Minecraft IntegrationSAFE
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
A Model Context Protocol (MCP) integration for Minecraft that enables AI assistants to interact with a Minecraft server. This integration allows AI models to observe and interact with the Minecraft world through a bot.
Prerequisites
- Minecraft Launcher
- Node.js 18 or higher
- Claude Desktop App
- Java 21.0.5 (recommended)
⚠️ Note: Currently only tested on macOS/Linux. Windows compatibility is not guaranteed.
Important Note
- Use the F3+P Shortcut:
Press F3 + P together. This toggles the "Pause on Lost Focus" feature. Once turned off, you can switch to claude desktop and Minecraft will continue running without pausing.
- Connection Issues on Claude Restart:
If you restart Claude while the Minecraft server is running, you may experience MCP connection issues on the next claude launch due to lingering java process. See Troubleshooting: MCP Connection Failed for resolution steps.
Installation Steps
- Download and Setup Minecraft Server
- Download Minecraft server v1.21 from mcversions.net/1.21
- Install Java 21.0.5 if not already installed (other versions are untested)
- Create a dedicated directory (e.g.,
~/minecraft-server/) - Place the downloaded
server.jarfile in this directory - Note down the absolute path to your
server.jarfile
- Install and Configure MCP Integration
Quick Install (Recommended):
npx -y @smithery/cli install mcp-minecraft --client claude
Follow the CLI prompts to complete the setup.
Or Manual Setup:
- Navigate to
~/Library/Application Support/Claude/claude_desktop_config.json - Add the MCP server configuration:
{
"mcpServers": {
"mcp-minecraft": {
"command": "npx",
"args": [
"-y",
"mcp-minecraft165b8ef13a0aOBSERVED · 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 mcp-minecraft -- npx -y [email protected]
{
"mcpServers": {
"mcp-minecraft": {
"command": "npx",
"args": [
"-y",
"[email protected]"
]
}
}
}Exposed tools (21)
16 read · 5 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
attack | read | Attack a nearby entity by name |
chat | write | Send a chat message |
digBlock | read | Break a block at specified coordinates |
equipItem | read | Equip an item by name |
followPlayer | read | Follow a specific player |
getBlockInfo | read | Get information about a block at specified coordinates |
getInventory | read | Get contents of bot |
getNearbyEntities | read | Get list of nearby entities within specified range |
getStatus | read | Get bot |
goToPosition | read | Navigate to specific coordinates |
jump | read | Make the bot jump |
lookAt | read | Make the bot look at specific coordinates |
moveBack | write | Make the bot move backward |
moveForward | write | Make the bot move forward |
placeBlock | read | Place a block at specified coordinates |
selectSlot | read | Select a hotbar slot (0-8) |
stopFollowing | write | Stop following current target |
stopUsingItem | write | Stop using/deactivate the current item |
turnLeft | read | Make the bot turn left |
turnRight | read | Make the bot turn right |
useItem | read | Use/activate the currently held item |
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
- none-observed
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (2)
@modelcontextprotocol/sdk, @types/node, minecraft-protocol, mineflayer, mineflayer-pathfinder, ts-node, typescript, yargs
public/screenshot.png
Gates applied: no_behavioural_pass.
165b8ef13a0afull audit observations/trust-audit/mcp-server/arjunkmrm__minecraft-integration.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 | 165b8ef13a0a | SAFE | B | 89 | first audit |
Questions
What tools does Minecraft Integration expose?
21 in total: 16 read-only, 5 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is Minecraft Integration 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 Minecraft Integration need?
No credential environment variables were found in its source, so it appears to need none.
How does Minecraft Integration run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as mcp-minecraft at 1.0.34.
How current is this page?
The grade is for one exact copy of the source (165b8ef13a0a), read on 2026-10-07. The repository is watched and re-audited when it changes.