polyhaven-scene-builder

Fail

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/polyhaven-scene-builder.js takes command-line arguments such as --ground and --pedestal-texture and interpolates them directly into Python code strings.
  • [COMMAND_EXECUTION]: The interpolated Python code is sent via a TCP socket to a Blender MCP instance (port 9876) using the execute_code command.
  • [COMMAND_EXECUTION]: Because there is no sanitization of the input arguments, an attacker could craft an input (e.g., "); import os; os.system('...'); #) to break out of the Python string literal and execute arbitrary system commands in the environment where Blender is running.
  • [REMOTE_CODE_EXECUTION]: The skill uses a dynamic execution pattern where it generates executable Python code at runtime based on external parameters, creating a high-risk attack surface.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 15, 2026, 11:50 PM
Security Audit — agent-trust-hub — polyhaven-scene-builder