Get Started
Research questionHow can research agents reuse operational know-how when repositories are too large to load during tasks?Research agents may know a method in principle but lack the practical details needed to make it work. Those details are scattered across repositories and papers and are too extensive to load wholesale during a task.
AI
AI Agents
Machine Learning
Latest papersRecent research connected to this question, newest first.Repo-To-Skill: Distilling GitHub Repositories Into AI4AI SkillsThe source describes DisCo, a research agent that creates and uses verified skills distilled from machine-learning repositories. Its evidence includes the AREX-Skill Library, containing more than 5,000 skills from 1,000 repositories, and reported downstream benchmark gains under a fixed backbone, harness, and execution budget; it does not establish broader deployment conditions beyond those experiments.research paper · Sep 2, 2026
Related questions
How can LLM agents reuse procedural skills across diverse long-horizon tasks without generic or instance-bound memory?How can data-centric agents reuse prior intent-to-object discoveries without causing retrieval interference?How can computer-use agents retain, refine, and reliably reuse procedural skills across repeated GUI tasks?How can AI-agent memory survive model upgrades without losing retrieval or repairability?
Home
Topics
Search
Library