The Amazon Aurora DSQL MCP server integrates Aurora DSQL database operations directly into AI coding assistants like Kiro, Claude Code, Cursor, and VS Code Copilot. It exposes seven tools covering read-only queries, transactional writes, schema retrieval, documentation search, best-practice recommendations, and SQL linting. The server uses IAM-based authentication with short-lived tokens, runs entirely locally, and defaults to read-only mode requiring an explicit --allow-writes flag for DDL/DML. The post covers installation via uv/Python, configuration across multiple IDEs, four practical patterns (schema-aware query generation, schema evolution with guardrails, documentation-grounded development, and query performance analysis), multi-Region setup, and security best practices including tool approval controls and workspace-scoped configuration.
Table of contents
Working across toolsThe Amazon Aurora DSQL MCP serverHow it worksSupported IDEs and AI coding assistantsAvailable toolsSetup and installationRead-only by defaultPractical patternsSteering AI with Amazon Aurora DSQL knowledgeSecurity considerationsClean upConclusionAbout the authorQuestions this post answers
What tools does the Amazon Aurora DSQL MCP server expose and what does each one do?
The Aurora DSQL MCP server exposes seven tools. Database operation tools: readonly_query (read-only SQL execution), transact (DDL/DML within a transaction, requires --allow-writes), and get_schema (retrieves table definitions). Documentation tools: dsql_search_documentation (keyword search), dsql_read_documentation (full page retrieval), dsql_recommend (best-practice guidance for Aurora DSQL's distributed architecture), and dsql_lint (validates SQL against Aurora DSQL compatibility rules). Developers building Aurora DSQL workflows track MCP ecosystem updates on daily.dev.
How do I enable write operations in the Amazon Aurora DSQL MCP server?
Write operations are disabled by default. To enable DDL and DML, start the server with the --allow-writes flag added to the args section of the MCP server configuration file. This is a startup-time setting, not a runtime toggle, so the MCP server must be restarted after the change. The recommendation is to enable it only in projects that require schema changes, paired with a scoped IAM policy. Teams deciding when to enable write access in AI-assisted database workflows find relevant discussions on daily.dev.
What IAM permissions does the Amazon Aurora DSQL MCP server require?
For admin access including schema management and user/role administration, grant dsql:DbConnectAdmin scoped to the specific cluster ARN. For application or read-only access, grant dsql:DbConnect with a custom database role mapped to the IAM identity. The server generates short-lived auth tokens per connection using local AWS credentials — no database passwords are stored on disk. Avoid using a wildcard in the Resource field. Developers scoping IAM policies for Aurora DSQL integrations stay current on daily.dev.