SkillJack：自演化智能体中的持久性技能后门

摘要

自进化智能体日益将交互历史转化为可复用的技能，这些技能能够超越单个任务而持续存在。尽管已有工作研究了记忆与检索投毒，但此类攻击仅在受污染记录被作为上下文检索时才会影响智能体。我们揭示了一个新的、更为根本的风险：受污染的体验可被智能体自身转化为持久的行为产物。我们提出了SkillJack，这是首个利用自进化智能体经验-技能流水线的攻击。SkillJack并非直接操纵运行时上下文，而是劫持智能体自身的学习过程，将恶意行为植入其可复用的技能库中。我们识别出这一转化的三个关键特性：清洗粉饰（sanitization whitewashing），即恶意意图在技能提取过程中被掩盖；跨层提升（cross-layer promotion），即瞬时经验转变为持久能力；以及持久性隔离（persistence isolation），即攻击在移除其原始来源记录后仍然存活。我们在两个代表性系统SkillX和Anything2Skill上评估了SkillJack，使用包含四个策略风险类别、共150条轨迹的共享数据集。结果表明，技能提取显著降低了攻击的可检测性：在SkillX中，安全检测率从针对受污染轨迹的98.5%骤降至针对提取技能的11.4%，而Anything2Skill也呈现出类似效果。与此同时，植入的技能保持有效性，在两个系统上分别实现了56.2%和89.2%的攻击成功率。此外，80.0%的技能介导攻击在删除原始受污染记录后仍然持续存在，且部分技能会在良性查询上意外触发。我们的发现揭示了技能演化是一个新的攻击面，并促使我们倡导具备来源感知的技能生命周期保护。我们的代码可在https://github.com/Tencent/AI-Infra-Guard/research/skilljack获取。

English

Self-evolving agents increasingly convert interaction histories into reusable skills that persist beyond individual tasks. While prior work studies memory and retrieval poisoning, such attacks only affect agents when poisoned records are retrieved as context. We uncover a new and more fundamental risk: poisoned experiences can be transformed by the agent itself into durable behavioral artifacts. We present SkillJack, the first attack that exploits the experience-to-skill pipeline of self-evolving agents. Instead of directly manipulating runtime context, SkillJack hijacks the agent's own learning process to implant malicious behaviors into its reusable skill repertoire. We identify three key properties of this transformation: sanitization whitewashing, where malicious intent is obscured during skill extraction; cross-layer promotion, where transient experiences become persistent capabilities; and persistence isolation, where the attack survives removal of its original source records. We evaluate SkillJack on two representative systems, SkillX and Anything2Skill, using a shared dataset of 150 trajectories across four policy-risk categories. Results show that skill extraction substantially reduces attack detectability: in SkillX, safety detection drops from 98.5\% for poisoned trajectories to 11.4\% for extracted skills, while Anything2Skill shows a similar effect. Meanwhile, the implanted skills remain effective, achieving attack success rates of 56.2\% and 89.2\% on the two systems, respectively. Furthermore, 80.0\% of skill-mediated attacks persist after deleting the original poisoned records, and some skills unintentionally activate on benign queries. Our findings reveal skill evolution as a new attack surface and motivate provenance-aware skill lifecycle protection. Our code is available at https://github.com/Tencent/AI-Infra-Guard/research/skilljack.