execute_query
Executes a SQL query on the Snowflake database.
Parameters: connection (SnowflakeConnection): The Snowflake connection object. query (str): The SQL query to execute.
Returns: result: The result of the query execution.
LABSCO SUMMARY
One tool means the model writes the SQL, and the grants on the connection are the only boundary.
execute_query takes a query and nothing else — no schema listing, no row cap, no separate read-only path — so whatever the Snowflake connection is allowed to do is what a client can do through it. That puts the interesting decision before installation rather than after it: the account and role you point it at define the blast radius. In return there is nothing to learn, since anything expressible in SQL is already in scope.
What it is
A single-tool bridge to a Snowflake database: one SQL query in, its result out.
What you get
- execute_query runs a SQL query on the Snowflake database over the connection the server holds and returns the result of the execution.
Requirements
A Snowflake database and a connection for the server to run queries over.
Setup effort
Build from source — git clone https://github.com/davidamom/snowflake-mcp.git
USE CASES
01
Answering a question whose data only lives in Snowflake
Write the SQL, run it over the server's connection, and read the result back in the same conversation.
1 tool
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool when we connected; the project documents 2.
Requires
- Python with pip (or Docker for the Docker route)
- SNOWFLAKE_USER, SNOWFLAKE_ACCOUNT, SNOWFLAKE_DATABASE, SNOWFLAKE_WAREHOUSE and SNOWFLAKE_ROLE for your Snowflake account
- SNOWFLAKE_PASSWORD, or SNOWFLAKE_PRIVATE_KEY_FILE for key-pair authentication
Optional
SNOWFLAKE_PRIVATE_KEY_PASSPHRASE in your environment
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/davidamom/snowflake-mcp.git
cd snowflake-mcp
pip install -r requirements.txtMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"snowflake": {
"command": "/usr/bin/python3",
"args": [
"<YOUR_PATH_TO_SNOWFLAKE_MCP>/server.py"
]
}
}
}/usr/bin/python3 has to be the Python you installed the requirements into, and <YOUR_PATH_TO_SNOWFLAKE_MCP> is the full path of your clone.
The Snowflake connection is set outside these blocks: SNOWFLAKE_USER, SNOWFLAKE_ACCOUNT, SNOWFLAKE_DATABASE, SNOWFLAKE_WAREHOUSE and SNOWFLAKE_ROLE, plus SNOWFLAKE_PASSWORD, or for key-pair authentication SNOWFLAKE_PRIVATE_KEY_FILE (and SNOWFLAKE_PRIVATE_KEY_PASSPHRASE if the key is encrypted). The Python route reads them from a .env file in the root of your clone; the Docker block passes them in its env.
Any MCP client — Docker
ANY MCP CLIENT — DOCKER · TERMINAL
git clone https://github.com/davidamom/snowflake-mcp.git
cd snowflake-mcp
docker build -t snowflake-mcp .ANY MCP CLIENT — DOCKER · JSON
{
"mcpServers": {
"snowflake-docker": {
"command": "docker",
"args": [
"run",
"-i",
"snowflake-mcp"
],
"env": {
"SNOWFLAKE_USER": "<YOUR_SNOWFLAKE_USER>",
"SNOWFLAKE_ACCOUNT": "<YOUR_SNOWFLAKE_ACCOUNT>",
"SNOWFLAKE_DATABASE": "<YOUR_SNOWFLAKE_DATABASE>",
"SNOWFLAKE_WAREHOUSE": "<YOUR_SNOWFLAKE_WAREHOUSE>",
"SNOWFLAKE_PASSWORD": "<YOUR_SNOWFLAKE_PASSWORD>",
"SNOWFLAKE_ROLE": "<YOUR_SNOWFLAKE_ROLE>"
}
}
}
}The Snowflake connection is set outside these blocks: SNOWFLAKE_USER, SNOWFLAKE_ACCOUNT, SNOWFLAKE_DATABASE, SNOWFLAKE_WAREHOUSE and SNOWFLAKE_ROLE, plus SNOWFLAKE_PASSWORD, or for key-pair authentication SNOWFLAKE_PRIVATE_KEY_FILE (and SNOWFLAKE_PRIVATE_KEY_PASSPHRASE if the key is encrypted). The Python route reads them from a .env file in the root of your clone; the Docker block passes them in its env.
