devskillsit/skills-mcp-glpi

开发工具

本页由 AI 翻译，技术细节请以原文为准。查看原文

7 stars 7 forks 质量 81 趋势 29MIT

Repository Install

MCP Server para integração com GLPI - 66 ferramentas para gestão de tickets, assets e usuários | Skills IT

概览

is a production-ready Model Context Protocol (MCP) server for GLPI (IT Service Management) with a revolutionary token-optimization architecture. Built by , this MCP consolidates 68 specialized tools into 17 active enterprise-grade tools (18 defined; the AI analysis tool is off by default) with Markdown responses, tool annotations, and MCP resources. Compatible with (legacy API v1) and (current stable). - — problems, changes, projects, contracts and suppliers are now first-class, through glpi_search_itil_records and glpi_manage_itil_records. - — glpi_search_records_by_criteria queries any itemtype, counts without paginating, and discovers which fields your instance actually exposes. - — assigned technician, assigned group, requester and category, each accepting a instead of forcing an id lookup first; plus urgency as its own axis, open_only, and caller-controlled sorting. - — unified timeline, tasks, approvals, group assignment, ticket linking and file attachments.

README

以下为原文 README 节选，完整内容请查看原文。

📖 About The Project

Skills MCP GLPI v2.2 is a production-ready Model Context Protocol (MCP) server for GLPI (IT Service Management) with a revolutionary token-optimization architecture. Built by Skills IT, this MCP consolidates 68 specialized tools into 17 active enterprise-grade tools (18 defined; the AI analysis tool is off by default) with Markdown responses, tool annotations, and MCP resources.

Compatible with GLPI 10.x (legacy API v1) and GLPI 11.x (current stable).

What’s new in 2.2 (August 2026):

ITIL beyond the ticket — problems, changes, projects, contracts and suppliers are now first-class, through glpi_search_itil_records and glpi_manage_itil_records.
Free-criteria search — glpi_search_records_by_criteria queries any itemtype, counts without paginating, and discovers which fields your instance actually exposes.
Filters people actually ask for — assigned technician, assigned group, requester and category, each accepting a name instead of forcing an id lookup first; plus urgency as its own axis, open_only, and caller-controlled sorting.
9 new ITIL actions on tickets — unified timeline, tasks, approvals, group assignment, ticket linking and file attachments.
Correctness fixes for answers that looked right and were not — text search silently discarding the other filters, entity filters that never applied, and pagination reporting totals it never had. See CHANGELOG.md.

🌟 Why v2.0?

The v1.0 approach (68 tools, raw JSON responses) caused token explosion — typical GLPI queries consumed 50-100K tokens with limited reusability. v2.0 fixes this with a consolidated, Markdown-first design:

Aspect	v1.0	v2.2
Tools	68 fragmented	17 consolidated ✨

在 GitHub 查看完整 README

安装

This server does not publish a one-line install command.

Open the repository installation guide 新手？先看 mcpskill.org 教程 →

配置

{
  "mcpServers": {
    "glpi": {
      "type": "streamable-http",
      "url": "http://localhost:8824/mcp",
      "headers": {
        "X-GLPI-User-Token": "your_user_token_here"
      }
    }
  }
}