Cybersecurity awareness training has historically adopted a one-size-fits-all approach, despite established individual differences in how users process and retain security information. Personality has been proposed as one axis along which training content might be tailored; yet no prior study has implemented and empirically evaluated a complete personality-conditional system end-to-end. This paper reports the design, implementation, and quasi-experimental evaluation of \emph{TailoredSec}, a mobile cybersecurity awareness application that routes training content based on a user's dominant Five-Factor Model (FFM) personality trait, as measured by the ten-item Big Five Inventory (BFI-10). Seventy-four UK-based adults were allocated to a traditional video-training condition ($n = 40$) or a personality-conditional condition ($n = 34$). Both groups completed a four-item scenario-based pre-assessment (scored 0--40), a single training session, and an equivalent post-assessment. The personality-conditional group additionally completed the BFI-10 (Big Five Inventory-10) and was routed to one of four training modules covering five FFM traits (Conscientiousness and Neuroticism share a module). Pre-assessment scores did not differ between groups ($t(69.1) = 0.43$, $p = .67$), confirming baseline equivalence. The personality-conditional group scored significantly higher on the post-assessment ($M = 35.88$, $SD = 5.00$ vs $M = 30.75$, $SD = 10.23$; Welch's $t(58.5) = 2.81$, $p = .007$; Cohen's $d = 0.62$; 95\% CI $[1.47, 8.79]$ marks), with a pass-rate of 100\% versus 77.5\% (Fisher's exact $p < .01$). These results offer preliminary support for personality-conditional content routing as a feasible design principle for cybersecurity awareness training.

翻译：网络安全意识培训历来采用"一刀切"模式，尽管用户在安全信息的处理与记忆方面存在显著的个体差异。人格特质已被提出可作为培训内容定制的维度之一，但尚无研究完整实现并实证评估端到端的人格条件驱动系统。本文报告了移动网络安全意识应用《TailoredSec》的设计、实现与准实验评估——该系统通过十项大五人格量表（BFI-10）测量用户的主导大五人格模型特质，并据此路由培训内容。74名英国成年被试被分配到传统视频培训组（n=40）或人格条件组（n=34）。两组均完成四项情景式前测（评分范围0-40分）、单次培训课程及等效后测。人格条件组额外完成BFI-10量表测试，并被分流至覆盖五种大五人格特质（尽责性与神经质共享模块）的四个培训模块之一。前测成绩组间无显著差异（t(69.1)=0.43, p=0.67），证实基线等价性。后测中人格条件组得分显著更高（M=35.88, SD=5.00 vs M=30.75, SD=10.23；韦尔奇t检验：t(58.5)=2.81, p=0.007；科恩d=0.62；95%置信区间[1.47, 8.79]分），通过率达100%对比77.5%（费希尔精确检验p<0.01）。这些结果为人格条件驱动内容路由作为网络安全意识培训的可行设计原则提供了初步实证支持。