@misc{indiciaef26198188727,
  title = {One Skill Does Not Fit All: Automatic Discovery and Taxonomy-Guided Routing of Frame-Selection Skills for Long-Video Question Answering},
  author = {Jian Hu and Zixu Cheng and Da Li and Wei Li and Ziquan Liu and Shaogang Gong},
  year = {2026},
  url = {https://arxiv.org/abs/2609.12517},
  note = {Source identifier: 2609.12517}
}