unreal-engine-skills-for-claude-code-plugin README

The project's README from github.com/EpicGames/unreal-engine-skills-for-claude-code-plugin. As one markdown file: README.md.

Unreal Engine Skills for Claude Code

Control Unreal Editor directly from Claude Code via MCP. Hundreds of tools exposed via Unreal's ToolsetRegistry across 30+ toolsets: actors, blueprints, materials, Niagara, Control Rigs, Sequencer, State Trees, widgets, Gameplay Ability System, automation testing, and more.

Contents

Skills

Hooks

Prerequisites

  1. Unreal Editor with the ModelContextProtocol and AllToolsets plugins enabled (AllToolsets provides the tools; the server exposes none without it)
  2. Editor running to execute tools, with the MCP server started. Run ModelContextProtocol.StartServer, or enable bAutoStartServer through skills/unreal-mcp/references/setup.md. The optional proxy can keep the client connection open while the editor is stopped.
  3. A bash shell on PATH - required for the SessionStart hook (see Platform Support)

Platform Support

Supported: macOS, Linux, and Windows via Git Bash or WSL.

The SessionStart hook is a bash script (hooks/unreal-context.sh) invoked by hooks/hooks.json as bash ${CLAUDE_PLUGIN_ROOT}/hooks/unreal-context.sh. It requires a working bash on PATH:

Installation

This plugin is published in Anthropic's official Claude Code plugin marketplace, claude-plugins-official, which is available automatically when you start Claude Code. Browse it at claude.com/plugins or in the Discover tab of the /plugin panel.

Install for a single developer

In Claude Code:

/plugin install unreal-engine-skills-for-claude-code@claude-plugins-official

If the official marketplace is not present, add it first, then run the install command above:

/plugin marketplace add anthropics/claude-plugins-official

Install for a team

Commit this to .claude/settings.json in the project that should use the plugin. Anyone who trusts the project folder is prompted to install it automatically:

{
  "enabledPlugins": {
    "unreal-engine-skills-for-claude-code@claude-plugins-official": true
  }
}

Verification

  1. Launch Unreal Editor, then run ModelContextProtocol.StartServer in the console to start the MCP server.
  2. Check the Output Log for MCP server startup messages.
  3. In Claude Code, run /plugin. The Installed tab should list unreal-engine-skills-for-claude-code as enabled. This confirms the plugin itself (skills, hooks) is loaded.
  4. Run /mcp. Check for unreal-mcp, or unreal-mcp-proxy if you installed the proxy. A connected proxy does not prove that Unreal is reachable.
  5. Try: "List all actors in the current level".

Configuration

The default port is 8000 with URL path /mcp. If the port is in use, run ModelContextProtocol.StartServer <port> in the console with a different port number.

Note: This plugin does not ship a static .mcp.json file. Run ModelContextProtocol.GenerateClientConfig ClaudeCode in the editor console to generate it from the current server port and URL; re-run after changing either.

Optional proxy for editor recovery

If your engine build includes Engine/Plugins/Experimental/ModelContextProtocol/Extras/Proxy, you can use its unreal_mcp_proxy executable.

The proxy keeps the client-facing MCP session open while Unreal is unavailable and reconnects when Unreal starts again. It stores Unreal's native tool catalog as a temporary fallback. Cached tools do not prove that live calls can succeed. Without a cache, it exposes only unreal_mcp_status until the client fetches the recovered catalog.

See proxy setup for platform binaries and installation commands. See proxy recovery for status, cached catalogs, and connection failures.

Tool search is enabled by default: the MCP server exposes three meta-tools (list_toolsets, describe_toolset, call_tool) instead of the full tool catalog, so Claude discovers toolsets on demand, the prompt cache stays warm, and discovered tools are callable on the same turn through call_tool. Toggle with the bEnableToolSearch setting in [/Script/ModelContextProtocolEngine.ModelContextProtocolSettings]; when disabled, every tool is registered upfront (path used by the hash-mapping commandlet). The model-facing usage contract lives in skills/unreal-mcp/SKILL.md.

Security

Installing this plugin gives Claude broad, live access to the running Unreal Editor. Treat that access the same way you would treat running arbitrary code from an assistant, because in practice it is.

Localhost is not a trust boundary. The MCP server binds to localhost:8000 with origin validation. Origin validation protects against a browser tab talking to the server, but any process running as the same user on the same machine can connect. Do not run the MCP server on shared or untrusted machines, and do not expose the port outside the loopback interface.

ProgrammaticToolset.execute_tool_script executes arbitrary Python inside the editor process. That script has full access to every toolset API, the project on disk, the asset database, and editor-privileged functions. Treat every invocation as a privileged operation that can mutate, move, or delete project content, and expect it to succeed without a second confirmation when approvals are disabled.

--dangerously-skip-permissions removes the per-tool approval gate. In that mode Claude can drive the editor and run Python through execute_tool_script with no further consent. A bad prompt in that mode can reach into the project and modify a large amount of state before you notice. Prefer not to run Claude Code with --dangerously-skip-permissions while this plugin is loaded; if you do, keep the blast radius small (narrow prompts, read-heavy tasks, a throwaway sandbox project).

Source-control hygiene. MCP tools edit live UObject state and can mutate, move, or delete VCS-tracked assets in a single call. Save and commit (or shelve) before any long MCP-driven session so the working copy is recoverable if Claude produces an unexpected result. Review the diff before submitting.

What's Available

All tools are auto-discovered by Claude Code via MCP. No manual configuration needed. Tools cover the full editor surface across these domains:

unreal-engine-skills-for-claude-code-plugin (EpicGames/unreal-engine-skills-for-claude-code-plugin)

Searching Database…

552 ·GitHub
Unreal-engine-skills-for-claude-code-plugin
What is this?This is a Claude Code plugin that lets Claude control the Unreal Engine Editor directly through the Model Context Protocol (MCP), exposing hundreds of editor tools across 30+ toolsets such as actors, blueprints, materials, Niagara VFX, Sequencer, and the Gameplay Ability System.
Is this a standalone app or a plugin for Claude Code?It is a Claude Code plugin containing a 'unreal-mcp' skill and a SessionStart hook; the actual editor control happens through Unreal's own ModelContextProtocol and AllToolsets editor plugins, which this Claude Code plugin connects to.
How do I install and start using it?Install it from Anthropic's official claude-plugins-official marketplace (via the /plugin panel for a single developer, or by committing it to .claude/settings.json for a team), then in Unreal Editor enable the ModelContextProtocol and AllToolsets plugins and run ModelContextProtocol.StartServer.
What do I need running in Unreal before this works?You need Unreal Editor open with the ModelContextProtocol and AllToolsets plugins enabled, the MCP server started via ModelContextProtocol.StartServer, and a bash shell on PATH (Git Bash or WSL on Windows) for the SessionStart hook to run.
Is it safe to let Claude drive the editor like this?Not without care: the MCP server only binds to localhost but any local process can still connect, and ProgrammaticToolset.executetoolscript runs arbitrary Python inside the editor with full project access, so a bad prompt (especially under --dangerously-skip-permissions) can mutate or delete assets before you notice.
Does the SessionStart hook work on Windows too?It works, but native PowerShell alone is not enough — Windows needs Git for Windows (Git Bash) or WSL for the bash-based hook; macOS and Linux work out of the box.
What can the exposed tools actually do in the editor?They cover the full editor surface: spawning/editing actors, blueprint graphs, materials, meshes/textures, Control Rig and State Tree animation, Sequencer, Niagara VFX, UMG widgets, gameplay tags and GAS, automation testing, and batched Python scripting.
What is the optional proxy for?It keeps the client-facing MCP session open and caches Unreal's tool catalog so Claude Code's connection survives even when the editor is temporarily stopped or restarting.
Does the plugin expose every single tool at once?No, by default 'tool search' is on: only three meta-tools (listtoolsets, describetoolset, calltool) are exposed and Claude discovers specific toolsets on demand, keeping the prompt cache warm; this can be disabled via the bEnableToolSearch setting.
What does the SessionStart hook actually change?It injects a short note at session start telling Claude the repo is an Unreal Engine project, so it defaults to UE conventions (C++/UObject, Slate, UHT) and prefers the unreal-mcp skill.
What are similar projects?everything-claude-code, coral, evenfire, open-steps, skills, copperhead, open-seo-mcp-skills, devspace
Which topics is it in?game-engine, claude-code-plugin, game-development, model-context-protocol, shell, agentic-ai, mcp-server, plugin, claude, ai-agent, agent, claude-code
Creator
EpicGames
Size
0.1 MB
Stars
552
Updated
CommitSep 17
Age
4mo
Commits · 90 Days