SkillAtlas: An Attack Trace Library for Agent Skills

T1Covered by 1 source · first reported by arXiv at 24 Sept, 04:00 UTC

arXiv:2609.13353v2 Announce Type: replace-cross Abstract: Agent skills are reusable units for language-model agents, but their risks emerge through model decisions, user context, tool calls, and execution feedback rather than through stable signatures or a single sandbox run. Existing static…

science

Sources

• T1SkillAtlas: An Attack Trace Library for Agent SkillsarXiv24 Sept, 04:00