lov-finder-action
Fail
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill generates and executes shell scripts to register Automator workflows, manipulate system services (pbs, killall Finder), and build Swift applications using xcodebuild.- [REMOTE_CODE_EXECUTION]: The instructions explicitly facilitate the creation of a 'Helper App' designed to bypass macOS Sandbox restrictions. This app reads commands from the system clipboard (pbpaste) and executes them via AppleScript (osascript), providing a path for arbitrary code execution.- [DATA_EXFILTRATION]: The skill configuration for Mode B (Finder Sync Extension) requests broad read/write access to the entire root directory ('/') using the 'com.apple.security.temporary-exception.files.absolute-path.read-write' entitlement, significantly expanding the data exposure surface beyond standard application boundaries.- [PROMPT_INJECTION]: The skill possesses a high surface area for indirect prompt injection (Category 8) because user-provided 'core commands' and 'action names' are directly interpolated into generated shell scripts and Swift source code (e.g., in run.sh and project.yml) without instructions for sanitization or boundary markers.- [PERSISTENCE_MECHANISMS]: The skill creates persistent artifacts on the user's system, including Automator workflows in '~/Library/Services/' and standalone helper applications in '/Applications/', which remain active and accessible across reboots.
Recommendations
- AI detected serious security threats
Audit Metadata