Corpus Input + Prompts for Skill Generator Baseline

[TheYoungling1]

Hi team,

Great work on the paper, I definitely think its the right direction to take from the original SkillBench paper.

However, I just have 2 questions:

1. The paper is about how to use raw corpus info to generate Skill files, however, in the code, there are only "Material" Folders under the task folders that most of the time only contains the raw readme file. Are these the "source materials/visible corpus" mentioned in the paper.
2. Can the authors expand more on how the prompts and settings used for the 4 Skill Generation baselines (SkillNet, Evoskills etc..), just curious as to how you passed in the corpus information to these agent/tools.

Thanks