RESOURCE2SKILL converts human-created multimodal resources, including tutorial videos, repositories, articles, and reference artifacts, into executable skills for software agents. It stores these skills in a hierarchical multimodal Skill Wiki combining structured text, code, visual examples, metadata, and provenance. At inference time, agents retrieve and compose relevant skills; when the library lacks coverage, the same construction operator can acquire new skills online. The paper reports an average overall improvement of 11.9 percentage points over agents without skills across seven practical authoring domains, outperforming strong harness baselines in 26 of 28 main model-domain aggregate cells. Ablations examine format, hierarchy, source diversity, selection, and online acquisition.
No heat snapshots are available in the last 24 hours.