This is a study of street-level bureaucrats, pretrial risk assessments, unintended consequences, and perverse effects in Virginia’s criminal legal system. The overarching questions motivating this work are: how do legal actors interact with risk assessments, and what effects do those interactions have on individual outcomes, agency dynamics, and the pretrial system overall? A growing body of research has emphasized the importance of empirical work on pretrial risk assessments, which considers how judicial decisions are influenced by risk assessments rather than drawing direct comparisons between judge decisions and risk assessment recommendations that judges may or may not follow. My work expands on prior research by incorporating a cumulative inequality framework, utilizing both ethnographic observations and statistical methods, and highlighting legal actors who are frequently overlooked despite their outsized influence on defendant outcomes and the pretrial system overall. Ultimately, this approach permits a far more comprehensive understanding of the ramifications of pretrial risk assessment implementation. I show that jurisdictions cannot depend on pretrial risk assessments alone to promote policy goals such as public safety, decarceration, and racial equity. Their impacts are unpredictable, both in the short term and long term. And at their worst, they introduce perverse effects—promoting harsher punishment, compounding discrimination, and eroding civil liberty.