Feedback (SA team, re: @yongliang Xie): the polym-eval-* skill family is too fragmented — proposal to consolidate.
@yongliang Xie 这个 eval 的 skill 太散乱了，是否可以合并下，比如 by 模型类型出几个 skill 即可，甚至合并成 1 个，完全根据用户意图去判断具体要做的任务是什么，是否也 ok？
Current state — the eval family spans many separate skills:
polym-eval-generate-{gemini,gpt-image,seedance,seedream,sora,veo}, polym-eval-score-{image,video}-gemini, polym-eval-run-evaluation, polym-eval-prompt-optimizer, polym-eval-report-generator, polym-eval-db.
Options to discuss:
- Group by model type → a handful of skills, or
- A single
polym-eval entry that routes by user intent (generate / score / run / report).
Feedback (SA team, re: @yongliang Xie): the
polym-eval-*skill family is too fragmented — proposal to consolidate.Current state — the eval family spans many separate skills:
polym-eval-generate-{gemini,gpt-image,seedance,seedream,sora,veo},polym-eval-score-{image,video}-gemini,polym-eval-run-evaluation,polym-eval-prompt-optimizer,polym-eval-report-generator,polym-eval-db.Options to discuss:
polym-evalentry that routes by user intent (generate / score / run / report).