hytale-model-creator
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- EXTERNAL_DOWNLOADS (HIGH): The 'First-Time Setup Flow' in
SKILL.mdinstructs the agent to guide users to clone an untrusted repository (https://github.com/enfp-dev-studio/blockbench-mcp.git). This repository is not associated with any trusted organizations or the official Hytale team. - REMOTE_CODE_EXECUTION (HIGH): The setup flow includes commands for
pnpm install,pnpm build, and ultimately executing a Node.js script (node .../dist/index.js) from the cloned repository. Executing third-party code from unverified sources poses a significant risk of arbitrary code execution on the user's machine. - COMMAND_EXECUTION (MEDIUM): The skill directs the agent to create and modify a
.mcp.jsonconfiguration file in the workspace root. This file is used to register a local MCP server, which gives the agent persistent capability to launch and interact with the third-party Node.js process. - METADATA_POISONING (MEDIUM): The skill presentation mixes official links (
hytale.com,cdn.hytale.com) with the unofficialenfp-dev-studiorepository. This may lead users to believe the MCP bridge and its associated scripts are official or endorsed tools when they are actually unverified community contributions.
Recommendations
- AI detected serious security threats
Audit Metadata