Counterfactual Trace Auditing of LLM Agent Skills 事件

Counterfactual Trace Auditing of LLM Agent Skills arXiv:2605.11946v2 Announce Type: replace Abstract: Large Language Model agents are increasingly augmented with agent skills. Current evaluation methods for skills remain limited. Most deployed benchmarks report only pass rate before and after a skill is attached, treating the skill as a black box change to agent behavior. We introduce Counterfactual Trace Auditing (CTA), a framework for measuring how a skill changes agent behavior. CTA pairs ea