shacl-mcp

MCP (Model Context Protocol) server for SHACL validation.

Overview

This crate provides an MCP server that exposes SHACL validation functionality through the Model Context Protocol. It allows AI assistants and other MCP clients to validate RDF data against SHACL shapes, parse shapes, and lint RDF graphs.

Features

The server exposes the following tools:

validate_graphs

Validate RDF data against SHACL shapes and return a validation report.

Parameters:

• dataGraph: RDF data graph as a string
• shapesGraph: SHACL shapes graph as a string
• dataFormat: Format of the data graph (e.g., 'ttl', 'nt', 'jsonld')
• shapesFormat: Format of the shapes graph (e.g., 'ttl', 'nt', 'jsonld')
• outputFormat: Format of the output report ('text', 'json', or RDF format like 'ttl')

Returns: Validation report in the specified format

validate_graphs_conforms

Check if RDF data conforms to SHACL shapes (returns only boolean result).

Parameters:

• dataGraph: RDF data graph as a string
• shapesGraph: SHACL shapes graph as a string
• dataFormat: Format of the data graph (e.g., 'ttl', 'nt', 'jsonld')
• shapesFormat: Format of the shapes graph (e.g., 'ttl', 'nt', 'jsonld')

Returns: { "conforms": true/false }

lint_graph

Validate RDF graph syntax.

Parameters:

• graph: RDF graph as a string
• format: Format of the graph (e.g., 'ttl', 'nt', 'jsonld')

Returns: { "valid": true } or error

parse_shapes_graph

Parse SHACL shapes graph and return parsed shape information.

Parameters:

• shapesGraph: SHACL shapes graph as a string
• shapesFormat: Format of the shapes graph (e.g., 'ttl', 'nt', 'jsonld')

Returns: Parsed shapes metadata including shape count and details

Installation

Building from Source

Build the MCP server:

cargo build --release -p shacl-mcp

The binary will be located at target/release/shacl-mcp.

Installation via Cargo

You can also install the MCP server globally using Cargo:

cargo install shacl-mcp

Usage

The MCP server communicates via JSON-RPC 2.0 over stdin/stdout. It's designed to be used by MCP clients such as Claude Desktop or other AI assistants.