mcp-lint

Analyzes MCP server tool schema definitions against a comprehensive ruleset to detect structural violations, naming inconsistencies, type mismatches, and compatibility issues before runtime. Uses AST-like traversal of JSON schema objects to validate against MCP specification constraints, identifying issues like missing required fields, invalid parameter types, malformed descriptions, and schema patterns that would cause client incompatibility.

Performs pre-execution validation of tool invocation requests before they reach the actual tool handler, checking that provided arguments match the schema definition, required parameters are present, and types conform to declared specifications. Intercepts tool calls at the MCP protocol layer and validates against the registered schema, returning structured validation errors that prevent malformed calls from executing and causing runtime failures.

Analyzes tool schemas to identify features or patterns that may not be supported by all MCP clients, such as advanced parameter types, nested object structures, or client-specific extensions. Generates a compatibility matrix showing which schema features are supported by different MCP client implementations and versions, helping developers understand where their tools may fail or degrade gracefully.

Enables definition and enforcement of custom policies that govern which tools can be called, under what conditions, and with what parameter constraints. Policies are defined declaratively (e.g., 'only allow file operations on paths under /tmp', 'require approval for network calls') and evaluated at runtime before tool execution, blocking or modifying calls that violate policy rules.

Validates that tool schemas include complete, consistent, and well-formed documentation across all tools in a server. Checks for missing descriptions, inconsistent terminology, formatting violations, and ensures documentation follows a defined style guide. Generates reports highlighting documentation gaps and suggests standardized descriptions based on tool patterns.

Processes multiple MCP server schemas in batch mode, generating comprehensive validation reports across all servers and tools. Supports batch validation of schema files, directories, or remote schema registries, producing aggregated reports with cross-server consistency checks and trend analysis over time.

Analyzes schemas to identify patterns that may cause issues with specific LLM agents (Claude, GPT-4, etc.) and their tool-calling implementations. Generates agent-specific warnings about schema features that particular agents handle poorly, such as deeply nested parameters, ambiguous type unions, or parameter descriptions that might confuse specific model versions.