Skip to main content
Glama
Sign In
EKMS MCP Server

EKMS MCP Server

by jihsin
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Remote

EKMS MCP Server

Enterprise Knowledge Management System 的 Model Context Protocol (MCP) 伺服器。

讓 Claude Code 能夠直接查詢和操作企業知識庫。

功能

  • search_knowledge - 搜尋知識庫（支援混合/語意/關鍵字搜尋）

  • recommend_knowledge - 根據上下文和客戶特徵推薦知識

  • get_knowledge_item - 取得特定知識項目完整內容

  • list_knowledge_types - 列出所有知識類型

  • list_customers - 列出所有客戶

  • get_knowledge_graph - 取得知識關聯圖譜

  • record_knowledge_feedback - 記錄知識使用反饋

Related MCP server: RAG MCP Server

安裝

cd ekms-mcp-server
npm install

配置

複製環境變數範本：

cp .env.example .env

編輯 .env 設定 EKMS API 位址：

EKMS_API_URL=http://your-ekms-api-server:3000
EKMS_API_KEY=your-api-key

使用方式

方式一：本地開發

  1. 啟動伺服器測試：

npm start

  1. 在 Claude Code 配置中加入：

{
  "mcpServers": {
    "ekms": {
      "command": "node",
      "args": ["/path/to/ekms-mcp-server/src/index.js"],
      "env": {
        "EKMS_API_URL": "http://localhost:3000"
      }
    }
  }
}

方式二：使用 Claude CLI

claude mcp add ekms node /path/to/ekms-mcp-server/src/index.js

開發

# 開發模式（自動重載）
npm run dev

# 執行測試
npm test

架構

ekms-mcp-server/
├── src/
│   ├── index.js              # MCP 伺服器入口
│   ├── tools/
│   │   └── definitions.js    # 工具定義
│   ├── handlers/
│   │   └── knowledge-handlers.js  # 工具處理邏輯
│   └── utils/
│       └── api-client.js     # EKMS API 客戶端
├── config/
├── migrations/               # 資料庫遷移腳本
└── package.json

License

MIT

Available Tools

7 tools
get_knowledge_graphC

取得特定知識項目的關聯圖譜，顯示與其他知識的關係

Inputs: knowledgeId (required), depth, relationTypes.

get_knowledge_itemB

根據 ID 取得特定知識項目的完整內容，包含所有元資料和關聯知識

Inputs: id (required), includeRelations.

list_customersB

列出知識庫中所有客戶名稱，用於過濾特定客戶的知識

Inputs: includeStats.

list_knowledge_typesB

列出所有可用的知識類型及其說明，用於了解知識庫結構

Inputs: no named parameters.

recommend_knowledgeA

根據對話上下文和客戶特徵推薦相關知識。 可根據 Big Five 人格特質和健康分數進行個人化推薦：

  • openness: 開放性（高=喜歡延伸資訊，低=只要核心答案）

  • conscientiousness: 盡責性（高=詳細步驟，低=簡潔重點）

  • extraversion: 外向性

  • agreeableness: 親和性

  • neuroticism: 神經質（高=需要更多安撫） 健康分數 0-100，低分客戶需要更多關懷

Inputs: context (required), customerName, personalityScores, healthScore, +1 more.

record_knowledge_feedbackB

記錄知識使用反饋，用於持續優化知識品質

Inputs: knowledgeId (required), feedbackType (required), sessionId, context, +2 more.

search_knowledgeA

搜尋企業知識庫。支援三種搜尋模式：

  • hybrid: 混合搜尋（關鍵字 + 語意），最佳平衡

  • semantic: 純語意搜尋，理解意圖但可能漏掉精確匹配

  • keyword: 純關鍵字搜尋，精確但可能漏掉同義詞

知識類型包含：product_info（產品資訊）、faq（常見問題）、policy（政策規定）、 announcement（公告）、tutorial（教學）、troubleshooting（故障排除）、 best_practice（最佳實踐）、case_study（案例研究）、glossary（術語）、other（其他）

Inputs: query (required), searchType, knowledgeTypes, customerName, +2 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A3.7/5.0

    Scored across 7 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose with no overlap: get_knowledge_graph retrieves relationship maps, get_knowledge_item fetches full content, list_customers and list_knowledge_types provide filtering metadata, recommend_knowledge offers personalized suggestions, record_knowledge_feedback captures usage data, and search_knowledge performs queries. The descriptions clearly differentiate their functions, preventing agent misselection.

    Naming Consistency5/5

    All tool names follow a consistent verb_noun pattern in snake_case (e.g., get_knowledge_graph, list_customers, search_knowledge). There are no deviations in naming conventions, making the set predictable and readable for agents.

    Tool Count5/5

    With 7 tools, the count is well-scoped for an enterprise knowledge management system (EKMS). Each tool serves a specific role in knowledge retrieval, recommendation, feedback, and metadata handling, with no redundant or missing tools for the domain.

    Completeness4/5

    The tool surface covers core EKMS workflows: retrieval (get, search), metadata (list), recommendation, and feedback. Minor gaps exist, such as no explicit create, update, or delete tools for knowledge items, but agents can work around this by using existing tools for read-only operations and feedback for updates.

    Resources

    Related MCP Connectors

    Related MCP Servers

    • RAG MCP Server

      torkian
      A
      license
      Not graded
      quality
      D
      maintenance
      Enables Claude Desktop to search custom knowledge bases using retrieval-augmented generation via a simple MCP tool.
      MIT

    • Cognify MCP Server

      S3YED
      A
      license
      Not graded
      quality
      B
      maintenance
      Enables document ingestion and typed knowledge graph queries through Claude MCP tools, allowing agents to extract, store, and retrieve typed entities and relations from documents.
      2
      MIT