There is a curious absence of legal constraints on U.S. government agencies undertaking potentially risky scientific research. Some of these activities may present a risk of killing millions or even destroying the planet. Current law leaves it to agencies to decide for themselves whether their activities fall within the bounds of acceptable risk. This Article explores to what extent and under what circumstances the law ought to allow private actions against such non-regulatory agency endeavors. Engaging with this issue is not only interesting in its own right, it allows us to test fundamental concepts of agency competence and the role of the courts. Two case studies provide a foundation for discussion: NASA's use of plutonium power supplies on spacecraft, which critics say could cause millions of cancers in the event of atmospheric disintegration, and a Department of Energy particle-collider experiment (the RHIC, or Relativistic Heavy Ion Collider) that allegedly poses a small risk of collapsing the Earth (via a "killer strangelet"). These extreme examples serve as a test-bed for applying insights from neoclassical economics, behavioral economics, risk-management studies, and cognitive psychology. The resulting analysis suggests that in low-probability/high-harm scenarios, agencies are likely to do a poor job of judging the acceptability of risk to the public. Instead, generalist judges working in a common-law vein may have surprising advantages. This in turns suggests that under certain circumstances the government should be subject to legal action that provides non-deferential review of discretionary agency actions that are non-regulatory in nature.