SkillsBench 1.1 Launch Party @ ACM CAIS

​BenchFlow is excited to announce SkillsBench 1.0 containing more than 100 expert curated tasks measuring how well agents use skills across diverse and complex domains.

​In collaboration with Kaggle, we're launching an afterparty for ACM CAIS Agent Skills'26 workshop featuring researchers and practitioners working on Skills design, benchmarking, optimization, security, and ecosystem infrastructure.

​We'll feature Live Demos/Talks:

• ​7:00pm: Networking & chatting (Ivan)

• ​7:30pm: Presentations start (Ivan to open)

  • ​7:30pm - 7.45pm:  ‘Community Is All You Need’, Xiangyi Li, creator of SkillsBench

  • ​7:45pm - 8:00pm: SkillsBench 1.1 -- Wenbo Chen, coauthor of SkillsBench

  • ​8:00pm - 8:15pm: A Taxonomy of RL environments for LLM Agents -- Han Lee

  • ​8:15pm - 8.30pm: ‘Agentic evaluation at scale – for everybody’. Kaggle (Nick)

• ​8:30pm: Networking

• ​9:00pm: Event close

​Limited spaces available, sign up today! Excited to see everyone at the venue :)

​Hosted in partnership with Kernel Labs!