byod-hardware-orchestrator

Fail

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill implements a 'Command Execution Tunnel' that allows the AI to send and execute terminal commands on a user's PC or Mac after a pairing process.
  • [EXTERNAL_DOWNLOADS]: The instructions direct users to run external commands like curl or npx from the internet to install a local agent, which is a common vector for executing unverified remote code.
  • [COMMAND_EXECUTION]: The skill's primary purpose is to 'break out of the cloud sandbox' to interact with local file systems, move the mouse cursor, and control installed applications (VS Code, Excel, Chrome) via automation scripts.
  • [DATA_EXFILTRATION]: By gaining access to the local file system and establishing a secure tunnel, the skill creates a technical pathway for reading and potentially transmitting local sensitive data to the AI's environment.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Apr 18, 2026, 05:56 PM
Security Audit — agent-trust-hub — byod-hardware-orchestrator