Metraipolis › archive › dc_4ecf16f011bf
SkillShift-Bench: Benchmarking Skill Learning under Environment Evolution for Web Agents — arXiv:2610.33528
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.33528
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-27T07:56:37.973Z
- SHA-256 of the archived text
eef846dcd6a1177d0b1e93da4ab6c3c1ef019046bbaa8ce466b2fd0681880647- Stable URL of this page
https://metraipolis.com/documento/dc_4ecf16f011bf
How to cite
SkillShift-Bench: Benchmarking Skill Learning under Environment Evolution for Web Agents — arXiv:2610.33528 — arXiv (en), https://arxiv.org/abs/2610.33528 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4ecf16f011bf (Metraipolis archive dc_4ecf16f011bf, sha256 eef846dcd6a1177d0b1e93da4ab6c3c1ef019046bbaa8ce466b2fd0681880647)
BibTeX
@misc{dc4ecf16f011bf,
title = {SkillShift-Bench: Benchmarking Skill Learning under Environment Evolution for Web Agents — arXiv:2610.33528},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.33528},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4ecf16f011bf; sha256 eef846dcd6a1177d0b1e93da4ab6c3c1ef019046bbaa8ce466b2fd0681880647},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.33528. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Skill learning enables web agents to reuse prior experience for long-horizon interactive tasks. However, learned skills are typically evaluated in closed environments, overlooking the fact that real-world web deployment involves continuous shifts in invocation context, perceptual grounding, and execution feasibility. To this end, we introduce SkillShift-Bench, the first benchmark for evaluating skill learning with controlled open-environment shifts. SkillShift-Bench simulates Contextual Shift, Perceptual Shift, and Execution Shift across three self-hosted web systems, and evaluates agents with diagnostics covering task success, interaction cost, skill use, and skill failure. Comprehensive experiments reveal systematic robustness degradation under controlled shifts, characterized by lower success rates, higher interaction costs, and more frequent skill failures, with Execution Shift exerting the strongest pressure. We further formulate shifted skill execution as a nonstationary local executability problem and provide both a dynamic-regret analysis and a practical test-time adaptation method, demonstrating the feasibility of improving learned-skill execution without relearning the skill library. Code, tasks, environments, documentation, and the leaderboard are available at this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.33528 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-27T07:55:30.642Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).