BeyondTrust announced on June 30, 2026, that AI Agent Security, a private beta module for its Pathfinder platform, will enforce endpoint privileges for AI coworkers and autonomous agents across Windows, macOS, Linux, and containers before those systems can act. The pitch is simple but consequential: the next privileged user on a corporate device may not be a user at all. It may be Claude Code, Microsoft Copilot, Cursor, OpenAI Codex, or another agent acting with borrowed authority...