skill-launcher

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION] (MEDIUM): The skill executes a local binary in the background using 'nohup'. This creates a persistent process that runs independently of the agent session, which could be used to hide malicious activity.
  • [DYNAMIC_EXECUTION] (MEDIUM): The instructions require a manual build step ('swift build') to compile code on the host system. Executing or compiling unverified local source code is a security risk.
  • [PRIVILEGE_ESCALATION] (LOW): The skill documentation explicitly notes that 'Accessibility' permissions are required. While common for launcher applications, this permission allows for monitoring keystrokes and controlling other applications.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 08:02 AM