Reproduce “SkillEarth: Teaching General-Purpose Agents to Act on Earth via Skill Evolution” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Haojun Tang, Wenda Zhao, Xiang Zhou, Qiwei Lin, Lida Huang, Yan Wang

SkillEarth converts demonstrated remote-sensing tool trajectories into reusable skill files and revises them using execution feedback, separate tool and answer judges, and a graph linking tool outputs to later arguments. On OpenEarth-Bench, its SameOrder score was 25.15, versus SkillX’s 16.77. Adding the graph raised image-task generation accuracy from 23.45 to 24.14; SkillEarth does not lead every task-accuracy comparison.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.