microsoft-powertoys

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes PowerShell and C# code segments designed to manage system processes and services. Specifically, the Configure-PowerToys.ps1 script uses Stop-Process and Start-Process to restart the PowerToys runner, while the C# plugin template uses ServiceController to start and stop Windows services.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and processing data from the local system environment which could be manipulated.
  • Ingestion points: Reads local configuration files (settings.json, zones-settings.json) and queries Windows service metadata (display names and service names) via C#.
  • Boundary markers: None; the scripts and instructions lack delimiters or explicit 'ignore embedded instructions' warnings when processing system-derived strings.
  • Capability inventory: The skill possesses capabilities to write to the file system (Set-Content), terminate processes (Stop-Process), and change the state of Windows services (svc.Start(), svc.Stop()).
  • Sanitization: None; there is no evidence of validation or sanitization for service names or configuration values before they are used in agent logic or UI display cards.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 12:20 AM