Database Query

这是一个用于Coding Agent MCP 服务的数据库查询工具，可以通过提示词，从当前项目配置文件中读取数据库连接信息并连接多种类型的数据库（MySQL、PostgreSQL、MSSQL、Oracle）并执行 SQL 查询。**该工具运行在只读模式下，确保不会意外修改数据库中的数据**。

javascript

GitHub Stars

javascript

Language

4 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "xuejike-coding-db-mcp": {
      "command": "coding-db-mcp",
      "args": [],
      "env": {
        "DB_HOST": "YOUR_HOST",
        "DB_NAME": "YOUR_DATABASE",
        "DB_PORT": "YOUR_PORT",
        "DB_USER": "YOUR_USER",
        "DB_PASSWORD": "YOUR_PASSWORD"
      }
    }
  }
}

You can query multiple databases directly from your MCP workflow using the Database Query MCP Tool. It reads your project configuration to obtain connection details and can talk to MySQL, PostgreSQL, MSSQL, and Oracle databases in a safe read-only mode, returning structured results without modifying data.

How to use

Install and run the tool in your MCP workflow to perform read-only SQL queries across supported databases. When the MCP client invokes this service, you provide host, port, user, password, database name, and the query you want to run. The tool will ensure only read-only queries are executed and will return structured results with column metadata and row data.

How to install

Prerequisites: Node.js and npm installed on your system.

Option A: Install globally and run directly.

npm install -g coding-db-mcp

db-query-mcp

Option B: Run with npx without installation.

npx coding-db-mcp

Configure the MCP client to load this service as an MCP server. The following examples show how to register the service under the name database_query using either a direct command or an npx-based command.

Configuration and security

Database connection information can be picked up from your project configuration automatically. If you prefer manual configuration, you can supply a local environment file with the necessary credentials.

Security considerations: operate in read-only mode to automatically block any data modification statements. Restrict database user permissions so that credentials have only the necessary read access. Do not hard-code credentials in code, and enable SSL in production environments.

Examples and error handling

When a query succeeds, you receive a structured response containing success, data, columns, and rowCount. If an error occurs, you receive success: false along with an error message and a code.

Troubleshooting

If you encounter connection errors, verify network reachability to the database host, correct port numbers, and that the credentials have read access for the target database. Check SSL/TLS settings if your database requires encrypted connections.

Tools exposed by the MCP service

query_mysql: Execute read-only SQL on MySQL databases. Parameters: host, port, user, pwd, db, querySql.

query_postgresql: Execute read-only SQL on PostgreSQL databases. Parameters: host, port, user, pwd, db, querySql.

query_mssql: Execute read-only SQL on MSSQL databases. Parameters: host, port, user, pwd, db, querySql.

query_oracle: Execute read-only SQL on Oracle databases. Parameters: host, port, user, pwd, db, querySql.

Available tools

query_mysql

Executes a read-only SQL SELECT-style query against a MySQL database and returns structured results.

query_postgresql

Executes a read-only SQL SELECT-style query against a PostgreSQL database and returns structured results.

query_mssql

Executes a read-only SQL SELECT-style query against a MSSQL database and returns structured results.

query_oracle

Executes a read-only SQL SELECT-style query against an Oracle database and returns structured results.

Built by

VeilStrat

AI signals for GTM teams