Tool-Use on 兜忆轩 | Cedric's Compiler Noteshttps://douyixuan.github.io/tags/tool-use/Recent content in Tool-Use on 兜忆轩 | Cedric's Compiler Notes兜忆轩 | Cedric's Compiler Noteshttps://douyixuan.github.io/images/papermod-cover.pnghttps://douyixuan.github.io/images/papermod-cover.pngHugo -- 0.146.0en-usThu, 24 Sep 2026 08:00:00 +0800AI Native Digest 2026-09-24https://douyixuan.github.io/posts/ai-native/2026-09-24/Thu, 24 Sep 2026 08:00:00 +0800https://douyixuan.github.io/posts/ai-native/2026-09-24/今天两篇分别聚焦 Agent 评测与 Skill 生态。NVIDIA 给出从 tool-call accuracy 到真实环境 task completion 的评测框架；UCL 的 GitSkills 数据集则把数百万个 SKILL.md 当成新的软件工程研究对象，揭示了 Skill 复制、维护与安全治理问题。