MCP Advisor
Access the Model Context Protocol specification through prompts and resources.
MCP Advisor
⚠️Warning: currently this server depends on the official LLMS.txt file which no longer matches the format expected by this server. This server also still needs to be updated for the latest spec. For the time being, I recommend using the official docs server and pointing to the schemas in the MCP repo instead.
An MCP server that provides comprehensive access to the Model Context Protocol specification through both prompts and resources. This server helps LLMs and humans understand and work with the MCP specification by providing the complete JSON schema and detailed documentation.
When would I use this type of solution instead of web search or other RAG solutions?
While the same information is already available on the web, if you have a use case where precise spec information is preferred, directly fetching the spec details as context should provide a more reliable result.
Since these resources should easily fit within a model's context window, using a more complex RAG solution is not really necessary unless you need to compare documents to other documents, or do more complex types of querying.
Other ways to do the same thing
If you prefer a more ad-hoc approach you can also use an MCP server like fetch to do the following:
- Fetch the contents of https://modelcontextprotocol.io/llms.txt to get the list of valid links
- Fetch content from links that are relevant to the current task
Installation
# Install as a dependency
npm install mcp-advisor
# Or install globally to use the CLI
npm install -g mcp-advisor
Usage
With Claude Desktop
- Add the following to your Claude Desktop configuration file:
{
"mcpServers": {
"mcp-advisor": {
"command": "npx",
"args": [
"-y",
"mcp-advisor@latest"
],
"env": {
"DEFAULT_SPEC_VERSION":"2025-06-18" // Optional - overrides the default version used for static Resources that correspond to a specific MCP version
}
}
}
}
- The server provides the following capabilities:
Prompts
explain: Comprehensive explanation of MCP topics with full documentation context.- Required argument:
topic- Specifies which MCP topic you would like explained in detail. - Optional argument:
version- Specifies which MCP specification version to use. Supported versions:draft,2024-11-05,2025-03-26,2025-06-18(default).
- Required argument:
evaluate_server_compliance: Evaluates Model Context Protocol (MCP) specification compliance for a given server repository.- Required argument:
path- Specifies the path to the MCP server repository to evaluate. - Optional argument:
version- Specifies which MCP specification version to use. Supported versions:draft,2024-11-05,2025-03-26,2025-06-18(default).
- Required argument:
Resources
The server provides access to different sections of the MCP specification and documentation:
Specification Resources
- Complete Specification: The complete Model Context Protocol specification including schema, architecture, base protocol, utilities, server features, and client features
- Schema Specification: The complete Model Context Protocol JSON schema specification
- Architecture Overview: Overview of the Model Context Protocol architecture
- Base Protocol: Core protocol details including transports, authorization, and lifecycle
- Utilities: Documentation for Ping, Cancellation, and Progress Reporting features
- Server Features: Comprehensive guide to Prompts, Resources, Tools, and Server Utilities including completion, logging, and pagination
- Client Features: Information about Roots and Sampling capabilities
All specification resources can be accessed with a specific version parameter.
Resource Templates
The server provides resource templates that allow accessing specification resources for different versions:
https://modelcontextprotocol.io/specification/{version}/index.md: Access the complete specification for any supported versionhttps://modelcontextprotocol.io/specification/{version}/schema.json: Access the JSON schema for any supported versionhttps://modelcontextprotocol.io/specification/{version}/architecture/index.md: Access the architecture specification for any supported versionhttps://modelcontextprotocol.io/specification/{version}/basic/index.md: Access the base protocol specification for any supported versionhttps://modelcontextprotocol.io/specification/{version}/basic/utilities/index.md: Access the utilities specification for any supported versionhttps://modelcontextprotocol.io/specification/{version}/server/index.md: Access the server features specification for any supported versionhttps://modelcontextprotocol.io/specification/{version}/client/index.md: Access the client features specification for any supported version
Supported versions: draft, 2024-11-05, 2025-03-26, 2025-06-18 (default)
Version Configuration:
- Resource Templates: Clients that support Resource Templates can specify the version in the URI template.
- Environment Variable: Set the
DEFAULT_SPEC_VERSIONenvironment variable to change the default version (e.g.,DEFAULT_SPEC_VERSION=draft). - Default Version: If neither of the above is specified, the server uses
2025-06-18as the default version.
Note on Backward Compatibility: Clients that only support Resources (and not Resource Templates) will still be able to access the regular Resources using the configured default version. The server maintains full backward compatibility with existing clients.
Additional Documentation Resources
- Getting Started (
/quickstart): Getting started guides for client developers, server developers, and users - Development (
/development): Development resources including contributing guidelines, roadmap, and updates - SDK Documentation (
/sdk): SDK documentation for various programming languages - Tutorials & Examples (
/tutorials): Tutorials, examples, and implementation guides - General Documentation (
/docs): General documentation including FAQs, introduction, and client list - Community Documentation (
/community): Community guidelines including SEP Guidelines, Communication, and Governance - Getting Started Guide (
/docs/getting-started): Introduction and getting started with MCP - Learning Resources (
/docs/learn): Architecture overview and core concepts - MCP Debugging Tools (
/legacy/tools): Debugging tools including the MCP Inspector - MCP Overview (
/overview): High-level overview of the Model Context Protocol
All specification content is fetched from a list provided by a standardized llms.txt file (except for the schema.json which is fetched from GitHub) and cached locally with a 1-hour TTL (time-to-live) to improve performance. If a fetch fails, the server will attempt to use expired cached content as a fallback when available.
Development
# Clone the repository
git clone https://github.com/olaservo/mcp-advisor.git
cd mcp-advisor
# Install dependencies
npm install
# Run in development mode
npm run dev
# Build the project
npm run build
# Start the built server
npm start
Testing URL Filtering and Version Support
The server includes URL filtering to ensure content matching the requested version is included:
npm run test
This verifies that the server correctly filters specification URLs based on the requested version. The server supports multiple versions including draft, 2024-11-05, 2025-03-26, and 2025-06-18, with 2025-06-18 being the default if no version is specified.
Links
- Model Context Protocol Specification (supports multiple versions)
相關伺服器
Alpha Vantage MCP Server
贊助Access financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Jira Context MCP
MCP server to provide Jira Tickets information to AI coding agents like Cursor.
MLflow MCP
MLflow MCP server for ML experiment tracking with advanced querying, run comparison, artifact access, and model registry.
GhidraMCP
An embedded MCP server for Ghidra, exposing program data and reverse engineering functionalities.
Build-Scout
Interact with various build systems including Gradle, Maven, NPM/Yarn, Cargo, Python, Makefile, and CMake.
Claude Swarm MCP Server
An MCP server for multi-agent orchestration using Claude AI via Claude Desktop.
Codebase MCP Server
An intelligent codebase search engine that transforms local codebases into a natural language queryable knowledge base.
Cygnus MCP Server
An MCP server demonstrating Cygnus tools for reading text files and invoking local APIs.
Floyd
Scheduling and booking engine for AI agents. Check availability, hold slots, and confirm appointments with two-phase booking and conflict-free resource management.
Manim MCP Server
Executes Manim Python scripts to generate and return rendered video animations.
TokRepo MCP Server
Search, install, and share AI skills, prompts, and MCP configs from a curated registry of 200+ assets via CLI or MCP protocol.