UE5 BridgeSAFE
MCP server bridging AI assistants to Unreal Engine 5 editor
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
[](tests/) [](vitest.config.js) [](package.json) [](LICENSE)
An MCP (Model Context Protocol) server that bridges AI assistants to Unreal Engine 5's editor, enabling direct manipulation of levels, actors, Blueprints, and Animation Blueprints.
Why a Standalone Repository?
This MCP server was originally part of the UnrealClaude plugin, but has been moved to its own repository because:
- Expanded Scope: The MCP protocol is an open standard now supported by many AI providers
- Easier Forking: Developers can fork and adapt this bridge for other AI services without pulling the entire Unreal plugin
- Community Contributions: Standalone repo makes it easier to accept PRs for new features, bug fixes, and support for additional MCP clients
Compatible AI Clients
This server uses the standard Model Context Protocol and works with any MCP-compatible AI client, including:
- Claude Code (CLI)
- Claude Desktop App
- OpenAI ChatGPT Desktop (MCP support added March 2025)
- Cursor (AI-powered IDE)
- Replit (AI coding platform)
- Sourcegraph Cody
- Windsurf
- And any other MCP-compliant tool
The protocol was donated to the Linux Foundation in December 2025 and continues to see broad adoption across the AI industry.
Important: What This Tool Is (and Isn't)
This is NOT a "text-to-game" tool. It will not generate 3D models, textures, animations, or replace the work of artists, animators, and asset creators.
This IS a workflow accelerator for Unreal Engine developers. It helps with:
- **Project setup monoto
f75772f7c397OBSERVED · 2026-10-07Connect
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 ue5-mcp-server -- npx -y [email protected]
{
"mcpServers": {
"ue5-mcp-server": {
"command": "npx",
"args": [
"-y",
"[email protected]"
]
}
}
}Exposed tools (14)
11 read · 3 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
bad_tool | read | |
blueprint_compile | read | Compile a Blueprint asset |
blueprint_modify | write | Modify a Blueprint asset |
character | read | Character blueprint and movement operations |
enhanced_input | read | Enhanced Input action and mapping context operations |
execute_script | write | Execute a Python script in Unreal |
get_actors | read | List actors in the current level |
material | read | Material instance and parameter operations |
minimal_tool | read | A minimal tool |
spawn_actor | read | Spawn an actor in the current level |
task_submit | write | Submit a task to the async queue |
unreal_get_project_context | read | Get full project context: C++ class list, source structure, level actors, asset counts. Call when you need project-specific details. |
unreal_get_ue_context | read | Get Unreal Engine 5.7 API context/documentation. Categories: ${listCategories().join( |
unreal_status | read | Check if Unreal Editor is running with the plugin. Currently: NOT CONNECTED. Please start Unreal Editor with the plugin enabled. |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (1 observation(s))
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (7)
console.log(`\n Estimated token savings per scenario set: ~${formatNum(charsToTokens(totalBefore - totalAfter))} tokens`);} from "../../context-loader.js";
} from "../../lib.js";
import { resolveUnrealTool, classifyTool, categorizeToolForStatus } from "../../tool-router.js";} from "../../context-loader.js";
} from "../../lib.js";
@modelcontextprotocol/sdk, @vitest/coverage-v8, vitest
Gates applied: no_behavioural_pass.
f75772f7c397full audit observations/trust-audit/mcp-server/natfii__ue5-bridge.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 |
|---|---|---|---|---|---|
| 2026-10-07 | f75772f7c397 | SAFE | B | 89 | first audit |
Questions
What is the UE5 Bridge MCP server?
MCP server bridging AI assistants to Unreal Engine 5 editor
What tools does UE5 Bridge expose?
14 in total: 11 read-only, 3 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is UE5 Bridge 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.
What credentials does UE5 Bridge need?
No credential environment variables were found in its source, so it appears to need none.
How does UE5 Bridge run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as ue5-mcp-server at 1.5.0.
How current is this page?
The grade is for one exact copy of the source (f75772f7c397), read on 2026-10-07. The repository is watched and re-audited when it changes.