openai-screenshot

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes platform-specific utilities including screencapture on macOS and scrot, gnome-screenshot, or import on Linux to perform screen captures.
  • [DYNAMIC_EXECUTION]: On macOS, the Python helper executes bundled Swift scripts to access CoreGraphics and AppKit APIs for window and permission management.
  • [DYNAMIC_EXECUTION]: The Windows PowerShell script dynamically compiles C# code snippets using Add-Type to interface with the Win32 API for foreground window detection and boundary calculation.
  • [PRIVILEGE_ESCALATION]: The skill instructions specify the use of the -ExecutionPolicy Bypass flag for PowerShell to facilitate the execution of the local capture script regardless of system-wide policy settings.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external data such as application names and window titles into command arguments, employing basic sanitization like quote escaping for AppleScript execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 04:27 AM
Security Audit — agent-trust-hub — openai-screenshot