mcp-schema-lint

Validates tool and resource schema definitions against the Model Context Protocol specification using a schema parser that checks structural conformance, required fields, type correctness, and naming conventions. The linter parses JSON/YAML schema files and compares them against MCP's official schema definitions to catch malformed or non-compliant schemas before deployment.

Processes multiple schema files in a single CLI invocation, recursively scanning directories or processing file globs to validate entire schema repositories. The linter aggregates results across files and produces consolidated reports showing which files pass/fail validation with detailed error locations.

Generates human-readable error messages that pinpoint exactly where schema violations occur, including file paths, line numbers, column positions, and contextual snippets of the problematic schema. Errors are categorized by type (missing required field, type mismatch, naming convention violation, etc.) to help developers quickly understand and fix issues.

Exposes schema validation as a command-line tool with configurable output formats (text, JSON, TAP) and standard exit codes (0 for success, non-zero for failures) that integrate seamlessly with shell scripts, CI/CD systems, and build pipelines. Supports flags for controlling verbosity, output destination, and validation strictness.

Validates that tool names, resource names, parameter names, and other identifiers in MCP schemas follow MCP's naming conventions (e.g., snake_case for parameters, specific patterns for tool names). Checks against a configurable set of naming rules that align with MCP best practices and protocol requirements.

Validates that parameter and response types in tool schemas conform to MCP's supported type system (string, number, boolean, object, array, etc.) and that type definitions are properly structured. Checks for type mismatches, unsupported types, and malformed type declarations that would cause runtime failures.