Attack success is a measured outcome indicating whether an offensive action achieved a defined objective in a test or operation. The objective might be gaining access, retrieving data, or completing a specified action. Because the meaning depends on how success is defined, reports should state the goal, test conditions, and how outcomes were counted.
In cybersecurity, this measure helps compare attack methods and evaluate defenses. For example, a test of deceptive context for AI agents might count how often an agent reaches its target despite encountering decoys. A reported reduction in attack success is difficult to interpret without a baseline and clear definition of the target outcome; it does not by itself prove a defense will work in other environments. Defenders should treat such results as one measure alongside controls such as least privilege, logging, and network restrictions.