HN 提问：有人测量过你们发布的技能被代理使用了多少次吗？
1 分•作者: sohaibtariq•4 天前
如今，许多SaaS和API平台都会发布代理技能。我们也发布了，但我无法判断它们对真实用户的实用性。
我们有评估来测试技能在理想条件下的运行情况，但我对它们在实际环境中的运行时行为更感兴趣。例如：
代理在多大程度上会回退到网络搜索，而不是遵循技能的指示？
对于普通用户来说，已安装的技能会被触发多少次？
除了skills.sh等平台报告的安装量之外，是否有人在跟踪实际使用情况？尝试获取这些洞察是否有意义？
查看原文
A lot of SaaS and API platforms now publish agent skills. We did too, but I cannot tell how useful they are for real users.<p>We have evals for testing that the skills work under ideal conditions, but I'm interested in runtime behaviour in the wild. For instance:<p>How often did an agent fall back to web search instead of following what the skill said?
How often does an installed skill get triggered for an average user?<p>Is anyone tracking real use, beyond the install counts that platforms like skills.sh report? Does it even make sense to try and get these insights?