Metraipolis › archive › dc_a29a07559b68

SkillCycle: Co-Evolving Agent Policies and Skill Banks — arXiv:2610.09430

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.09430
License (as archived)
CC0-1.0
Archived
2026-10-08T12:47:16.417Z
SHA-256 of the archived text
238456bc2942b175b2ee4d2b6de5ee501cca32ad7d96e34c0b0b6d6e7a4554ef
Stable URL of this page
https://metraipolis.com/documento/dc_a29a07559b68

How to cite

SkillCycle: Co-Evolving Agent Policies and Skill Banks — arXiv:2610.09430 — arXiv (en), https://arxiv.org/abs/2610.09430 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a29a07559b68 (Metraipolis archive dc_a29a07559b68, sha256 238456bc2942b175b2ee4d2b6de5ee501cca32ad7d96e34c0b0b6d6e7a4554ef)
BibTeX
@misc{dca29a07559b68,
  title = {SkillCycle: Co-Evolving Agent Policies and Skill Banks — arXiv:2610.09430},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.09430},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a29a07559b68; sha256 238456bc2942b175b2ee4d2b6de5ee501cca32ad7d96e34c0b0b6d6e7a4554ef},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09430. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Internalizing external skills changes a language agent's capabilities and, with them, the value of its remaining guidance: rules can become redundant, misleading, or insufficient for newly encountered decisions. This creates a coupled problem of learning from skills and adapting the skills that supervise further learning. We introduce SkillCycle, a framework for co-evolving agent policies and skill banks through a feedback loop between skill internalization and rule revision. Our central contribution is to give distillation feedback a second role: token-level contextual differences help locate rules for inspection, while interaction outcomes guide edits to their content and applicability. SkillCycle alternates between two phases: policy learning with a fixed skill bank and router, and rule revision with a frozen policy. Candidate edits undergo rule-level and whole-bank environment comparisons before they guide the next learning cycle. On WebShop, SkillCycle with a 3B model achieves a success rate of 74.74% and a score of 88.37 without inference-time skill inputs, representing relative improvements of 0.73% and 3.96% over the state-of-the-art (SOTA) model, respectively. In Cycle 3 ablations on ALFWorld and WebShop, SkillCycle's no-skill success rates improve by 10.18% and 18.11% relative to a static skill bank, and by 2.41% and 2.50% relative to a single bank update, respectively. These results show that continually revising skill guidance as the agent's capabilities change helps transform external skills into policy capabilities that require no skill inputs at inference. We will release code, configurations, skill banks, and evaluation protocols.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.09430 — 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 2026-10-08T12:18:15.360Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).