databricks-workspace-ops

# Databricks Workspace Ops Skill ## Purpose Manage Databricks workspace artifacts via the Databricks CLI (v0.205+). Covers workspace ls/import/export, repos, secrets, files, and notebooks. ## Owner databricks-cli-operator ## Preconditions - Databricks CLI v0.205+ installed (`databricks --version`) - Authentication configured via profile or environment variables - `DATABRICKS_HOST` and `DATABRICKS_TOKEN` (or profile) set - Target workspace accessible ## Covered Operations ### Workspace - `databricks workspace ls <path>` — list workspace objects - `databricks workspace import <local-path> <remote-path>` — import notebook/file - `databricks workspace export <remote-path> <local-path>` — export notebook/file - `databricks workspace mkdirs <path>` — create workspace directory - `databricks workspace delete <path>` — delete workspace object (destructive) ### Files (DBFS and Unity Catalog Volumes) - `databricks fs ls <path>` — list files - `databricks fs cp <src> <dst>` — copy files (local to remote, remote to local) - `databricks fs mkdir <path>` — create directory - `databricks fs rm <path>` — remove file (destructive) ### Repos - `databricks repos list` — list Git repos - `databricks repos create --url <url> --provider <provider>` — create repo link - `databricks repos update --repo-id <id> --branch <branch>` — update repo branch - `databricks repos delete --repo-id <id>` — delete repo link (destructive) ### Secrets - `databricks secrets list-scopes` — list secret scopes - `databricks secrets list --scope <scope>` — list secrets in scope (metadata only) - `databricks secrets create-scope --scope <scope>` — create secret scope - `databricks secrets put-secret --scope <scope> --key <key>` — store secret (stdin) - `databricks secrets delete-secret --scope <scope> --key <key>` — delete secret (destructive) ## Disallowed Operations - Interactive notebook execution (requires Databricks UI or Jobs API) - Direct SQL execution (use databricks-identity-sql-ops for SQL warehouses) - Cluster management (use databricks-compute-jobs-ops) ## Output Contract - All commands use `--output json` for programmatic consumption - Workspace listing returns array of objects with `object_type`, `path`, `language` - File operations return success/failure status - Secret operations never expose secret values in output ## Verification - After import: `databricks workspace ls <path>` confirms object exists - After file copy: `databricks fs ls <path>` confirms file exists - After repo create: `databricks repos list` confirms repo appears