Repo-To-Skill：將 GitHub 儲存庫蒸餾為 AI4AI 技能

摘要

自主代理正開始端到端地執行機器學習（ML）研究。這些代理將模型主幹與規劃、執行、記憶和驗證的調度框架相結合，但這種架構仍將領域特定的專業知識留在代理之外。我們將這一缺失層稱為操作知識——即區分「知道某種方法」與「讓方法真正運作」的訣竅。這種知識並非不存在於該領域之中。它出現在程式庫和論文中，但以面向人類讀者的形式編寫，且規模過大，無法在任務進行中載入。一旦將其蒸餾為精簡且經過驗證的技能，這種知識便能在任務之間重複使用，而非在每次運行中重新發現。我們提出 DisCo，一個以技能驅動的研究代理，能夠建立技能並在研究過程中使用它們。其蒸餾以兩種互補形式進行：任務無關形式，將該領域廣泛使用的程式庫濃縮為可重用的技能；以及任務導向形式，產出具體任務所需的技能。前者應用於開源生態系統，產出 AREX-Skill 技能庫，其中包含從 1,000 個廣泛使用的 ML 程式庫中蒸餾出的 5,000 多項經驗證技能，並組織為 20 個領域和 178 個能力家族。在 GPT-5.5 主幹、研究調度框架和下游執行預算保持固定的條件下，配備技能的研究代理在 MLE-bench 上得分提高 134.3%，在 PaperBench 上提高 34.4%，在 FrontierCS 上提高 9.2%，在 PassNet 上提高 14.0%，相較於未配備技能的同一代理。這些提升來自於在該固定設置下添加了蒸餾後的操作背景知識。

English

Autonomous agents are beginning to carry out machine-learning (ML) research end to end. These agents combine a model backbone with a harness for planning, execution, memory, and verification, but this architecture still leaves domain-specific know-how outside the agent. We call this missing layer operational knowledge, the know-how that separates knowing a method from making it work. That knowledge is not absent from the field. It appears in repositories and papers, but in forms written for human readers and too large to load during a task. Once distilled into compact, verified skills, this knowledge can be reused across tasks rather than rediscovered during each run. We present DisCo, a skill-powered research agent that creates skills and uses them during research. Its distillation runs in two complementary forms: task-agnostic, condensing the field's widely used repositories into reusable skills, and task-oriented, producing the skills a concrete task calls for. The former, applied across the open ecosystem, yields the AREX-Skill Library, with 5,000+ verified skills distilled from 1,000 widely used ML repositories and organized into 20 areas and 178 capability families. With the GPT-5.5 backbone, research harness, and downstream execution budget held fixed, the skill-equipped research agent scores 134.3% higher on MLE-bench, 34.4% higher on PaperBench, 9.2% higher on FrontierCS, and 14.0% higher on PassNet than the same agent without skills. These gains come from adding distilled operating context under that fixed setup.