{"id":8267,"date":"2026-03-06T04:54:00","date_gmt":"2026-03-05T19:54:00","guid":{"rendered":"https:\/\/devneko.jp\/wordpress\/?p=8267"},"modified":"2026-02-28T15:03:09","modified_gmt":"2026-02-28T06:03:09","slug":"skillsbench-benchmarking-how-well-agent-skills-work-across-diverse-tasks","status":"publish","type":"post","link":"https:\/\/devneko.jp\/wordpress\/?p=8267","title":{"rendered":"SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks\u00a0"},"content":{"rendered":"\n<ul class=\"wp-block-list\">\n<li><strong>SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks\u00a0<\/strong>[61.9]<br>\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u30b9\u30ad\u30eb\u306f\u3001LLM\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3092\u63a8\u8ad6\u6642\u306b\u5897\u5f37\u3059\u308b\u624b\u7d9a\u304d\u7684\u77e5\u8b58\u306e\u69cb\u9020\u5316\u30d1\u30c3\u30b1\u30fc\u30b8\u3067\u3042\u308b\u3002 SkillsBench\u306f\u300111\u306e\u30c9\u30e1\u30a4\u30f3\u306b\u308f\u305f\u308b86\u306e\u30bf\u30b9\u30af\u306e\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3067\u3042\u308a\u3001\u30ad\u30e5\u30ec\u30fc\u30c8\u3055\u308c\u305f\u30b9\u30ad\u30eb\u3068\u6c7a\u5b9a\u8ad6\u7684\u691c\u8a3c\u3068\u306e\u7d44\u307f\u5408\u308f\u305b\u3092\u793a\u3059\u3002 \u30ad\u30e5\u30ec\u30fc\u30c8\u3055\u308c\u305f\u30b9\u30ad\u30eb\u306f\u5e73\u5747\u30d1\u30b9\u30ec\u30fc\u30c8\u309216.2\u30dd\u30a4\u30f3\u30c8(pp)\u4e0a\u6607\u3055\u305b\u308b\u304c\u3001\u52b9\u679c\u306f\u9818\u57df\u306b\u3088\u3063\u3066\u5927\u304d\u304f\u7570\u306a\u308b\u3002 \u81ea\u5df1\u751f\u6210\u30b9\u30ad\u30eb\u306f\u5e73\u5747\u7684\u306b\u4f55\u306e\u5229\u76ca\u3082\u4e0e\u3048\u305a\u3001\u30e2\u30c7\u30eb\u304c\u6d88\u8cbb\u304b\u3089\u5f97\u3089\u308c\u308b\u624b\u7d9a\u304d\u7684\u306a\u77e5\u8b58\u3092\u78ba\u5b9f\u306b\u8aac\u660e\u3067\u304d\u306a\u3044\u3053\u3068\u3092\u793a\u3057\u3066\u3044\u308b\u3002<br><a href=\"http:\/\/arxiv.org\/abs\/2602.12670v1\">\u8ad6\u6587<\/a>\u00a0\u00a0<a href=\"https:\/\/fugumt.com\/fugumt\/paper_check\/2602.12670v1\">\u53c2\u8003\u8a33\uff08\u30e1\u30bf\u30c7\u30fc\u30bf\uff09<\/a>\u00a0 \u00a0(Fri, 13 Feb 2026 07:06:06 GMT)<\/li>\n\n\n\n<li>Skills\u306b\u95a2\u3059\u308b\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u3002Findings\u304c\u300c(1) curated Skills provide substantial but variable benefit (+16.2 percentage points average, with high variance across domains and configurations); (2) self-generated Skills provide negligible or negative benefit (\u20131.3pp average), demonstrating that effective Skills require human- curated domain expertise; (3) less is more\u2014focused Skills with 2\u20133 modules outperform comprehensive documentation; and (4) Skills can partially substitute for model scale, enabling smaller models to match larger ones on procedural tasks.\u300d\u304c\u8208\u5473\u6df1\u304f\u3001(2)\u3082\u305d\u3046\u3060\u308d\u3046\u3068\u601d\u308f\u306a\u304f\u306f\u306a\u3044\u304c\u3001\u4eca\u5f8c\u306e\u5de5\u592b\u3067\u3069\u3046\u306b\u304b\u306a\u308b\u306e\u304b\u304c\u6c17\u306b\u306a\u308b\u3068\u3053\u308d\u3002<\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[717,517],"class_list":["post-8267","post","type-post","status-publish","format-standard","hentry","category-arxiv","tag-skills","tag-517"],"_links":{"self":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8267","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=8267"}],"version-history":[{"count":1,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8267\/revisions"}],"predecessor-version":[{"id":8268,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=\/wp\/v2\/posts\/8267\/revisions\/8268"}],"wp:attachment":[{"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=8267"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=8267"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devneko.jp\/wordpress\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=8267"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}