blockbench-hytale

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION] (SAFE): The skill contains only instructional and descriptive content for Hytale modeling. No override markers, role-play injections, or attempts to bypass safety filters were found.
  • [DATA_EXFILTRATION] (SAFE): No network operations (curl, wget, fetch) or access to sensitive file paths (~/.ssh, ~/.aws) are present. The skill uses an internal URI scheme (hytale://) for resource referencing.
  • [REMOTE_CODE_EXECUTION] (SAFE): There are no commands to download or execute remote scripts. The mention of a prerequisite plugin is an instruction for the user to set up their environment manually.
  • [COMMAND_EXECUTION] (SAFE): The skill defines MCP-style tools but includes no shell command execution, subprocess spawning, or system-level operations.
  • [EXTERNAL_DOWNLOADS] (SAFE): No package managers (npm, pip) or external script downloads are utilized.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:33 PM