Metraipolis › archive › dc_ca6dd05d4243

SkillGym: Internalizing Human Skills into LLMs for Real-World Problem Solving — arXiv:2609.27717

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.27717
License (as archived)
CC0-1.0
Archived
2026-10-08T10:05:05.349Z
SHA-256 of the archived text
c34c9bb776fa67a06a55abd02306df00d04b4c5d234d9c7d881f21b508c0464c
Stable URL of this page
https://metraipolis.com/documento/dc_ca6dd05d4243

How to cite

SkillGym: Internalizing Human Skills into LLMs for Real-World Problem Solving — arXiv:2609.27717 — arXiv (en), https://arxiv.org/abs/2609.27717 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ca6dd05d4243 (Metraipolis archive dc_ca6dd05d4243, sha256 c34c9bb776fa67a06a55abd02306df00d04b4c5d234d9c7d881f21b508c0464c)
BibTeX
@misc{dcca6dd05d4243,
  title = {SkillGym: Internalizing Human Skills into LLMs for Real-World Problem Solving — arXiv:2609.27717},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.27717},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ca6dd05d4243; sha256 c34c9bb776fa67a06a55abd02306df00d04b4c5d234d9c7d881f21b508c0464c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.27717. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Human-written agent skills encode rich workflows for real-world problem solving, but are typically used as external inference-time instructions rather than internalized as reusable model capabilities. We introduce \texttt{SkillGym}, a framework that transforms these skills into executable, verifiable training environments for large language model agents. Its skill-to-task pipeline instantiates concrete tasks, verifies outcomes with code-based checkers, and assesses empirical skill dependence through contrastive executions. We construct and release 2,756 environments across 12 categories and collect 8,364 successful trajectories from multiple models and harnesses, averaging 49 tool calls and over 60k logged text tokens. These resources support supervised fine-tuning on verified workflows and reinforcement learning with outcome-based rewards. Under Claude Code, supervised fine-tuning improves Qwen3.5-35B-A3B by 199 Elo on GDPval-AA v2, 19.10 percentage points on Terminal-Bench 2.1, and 28.13 and 12.38 points on SkillsBench v1.1 with and without skills, respectively. Our 35B \texttt{SkillGym-Agent} reaches 51.47\% on skill-assisted SkillsBench, exceeding reported scores for Claude Sonnet 4.6, GPT-5.4 Mini, and DeepSeek V4 Pro. Without skills, it also surpasses skill-assisted bases under Codex and Claude Code, suggesting reusable procedural competence.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-24 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.27717 — 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 2026-10-08T09:11:49.632Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-24?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).