Databricks
Execute SQL statements and manage Unity Catalog catalogs and tables in a Databricks workspace.
About
Databricks ships in the w6w first-party pack. It declares 8 actions, 2 health checks, and the host runs its code in a sandbox that never sees the credential.
- App id
io.w6w.databricks- Version
- 0.1.3
- Author
- w6w
- Licence
- MIT
- Categories
- Data Warehousing & ETL
Overview
Databricks runs SQL directly against a workspace from a workflow — submitting a statement and getting back either a finished result or a pending status with an id to poll, so a workflow can wait on a query without holding a connection open the whole time.
Unity Catalog is a first-class part of the surface too: catalogs can be listed, fetched, created and deleted, and tables can be listed and inspected, so a workflow can discover what data exists in a workspace rather than assuming a fixed set of table names.
A fit for pulling data out of a lakehouse for use elsewhere in a workflow, running scheduled or triggered SQL against a workspace, or managing catalog structure as part of provisioning a new data source.
Build with Databricks
Three routes to the same 8 actions. The Workflow tab is generated from Databricks's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
Execute SQL Statement
sql-statement-execute Run a SQL statement against a SQL warehouse. May return PENDING/RUNNING for long queries — poll with Get SQL Statement.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Databricks's real ids.
{
"manifestVersion": "2",
"name": "databricks-example",
"steps": [
{
"id": "catalog-create",
"uses": {
"app": "io.w6w.databricks",
"action": "catalog-create",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"name": "<name>"
}
}
]
}Here are some of the things you can do
-
Create Catalog
performcatalog-create -
Get Catalog
readcatalog-get -
List Catalogs
searchcatalog-list -
Execute SQL Statement
performsql-statement-execute -
Get SQL Statement
readsql-statement-get
+3 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Databricks, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Databricks. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "sql-statement-execute",
payload: {
warehouseId: "<value>",
statement: "<value>",
// catalog: "<value>",
// schema: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value);npm install -g @w6w/cliw6w run conn_YOUR_CONNECTION_ID --action sql-statement-execute --payload '{"warehouseId":"<value>","statement":"<value>"}'Give an AI agent Databricks — without giving it Databricks's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.databricks#catalog-create",
"input": {
"name": "<name>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Databricks connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
What the agent gets
Credentials it can't read
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
A tool surface scoped to the caller
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
A durable workflow in one call
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Health-aware discovery
Databricks's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Databricks itself is MIT, and the runtime that executes it is source-available (FSL).
Health checks
Databricks declares its own checks, so its health is a property of the app rather than something the host guesses at.
Databricks platform status
Declared absent — Databricks workspaces are per-customer deployments with no aggregate status feed. See the `workspace` dependency check for this connection's own reachability.
Workspace reachable
Unauthenticated request to this connection's Databricks workspace host. A 401 passes — it proves the workspace is serving; credential validity is the `auth:*` check's job.