Pluto Security discloses $42 million for controls that interrupt risky AI-agent actions

The startup says it can intervene while an agent works. Its CEO says most risks it detects come from AI-built tools made by non-developers.