UnrealSAFE
pure python Unreal Engine MCP server
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
This repository contains a Model Context Protocol (MCP) Python server that allows Claude Desktop to interact with Unreal Engine 5.3 (via Remote Control API), creating and manipulating 3D objects based on text prompts. This integration enables Claude to build and modify 3D scenes in Unreal Engine through natural language, representing an early step toward text-to-game-generation technology.
Current Features:
- use Claude Desktop text prompts to arrange assets in Unreal Engine Editor
- create static meshes for assembling primitive shapes
- look up Unreal project folder for assets
- list actors in the current level and inspect the current editor selection
- save the current level from MCP
- call arbitrary Unreal Remote Control functions
- read and write Unreal object properties through the Remote Control property endpoint
- direct CLI with per-command
--helpfor non-MCP agents - shared castle workflow, sample agent skill, and TLA+ use-case model
- castle reset workflow and first-class castle MCP tools
Quick Start
1. Requirements
- Python 3.10+
- Unreal Engine 5.3 with Remote Control API (plugin) enabled
2. Installation
Clone the repository and install in editable mode:
git clone https://github.com/runeape-sats/unreal-mcp.git cd unreal-mcp pip install -e .
This installs the unreal-mcp-cli command and makes the unreal_mcp package importable.
3. Configure MCP server
- if you are using Claude Desktop → File → Settings → Developer → Edit Config
claude_desktop_config.jsonand add the following, adjusting the path to your local repository:
{
"mcpServers": {
"unreal-mcp": {
"command": "uv",
"args": ["--directory", "\\path\\to\\unreal-mcp", "run", "unreal_mcp_server.py"],
"ffd828478182OBSERVED · 2026-10-08Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.
claude mcp add unreal-mcp-dev -- uvx unreal-mcp-dev
{
"mcpServers": {
"unreal-mcp-dev": {
"command": "uvx",
"args": [
"unreal-mcp-dev"
]
}
}
}Exposed tools (21)
12 read · 6 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
create_basic_castle | write | |
create_static_mesh_actor | write | |
delete_actor | destructive | |
get_actor_info | read | |
get_level_info | write | Get information about the current Unreal Engine level and update spatial context. |
get_object_property | read | |
get_selected_actors | read | Get the actors currently selected in the Unreal Editor. |
get_spatial_context | read | Return the current spatial context of all actors as a JSON string. |
list_available_assets | read | |
list_level_actors | read | |
modify_actor | write | |
remote_call | read | |
reset_basic_castle | destructive | |
reset_spatial_context | destructive | Reset the spatial context, clearing all tracked actors. |
save_current_level | write | Save the currently open Unreal level. |
search_assets_recursively | read | |
select_actors | read | |
set_object_property | write | |
spawn_actor_from_blueprint | read | |
spawn_static_mesh | read | |
verify_basic_castle | read |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- declared (2 observation(s))
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (1)
delete_actor, reset_basic_castle, reset_spatial_context
Gates applied: no_behavioural_pass.
ffd828478182full audit observations/trust-audit/mcp-server/runeape-sats__unreal-2.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-08 | ffd828478182 | SAFE | B | 89 | first audit |
Questions
What is the Unreal MCP server?
pure python Unreal Engine MCP server
What tools does Unreal expose?
21 in total: 12 read-only, 6 that write, and 3 that can delete or overwrite (delete_actor, reset_basic_castle, reset_spatial_context). Every one is listed on this page with its risk.
Is Unreal safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 3 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Unreal need?
No credential environment variables were found in its source, so it appears to need none.
How current is this page?
The grade is for one exact copy of the source (ffd828478182), read on 2026-10-08. The repository is watched and re-audited when it changes.