arXiv：cs.CL· Serin Kim, Kwangwook Seo, Dokyung Song, Jinyoung Yeo, Dongha Lee·· 3 小时前AI 评分36
SkillSandbox：通过动态场景合成验证智能体技能
SkillSandbox: Skill Verification via Dynamic Scenario Synthesis
AI 导读
SkillSandbox 是一个为每个技能动态合成技能相关但新颖的任务与环境、以验证其可复用性的框架，由 Proposer、Builder 和 Verifier 三部分组成，Verifier 从可执行性、实用性和效率评估后给出 Keep 或 Reject 判定。在 ALFWorld 和 WebShop 上使用三个模型测试，SkillSandbox 持续取得最强下游性能并提升执行效率。
正文
Abstract:Self-evolving agents distill task-solving experience into skills for future reuse, but these skills can encode incorrect procedures or non-transferable knowledge. It is therefore critical to verify each skill's reusability: whether its guidance remains useful beyond the experience from which it was distilled. Such verification requires observing how a skill affects execution in new tasks, yet existing tasks may not expose the situations where the target skill can actually be exercised. To construct such situations, we propose SkillSandbox, a framework that dynamically synthesizes a task and its environment for each skill that are skill-relevant yet novel. A Proposer specifies the conditions to preserve and the source-specific details to vary, a Builder constructs an executable scenario, and a Verifier compares executions with and without the skill. The Verifier assesses executability, utility, and efficiency to assign a Keep or Reject verdict, determining whether the skill enters the library. Across ALFWorld and WebShop with three models, SkillSandbox consistently yields the strongest downstream performance and improved execution efficiency. Further analyses examine whether these gains reflect accurate assessment of skill reusability and identify which components of SkillSandbox contribute to them.
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.10088 [cs.AI] |
| (or arXiv:2610.10088v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10088 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Serin Kim [view email]
[v1]
Wed, 7 Oct 2026 13:48:58 UTC (2,946 KB)
来源：arXiv：cs.CL · arxiv.org