MCP 服务器

MCP Spectral

An MCP server that wraps Spectral to lint OpenAPI specifications, enabling LLMs to validate and fix API definitions.

README

MCP Spectral

Model Context Protocol (MCP) Server for Spectral to lint OpenAPI specifications

Overview

Allow Large Language Models (LLM) to check whether an OpenAPI Specifications is valid using a Model Context Protocol (MCP) Server.

The MCP Spectral Server wrappers the Spectral command line within a Docker image.

Using an OpenAPI specification to describe an API provides consistency across APIs, the ability to document the API, a contract on how to use the API, the ability to generate code stubs to consume and produce APIs, and API versioning. These features enable quicker API development, quicker uptake of API services, and a reduction in API maintenance.

Using MCP Spectral should reduce API development time further by helping with the OpenAPI specification phase.

Available tools

Tools provided by MCP Spectral

lint_openapi - lints OpenAPI specification
lint_openapi_with_ruleset - lints OpenAPI specification with Spectral ruleset
spectral_version - returns version of Spectral being used

Set up instructions

These instruction assume Anthropic Claude Desktop will interact with the MCP Spectral server. If another Large Language Model tool is being used change the set up instructions accordlingly.

Installation prerequisites

Ensure the following applications are installed on the local machine

Update Claude Desktop

Add following to the claude_desktop_config.json file within the Claude Desktop installation which tells Claude Desktop how to access MCP Spectral

{
  "mcpServers": {
    "OpenAPI linting (Spectral)": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "marksivill/mcp-spectral"
      ]
    }
  }
}

For example if MCP Spectral is also installed with mcp/fetch the complete claude_desktop_config.json would be

{
  "mcpServers": {
    "fetch": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "mcp/fetch"
      ]
    },
    "OpenAPI linting (Spectral)": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "marksivill/mcp-spectral"
      ]
    }
  }
}

Once the file has been updated, restart the Claude Desktop.

Claude Desktop will automatically pull the latest docker version of MCP Spectral from Docker Hub

Local builds

Optionally the docker image can be built locally instead of pulling from Docker Hub using

docker build --tag marksivill/mcp-spectral:latest .

This is useful if the source code or default ruleset needs to be changed for example

Example Claude Desktop prompts

Below are example prompts for Claude Desktop which may also work in other Large Language Model (LLM) tools that are calling MCP Spectral.

Note mcp/fetch may also need to be installed into Claude Desktop to fetch URLs.

Prompt 1

What version of Spectral do we have access to

Prompt 2

Validate the OpenAPI specification from OpenWeatherMap API.
If any errors are found, fix them, then revalidate, repeat the last steps until the OpenAPI specification is valid.

Prompt 3

Using the API documentation found at https://wheretheiss.at/w/developer build an OpenAPI specification.
Ensure that there are no validation errors.
If any errors are found, fix them, then revalidate, repeat the last steps until the OpenAPI specification is valid.

Prompt 4

Build an OpenAPI specification version 3.1 for the Rick and Morty API.
Ensure that there are no validation errors.
If any errors are found, fix them, then revalidate, repeat the last steps until the OpenAPI specification is valid.

Prompt 5

Using the API documentation found at https://wheretheiss.at/w/developer build an OpenAPI specification version 3.1 using the rules
1) ensure all operationIds use kebab case
2) add info and contact information using your details
3) ensure servers are defined
4) ensure all operations contains a description
5) ensure all operations have one or more relevant tags
6) ensure there is a tag section

Ensure that there are no validation errors.
If any errors are found, fix them, then revalidate, repeat the last steps until the OpenAPI specification is valid.