skills/hironow/skills/screenshot/Gen Agent Trust Hub

screenshot

Warn

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill instructions in "SKILL.md" advise using "powershell -ExecutionPolicy Bypass" to execute the Windows capture script, which bypasses the default PowerShell script execution security policy.
  • [DYNAMIC_EXECUTION]: The Windows script "scripts/take_screenshot.ps1" uses the "Add-Type" cmdlet to compile and load C# code at runtime for interacting with the Windows API to perform screen and window capture.
  • [COMMAND_EXECUTION]: The script "scripts/take_screenshot.py" uses "subprocess.run" to call multiple OS utilities including "screencapture", "scrot", "gnome-screenshot", "import", "osascript", and "swift" to perform screenshots and gather system information.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing user-provided inputs without robust validation.
  • Ingestion points: User-specified file paths, application names, and window titles are ingested into the context via command-line arguments.
  • Boundary markers: Absent. The skill does not use delimiters to isolate untrusted user data from shell commands.
  • Capability inventory: The skill can write files to the file system and execute shell commands via subprocess calls.
  • Sanitization: Only basic character escaping (double quotes) is performed in specific command strings in "take_screenshot.py", which is insufficient to prevent command injection if malicious strings are provided as application names or paths.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 7, 2026, 02:48 PM
Security Audit — agent-trust-hub — screenshot