Schema
Value Proposition Secure your AI operations with strict boundary controls and standardized schema validation, improving overall security and data integrity. Read the full value proposition.
Schema validation is treated as a strict data boundary. Every input parameter is strictly validated, ensuring reliable workflows.
These schemas are strictly enforced at the MCP boundary for secure integrations:
-
Strict Error Responses: Failing tools return a structured
MySQLMcpErrorobject. This empowers autonomous workflows to recover gracefully from failures. -
Deterministic Table Queries: Table queries return
{ exists: false, table: string }for missing tables. This enables graceful fallback handling in AI workflows. -
Sandboxed File Operations: File interactions are bounded using
ALLOWED_IO_ROOTS. This prevents directory traversal and ensures isolated, secure execution. - Safe Parsing & Interoperability: Tools use strict standard schema validation libraries for safe parsing. This supports aliases and type coercion while preserving JSON Schema compatibility.
- Static Schemas: All schemas must be defined statically as constants at the module level.
-
Resilient Environment Validation: Strictly validates all
process.envconfigurations at startup using strict standard schema validation libraries to proactively prevent runtime misconfigurations and ensure consistent deployments. - Distributed Rate Limiting: HTTP requests and Code Mode executions are strictly rate limited, with automatic Redis synchronization and in-memory fallbacks.
See the repository source for Zod schemas. It is recommended to use the Standard Schema interface (schema['~standard'].validate(input)) to handle validation errors gracefully without exceptions.
For runtime schema introspection, see the Code Mode documentation. It lets you discover API parameters programmatically using the .schema() introspection method.
- Code Mode - Code Mode API documentation
- Configuration - MCP client configuration