Connection verified with
```
databricks current-user me
```

Error Handling

Error	Cause	Solution
`INVALID_TOKEN`	Token expired or revoked	Generate a new PAT or re-run `databricks auth login`
`Could not resolve host`	Wrong workspace URL	Verify URL format: `https://adb-<id>.<region>.azuredatabricks.net`
`PERMISSION_DENIED`	Token lacks required entitlements	Ensure user/SP has workspace access in Account Console
`SSL: CERTIFICATE_VERIFY_FAILED`	Corporate proxy intercepts TLS	Set `REQUESTS_CA_BUNDLE=/path/to/cert.pem`
`Connection refused`	VPN or firewall blocking	Check corporate firewall rules for workspace domain
`No matching profile`	Profile name typo in `~/.databrickscfg`	Run `databricks auth profiles` to list available profiles

Examples

Account-Level Client (Multi-Workspace Management)

from databricks.sdk import AccountClient

# Account-level operations (manage workspaces, users, billing)
a = AccountClient(
    host="https://accounts.cloud.databricks.com",
    account_id="00000000-0000-0000-0000-000000000000",
    client_id="sp-client-id",
    client_secret="sp-secret",
)

for ws in a.workspaces.list():
    print(f"{ws.workspace_name}: {ws.deployment_name}")

Azure AD Managed Identity

from databricks.sdk import WorkspaceClient

# Uses Azure Default Credential chain (works in Azure VMs, AKS, Functions)
w = WorkspaceClient(
    host="https://adb-1234567890123456.7.azuredatabricks.net",
    azure_workspace_resource_id="/subscriptions/<sub>/resourceGroups/<rg>/providers/Microsoft.Databricks/workspaces/<ws>",
)

Resources

Next Steps

After successful auth, proceed to

databricks-hello-world

for your first cluster and notebook.