Workday 报告：Agent Skills 渐进式加载的实测影响
Report: Progressive Disclosure of Agent Skills
Workday 把生产环境里 agent 技能库按需加载的真实数据写成了报告，想优化成本和延迟的可以看下取舍结果。
Workday 的 LLM 智能体用户常通过定义命名技能（skills）来扩展智能体能力。随着技能库变大，运营成本随之上升。该报告实测渐进式披露（lazy-loading）的效果：技能检索质量有所提升，但整体延迟轻微变差。
Report: Progressive Disclosure of Agent Skills
Users of Workday's deployed LLM-based agents often request features which can be addressed by defining named procedures, also known as skills, in the LLM context, effectively augmenting agents' capabilities. However, as an agent's skills library grows in size, so does the agent's operational cost. Progressive disclosure (lazy-loading) of skills as needed may reduce operational costs, but its impact on overall latency and skill-retrieval quality remains unclear. In this report, we investigate the impact empirically and find that progressive disclosure improves skill-retrieval quality but marginally degrades overall latency.
