How an assessor could investigate control and monitoring of internal AI agents

Our proposal is to analyze whether a developer’s internal agents could get away with serious harm without being caught.