In the last fifteen years, individuals have brought hundreds of cases challenging government national security practices for violating human rights or civil liberties. Courts have reviewed relatively few of these cases on the merits, often deferring broadly to the executive branch on the grounds that they lack expertise, political accountability, or the ability to protect national security secrets. Yet in cases where courts have permitted civil suits to proceed far enough to decide legal questions, influence policy, or afford litigants relief, they have often experimented with new methods for managing the secret information implicated in many national security cases. These procedures include centralizing cases through Multidistrict Litigation, conducting in camera review of sensitive documents, pressing the government to provide opposing counsel access to secret evidence, appointing special experts of their own, facilitating interlocutory review, and deciding cases in an incremental and dynamic fashion. Illuminating this procedural experimentation, this Article contends that courts can address secrecy in national security adjudication in a tailored, pragmatic fashion, rather than deferring to the executive at the threshold. But this account also shows the limits of such strategies: where misapplied, some procedures may fall short of due process, undermine norms of public access and transparency in the courts, reduce pluralism in the adjudication of disputes, or import bias into judicial decision-making. Together, this suggests that courts should adopt these procedures cautiously and with case-specific assessment of their costs and benefits. Panning out from national security litigation, the Article also offers a set of secondary insights for civil procedure more generally: it highlights the role of the executive branch in making procedural law, the costs of certain transsubstantive procedures, and distorted perceptions across the civil–criminal procedure divide.