An Empirical Study: AI Agent Rules Need Context and Layered Enforcement

An empirical study on the enforcement gap between AI agent rules and the subset a system can actually check, highlighting the importance of context and layered enforcement in ensuring AI agent safety.