MCP Server - A Node In Layers Package for building MCP Servers
@Node-In-Layers
MCP Server - A Node In Layers Package for building MCP Servers について
概要はまだありません
基本情報
設定
ツール
ツールは検出されませんでした
ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。
概要
What is MCP Server - A Node In Layers Package for building MCP Servers?
A Node In Layers package for building MCP servers that exposes domains, features, and model CRUDs as MCP tools so an AI can discover and call them. It is self-describing, automatically organizes tools into an optimized system, and includes built-in prompting and examples to improve AI tool selection and formatting accuracy.
How to use MCP Server - A Node In Layers Package for building MCP Servers?
Install with npm install @node-in-layers/mcp-server. Configure in /config.base.mts by adding the MCP server module to the apps list, including the mcp layer in the layer order, and setting server connection details. Run the server via a script that calls system.mcp[McpNamespace].start(context).
Key features of MCP Server - A Node In Layers Package for building MCP Servers
- Automatically exposes features and model CRUDs as MCP tools.
- Self-describing system with a configurable START_HERE navigation tool.
- Domain-organized tool surface with
list_domains,list_features,describe_feature,execute_feature, and model CRUD tools. - Flat mode to expose feature functions as individual tools without navigation tools.
- Component hiding via
hideComponentsto restrict visibility of domains, features, or models. - Custom tool extension using
addTool()andaddAnnotatedFunction().
Use cases of MCP Server - A Node In Layers Package for building MCP Servers
- Build an MCP server that exposes a business domain’s features and data models to an AI assistant.
- Create a self-documenting API surface where the AI can discover and execute operations without hardcoded tool names.
- Reduce tool selection errors in systems with hundreds of tools through automatic organization and prompting.
- Expose a minimal tool surface (flat mode) for systems with a single function or small tool set.
- Hide sensitive or experimental domains/features while keeping the rest of the system available.
FAQ from MCP Server - A Node In Layers Package for building MCP Servers
What is the difference between normal mode and flat mode?
In normal mode the server exposes domain-, feature-, and model-level navigation tools (START_HERE, list_domains, etc.). In flat mode each feature function becomes its own MCP tool named domain_featureName, and navigation tools are not exposed.
How do I hide certain domains, features, or model CRUDs?
Configure hideComponents in the MCP namespace with domains (hide entire domains), paths (dot-separated paths to specific features or models), or allModels (hide all model CRUD tools).
What is the START_HERE tool and how is it configured?
START_HERE is the first tool an AI should call to learn how to navigate the system. Its response includes system metadata, default system entries, and optionally pre‑fetched domain/feature lists. You can also add custom examplesOfUse to document higher‑level flows.
How do I add custom tools beyond the auto‑exposed ones?
Use the addTool() method on the MCP layer (accessible via mcp[McpNamespace] inside a layer that runs after the MCP layer). Provide a name, description, input/output schema, and an execute function.
What are the runtime requirements?
The package requires Node.js, npm, TypeScript, and the @node-in-layers/core library. It is used within a Node In Layers system configuration.
「その他」の他のコンテンツ
Awesome Mcp Servers
punkpeyeA collection of MCP servers.
Awesome-MCP-ZH
yzflyMCP 资源精选, MCP指南,Claude MCP,MCP Servers, MCP Clients
AutoBrowser MCP
autobrowser-aiBrowser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser
FastMCP v2 🚀
jlowin🚀 The fast, Pythonic way to build MCP servers and clients.
MCP Go 🚀
mark3labsA Go implementation of the Model Context Protocol (MCP), enabling seamless integration between LLM applications and external data sources and tools.
コメント