Connect Databricks to Your AI Agent

Find the failed task and the evidence needed for the next investigation step.

Try a Databricks Task

Copy a prompt into your agent with the oo Skill installed. Add your page links and source notes before running it.

oo Skill Databricks

Explain recent job failures

Summarize execution states and useful error details for a selected job.

/oo Databricks: Review job [job ID] during [time range]. Inspect failed runs and summarize affected tasks, errors, and follow-up checks.

Start a specified existing job

Run the chosen workload with the supplied parameters and report its run identifier.

/oo Databricks: Start existing job [job ID] with [parameters] for [purpose]. Check its current progress and tell me whether it has finished.

Start Working with Databricks

Install OOMOL

Set up oo CLI and the oo Skill in your agent.
Install OOMOL →
Connect Databricks

Connect with a Databricks personal access token and workspace URL.
Connect Databricks →
Give Your Agent a Task

Use /oo with the pages to read, the changes to make, and where to save the result.
Connection Guide →

Choose Your Agent

Codex Claude Code Cursor Hermes

The token needs access to the jobs and runs you want to inspect.
Starting a job executes its workload and may incur compute costs. Run output depends on the workload and may be limited.

Databricks Tools

Inspect job definitions, run states, and available output, or start an existing workload with supplied parameters.

Job and run inspection

Read workload definitions, execution states and run output.

Existing job execution

Start a job with its identifier and optional workload parameters.

36 tools

Get Current User

Get the current Databricks workspace principal profile for the connected personal access token.

databricks.get_current_user

List Jobs

List Databricks jobs in the connected workspace with optional pagination, name filtering, and task expansion.

databricks.list_jobs

Get Job

Get one Databricks job, including its current settings and optional continuation page token.

databricks.get_job

Create Job

Create a Databricks job from a raw official Jobs API settings object and return the new job ID.

databricks.create_job

Update Job By Id

Update an existing Databricks job by ID using the official partial-update payload fields.

databricks.update_job_by_id

Delete Job

Delete a Databricks job by ID.

databricks.delete_job

Run Now Job

Trigger an immediate run for a Databricks job with optional parameter maps and idempotency token.

databricks.run_now_job

List Runs

List Databricks job runs with optional job, state, time-range, and pagination filters.

databricks.list_runs

Get Run By Id

Get one Databricks job run by run ID.

databricks.get_run_by_id

Get Run Output

Get the output payload for one Databricks run, including notebook output or error details when available.

databricks.get_run_output

Cancel Run

Cancel a Databricks job run by run ID.

databricks.cancel_run

Submit Run

Submit a one-time Databricks run from a raw official Jobs API submit payload and return the run ID.

databricks.submit_run

List Clusters

List Databricks clusters with optional paging, filtering, and sorting controls from the official clusters API.

databricks.list_clusters

Get Cluster

Get one Databricks cluster by cluster ID.

databricks.get_cluster

Create Cluster

Create a Databricks cluster from a raw official clusters create payload and return the new cluster ID.

databricks.create_cluster

Edit Cluster

Edit an existing Databricks cluster by cluster ID using a raw official clusters edit payload.

databricks.edit_cluster

Start Cluster

Start a terminated Databricks cluster by cluster ID.

databricks.start_cluster

Permanent Delete Cluster

Permanently delete a Databricks cluster by cluster ID.

databricks.permanent_delete_cluster

List Cluster Node Types

List Databricks cluster node types that are available in the connected workspace region.

databricks.list_cluster_node_types

List Cluster Zones

List Databricks cluster availability zones for the connected workspace when the API is supported.

databricks.list_cluster_zones

List Cluster Spark Versions

List Databricks Runtime and Spark versions that can be used when creating or editing clusters.

databricks.list_cluster_spark_versions

Workspace List

List the direct Databricks workspace objects under a workspace path.

databricks.workspace_list

Workspace Get Status

Get metadata for one Databricks workspace object by absolute path.

databricks.workspace_get_status

Workspace Export

Export one Databricks workspace object as base64 content or direct text content, depending on the export mode.

databricks.workspace_export

Workspace Import

Import base64 content into the Databricks workspace at an absolute path.

databricks.workspace_import

Workspace Mkdirs

Create a Databricks workspace directory, including any missing parent directories.

databricks.workspace_mkdirs

Workspace Delete

Delete a Databricks workspace object or directory by absolute path.

databricks.workspace_delete

Create Repo

Create a Databricks workspace repo that links a Git remote into the connected workspace.

databricks.create_repo

Update Repo

Update a Databricks workspace repo by switching branch or tag, or by changing sparse-checkout settings.

databricks.update_repo

Delete Repo

Delete a Databricks workspace repo by repo ID.

databricks.delete_repo

List Secret Scopes

List Databricks secret scopes in the connected workspace.

databricks.list_secret_scopes

List Secrets

List Databricks secret metadata rows in one secret scope.

databricks.list_secrets

Create Secret Scope

Create a Databricks secret scope with an optional backend configuration and initial manager principal.

databricks.create_secret_scope

Delete Secret Scope

Delete a Databricks secret scope by scope name.

databricks.delete_secret_scope

Put Secret

Create or overwrite a Databricks secret value inside one Databricks secret scope.

databricks.put_secret

Delete Secret

Delete one Databricks secret value from a Databricks secret scope.

databricks.delete_secret

Frequently Asked Questions

Will starting a job use compute resources?

Yes. It executes the selected workload and may incur compute costs.

Can this connection run an arbitrary SQL query?

The available operations cover jobs, runs, and workspace resources. They do not include direct SQL warehouse queries.