Agent Skills allow LLM agents to combine reusable capabilities, but checking each Skill in isolation cannot rule out policy violations when they are composed…
机构:中科院
来源:arXiv 2610.35082 | AI4Papers 论文推荐平台