Principal–agent theory and game theory are applied to the precautionary principle (PP) to open up a new research agenda. Principals assess whether the threat is uncertain above a threshold. If it is, the principals choose, pay, and command agents to decrease the uncertainty below the threshold. The agents perform the action. The process is repeated through a feedback loop impacting the threat, after which the process is renewed. The four dimensions of the PP, that is, threat, uncertainty, command, and action, are described. Games and game characteristics in the four dimensions are recognized. Games are possible between natural, technological, and human factors causing the threat and between principals, agents, and external actors. Moral hazard and adverse selection in principal–agent theory related to the PP are considered. Twelve kinds of uncertainty are identified for principal–agent theory in the PP, that is, the natures of the threat, uncertainty, and threshold; states of nature, technology, knowledge, and information; whether a game is played; players; which game is played; strategy sets; utilities; beliefs; incomplete information; imperfect information; risk attitudes; and bounded rationality.