Potencia tu agente
El comando propio del publicador. No es un endpoint de PlugMCP.
Host a copy with our keys and logs.
npx -y @model-graph-tools/[email protected]Claude Code no tiene enlace de instalación para un comando local. Copia el comando de abajo y ejecútalo en una terminal.
claude mcp add model-graph-tools-mcp-server -- npx -y @model-graph-tools/mcp-server@0.12.0Ejecútalo en una terminal.
{ "mcpServers": { "model-graph-tools-mcp-server": { "command": "npx", "args": [ "-y", "@model-graph-tools/[email protected]" ] } }}claude_desktop_config.json
¿URL remota o un comando en tu máquina? En qué se diferencian las dos instalaciones
Ejecutar en PlugMCP
A new server you create. It is not this listing's official endpoint. Use the publisher install when you want their URL and their credentials.
AbrirDel publicador
Explore the WildFly management model via natural language — from any AI agent that supports MCP and Agent Skills.
WildFly exposes its entire configuration and runtime state through a management model — a tree of resources, each with attributes, operations, and capabilities. This project lets you search, browse, and compare that model across WildFly versions and feature packs, powered by Neo4j graph databases and the CLI.
MCP server for querying the WildFly management model graph
The mgt CLI is bundled automatically with the MCP server via platform-specific npm packages — no separate installation needed.
claude plugin marketplace add https://github.com/model-graph-tools/claude-plugin
claude plugin install mgt@model-graph-tools
The MCP server is fetched automatically from npm via npx — no build step required.
git clone https://github.com/model-graph-tools/claude-plugin.git
cd claude-plugin/mcp-server
npm install
npm run build
Then load the plugin locally:
claude --plugin-dir /path/to/claude-plugin
The MCP server and skill follow open standards (MCP and Agent Skills) and work with any compatible AI agent. Setup has two parts:
skills/wildfly-model-graph/ directory from this repository into your agent's skills folderFor most agents, the universal skills directory is .agents/skills/ (project-level) or ~/.agents/skills/ (user-level). See the Agent Skills client list for all supported agents.
Add to ~/.gemini/settings.json (or project-level .gemini/settings.json):
{
"mcpServers": {
"wildfly-model-graph": {
"command": "npx",
"args": ["--yes", "@model-graph-tools/mcp-server"]
}
}
}
Install the skill:
# User-level (available in all projects)
cp -r skills/wildfly-model-graph ~/.agents/skills/
# Or project-level
cp -r skills/wildfly-model-graph .agents/skills/
Add to ~/.codex/config.toml (or project-level .codex/config.toml):
[mcp_servers.wildfly-model-graph]
command = "npx"
args = ["--yes", "@model-graph-tools/mcp-server"]
Install the skill:
cp -r skills/wildfly-model-graph ~/.agents/skills/
Add to .vscode/mcp.json in your workspace:
{
"servers": {
"wildfly-model-graph": {
"command": "npx",
"args": ["--yes", "@model-graph-tools/mcp-server"]
}
}
}
Install the skill:
# Project-level
cp -r skills/wildfly-model-graph .agents/skills/
# Or user-level
cp -r skills/wildfly-model-graph ~/.agents/skills/
Add to ~/.cursor/mcp.json (or project-level .cursor/mcp.json):
{
"mcpServers": {
"wildfly-model-graph": {
"command": "npx",
"args": ["--yes", "@model-graph-tools/mcp-server"]
}
}
}
Install the skill:
cp -r skills/wildfly-model-graph ~/.agents/skills/
Add to .bob/mcp.json in your project:
{
"mcpServers": {
"wildfly-model-graph": {
"command": "npx",
"args": ["--yes", "@model-graph-tools/mcp-server"]
}
}
}
Install the skill:
# Project-level
cp -r skills/wildfly-model-graph .agents/skills/
# Or user-level
cp -r skills/wildfly-model-graph ~/.agents/skills/
For any agent that supports MCP over stdio, configure it to run:
npx --yes @model-graph-tools/mcp-server
Copy the skill directory to your agent's skills folder (typically .agents/skills/ or ~/.agents/skills/). See the Agent Skills specification for details.
[!NOTE] Slash commands are a Claude Code feature. Other agents use the skill (which activates automatically based on context) and MCP tools directly — see Natural Language below.
Use /mgt:model followed by your question:
Model overview
/mgt:model Give me an overview of the WildFly 39 management model.
/mgt:model How many resources and attributes does WildFly 39 have?
Browse and search
/mgt:model How do I add a new datasource?
/mgt:model How does the logging subsystem work?
/mgt:model How is TLS configured?
/mgt:model What is an authentication context?
/mgt:model Tell me more about credential stores.
Stability levels
/mgt:model Are there any community features in WildFly 39?
/mgt:model Show me all preview resources.
/mgt:model What community stability attributes exist?
Capabilities and deprecation
/mgt:model How do capabilities work?
/mgt:model What capabilities does elytron provide?
/mgt:model What operations have been deprecated recently?
Compare versions
/mgt:model Compare WildFly 38 and 39.
/mgt:model What attributes changed between WildFly 38 and 39?
/mgt:model Compare the IO subsystem between WildFly 23 and the latest version.
Feature packs
/mgt:model What resources does the AI feature pack provide?
/mgt:model Compare the first and the latest version of the AI feature pack.
The skill also activates automatically when the agent detects WildFly management model context in your conversation — just ask about subsystems, resources, attributes, or operations naturally. So most of the questions above should also work without using the slash command.
Each WildFly version or feature pack runs as a separate Neo4j container pulled from
quay.io/modelgraphtools/model. The MCP server manages these through mgt:
mgt start) persist and are not affected/mgt:model stop WildFly 39This project is part of the model-graph-tools ecosystem:
Contributions are welcome! Here's how to get started:
git checkout -b feat/my-featurecd mcp-server && npm install && npm run buildclaude --plugin-dir /path/to/your/forkfeat:, fix:, docs:, etc.git clone https://github.com/model-graph-tools/claude-plugin.git
cd claude-plugin/mcp-server
npm install
npm run dev # runs with tsx, no compile step needed
claude-plugin/
├── .claude-plugin/
│ ├── plugin.json # Plugin manifest (skills, mcpServers)
│ └── marketplace.json # Marketplace manifest
├── skills/
│ └── wildfly-model-graph/
│ ├── SKILL.md # Skill prompt with domain knowledge
│ └── references/
│ ├── cypher-queries.md # Cypher query reference
│ └── graph-schema.md # Neo4j graph schema reference
├── commands/
│ └── model.md # /mgt:model slash command
├── mcp-server/
│ ├── src/
│ │ ├── index.ts # Server entry point, tool registration
│ │ ├── neo4j.ts # Neo4j connection management
│ │ ├── mgt.ts # mgt CLI wrapper
│ │ ├── session.ts # Session state management
│ │ └── tools/ # One file per MCP tool (20 tools)
│ ├── build.mjs # Custom build script
│ ├── package.json
│ └── tsconfig.json
├── CLAUDE.md # Claude Code project instructions
└── README.md
mcp-server/src/tools/mcp-server/src/index.ts with a Zod input schemaskills/wildfly-model-graph/SKILL.md with usage guidanceMCP server for querying the WildFly management model graph
{ "mcpServers": { "model-graph-tools-mcp-server": { "command": "npx", "args": [ "-y", "@model-graph-tools/[email protected]" ] } }}Herramientas
Aún no hay lista de herramientas. Los servidores remotos la reciben tras la primera comprobación de estado.
Preguntas frecuentes
Copia en esta página la configuración stdio para Cursor, Claude Desktop, Claude Code o VS Code. Usa la instalación del publicador, no un proxy de PlugMCP.
Este registro no indica ninguna variable de entorno obligatoria.
@model Graph Tools/ se ejecuta localmente por stdio.
Última comprobación: 2026-10-28. MCP server for querying the WildFly management model graph