Safety middleware for AI computer use agents that monitors, enforces policies, and provides guardrails to enable safe desktop interaction.
