Several studies in the last two decades have revealed that false criminal informant testimony is a leading factor in wrongful convictions, along with false confessions, eyewitness misidentification, and faulty forensic science. Although a great deal more remains to be done, many jurisdictions have implemented evidence-based reforms to these last three categories of evidence. Policy about criminal informants, however, seems to be stubbornly stagnant, and relevant social science is virtually nonexistent. This Article questions the relative lack of attention to informant testimony and suggests that the dangers posed by informant testimony are both greater and different than previously thought. Unlike much of the prior literature on the subject, this Article carefully distinguishes between jailhouse informants and other types of informant witnesses, especially accomplices. Although both categories of informants pose some of the same risks to the reliability of proceedings, accomplice witnesses pose additional risks — many enabled by rules of evidence, trial practices, and psychological phenomena — that have not yet been fully appreciated in the literature. After identifying these concerns, this Article concludes with recommendations for reform — and areas requiring further study — with the aim of developing a set of evidence-based best practices for the use of informant testimony.