Schema

Schema

🚀 Value Proposition

Create strong API boundaries with our Enterprise-grade validation engine. By wielding Zod 3 for type-safety and natively supporting the Standard Schema specification, we guarantee flawless interoperability across all modern AI frameworks. Our Dual-Schema Pattern ensures strict .safeParse() execution, delivering "strict schema validation" and powering autonomous, self-healing agent workflows with confidence.

🛡️ Enforce Zod Best Practices

When building extensions or custom tools, adhere strictly to these Zod patterns:

• Strict Validation: Always use .strict() on root objects to prevent unrecognized properties from silently passing.
• Type Discrimination: Prefer discriminated unions (z.discriminatedUnion) over simple unions (z.union) for predictable error reporting.
• Safe Parsing: Always use schema.safeParse() instead of .parse() at I/O boundaries to avoid throwing exceptions and to gracefully handle invalid inputs.

Enforce Reliability: Enterprise Validation Rules

To provide a secure and robust integration experience, we strictly enforce the following schemas at the MCP boundary:

1. Strict Error Responses: All tools are guaranteed to return a structured error response upon failure. All errors extend MySQLMcpError (9 categories) and return enriched ErrorResponse via formatHandlerError(). Our resilient design ensures you will never encounter raw, unhandled exceptions, empowering self-healing autonomous workflows. The schema is: { success: false, error: string, code: string, category: string, suggestion: string, recoverable: boolean }
2. Deterministic Table Queries: Tools that query tables will return { exists: false, table: string } if the table does not exist, enabling graceful fallback handling in your LLM workflows.
3. Sandboxed File Operations: File system interactions are strictly bounded by ALLOWED_IO_ROOTS to prevent directory traversal and ensure isolated, secure execution.
4. Dual-Schema Pattern: Tools use a plain z.object() Base schema for MCP parameter visibility, and a z.preprocess() wrapper for handler parsing to support aliases and coercion without breaking standard JSON Schema generation.

Build with Confidence: Integration Guide

Please refer to the repository source for Zod schemas (src/adapters/mysql/schemas/). We highly recommend using schema.safeParse() when building custom MCP tool execution handlers or client integrations to handle validation errors gracefully without throwing exceptions.

For runtime schema introspection, see the Code Mode documentation, which allows you to programmatically discover API parameters and return types using the .schema() introspection method.