Connect Snowflake to Cursor
Give Cursor governed access to Snowflake’s 8 tools through Runlayer, with identity, policy, scanning, and audit on every call.
Set up
Add Snowflake to Cursor in four steps
Admins approve the server once. Everyone in Cursor then gets it through the same governed connection.
- 1
Approve Snowflake in Runlayer
Add Snowflake from the catalog and choose who can use it and which of its tools are allowed.
- 2
Connect Cursor to Runlayer
Runlayer’s CLI installs the Runlayer Plugin natively in Cursor, and admins can roll it out org-wide with Auto Sync plus a Cursor team rule. With direct tool listing turned on for your deployment, Snowflake’s 8 tools show up in Cursor’s own tool list.
Cursor setup guide - 3
Sign in to Snowflake
Each user signs in to Snowflake with OAuth. Runlayer keeps the tokens out of the client’s configuration.
- 4
Use Snowflake from Cursor
Snowflake’s approved tools appear in Cursor. Every call is checked against policy and recorded.
Tools
8 Snowflake tools in Cursor
Runlayer policy decides which of these Cursor users can call.
Describe semantic view
Describe a semantic view.
Get semantic view ddl
Get the DDL for a semantic view.
List semantic views
List all semantic views in the account, database, or schema.
Query semantic view
Writes and runs a statement to query a semantic view using DIMENSIONS, METRICS, and/or FACTS.
Run Snowflake query
Run a SQL query in Snowflake.
Show semantic dimensions
Show all semantic dimensions in the account, database, or schema.
Show semantic metrics
Show all semantic metrics in the account, database, or schema.
Write semantic view query tool
Writes a query statement to query a semantic view using DIMENSIONS, METRICS, and/or FACTS.
Control
What Runlayer can see and stop in Cursor
Some controls apply to every client at the gateway. Others depend on what Cursor exposes to Runlayer.
Gateway policy (available)
Every Snowflake call from Cursor passes through Runlayer, where your policies can allow, alert, mask, or block it.
Audit trail (available)
Each call is recorded with the user, Cursor as the client, the tool, and the policy and scanner decisions.
Client hooks (available)
Capture supported prompts, reasoning, local and MCP tool calls, responses, and scanner decisions in real time.
Stop a call before it runs (available)
Runlayer can deny a tool call before Cursor runs it, including calls to MCP servers.
Block output after it runs (available)
Runlayer can block a tool’s output before Cursor’s model sees it.
Redact output in place (not available)
Cursor cannot take masked output, so Runlayer can block a tool’s output but not mask it. The gateway still applies policy to the call.
In Cursor
- Runlayer cannot tell Agents Window sessions apart from regular IDE sessions.
- The Cursor CLI sends fewer session events than the IDE, but Runlayer’s allow and deny decisions still apply.
Frequently asked questions
Connect more
