一句话：提出基于领域策略的 xPloiT 探索性试玩框架，结构化记录 playtraces，为类人智能体试玩建模提供数据。
视频游戏中的缺陷会挫伤玩家体验并影响产品成功。面对复杂动态的游戏系统，系统化测试实践仍依赖人工试玩，因为自动化难以捕捉玩家体验与涌现行为。本文提出 xPloiT：一个基于策略的探索性试玩框架。它通过领域特定策略组织测试，指导试玩者规划、执行并记录测试会话，形成结构化的 playtraces。这些 playtraces 将作为数据，用于建模能够模拟探索行为的智能体试玩系统，从而支持类人智能体复现真实测试行为，并推动复杂游戏环境中的人因驱动与半自动化测试。
Bugs in video games can frustrate players and compromise product success. Ensuring quality in these complex, dynamic systems requires systematic testing practices grounded in software engineering. While automated testing has advanced in other domains, games still depend on manual playtesting to capture the player’s experience and detect emergent behaviors that automation cannot easily replicate. This research proposes xPloiT, a strategy-based framework for exploratory playtesting. The framework organizes testing through domain-specific strategies that guide playtesters in planning, executing and documenting test sessions. The resulting structured playtraces will serve as data for modeling agentic playtesting systems that emulate exploratory behaviors. xPloiT aims to create human-like agents capable of reproducing real testing behaviors, contributing to both human-informed and semi-automated testing in complex game environments.