SkillAligner：在执行时将检索到的技能视为可调整的草稿
SkillAligner: Treating Retrieved Skills as Adaptable Drafts at Execution Time
浏览论文内容
中文总结 AI 辅助
SkillAligner是无训练的执行时技能适配框架，将检索技能视为可调整草稿，经联合适配整合为执行指南，在多基准实验中提升任务性能、降低退化与推理成本。
中文摘要 AI 辅助
通用技能可为语言智能体提供可复用的过程性知识，但语义相关性并不能保证执行效用：检索到的技能可能包含与当前任务、执行环境或其他检索到的技能相冲突的假设。我们将此问题形式化为技能-执行失配。为解决该问题，我们提出SkillAligner，这是一个无训练的执行时技能适配框架，它将检索到的技能视为可调整的草稿而非固定指令。在执行前，SkillAligner执行一次联合适配，将有用的技能片段专门适配到任务需求，使其过程假设与可用的执行接口对齐，并通过解决技能间的依赖、冲突和冗余来组合得到的指导。适配后的内容被整合为紧凑的执行指南，并在后续轨迹中重复使用。在不同的智能体基准和模型主干上进行的大量实验表明，SkillAligner相比现有技能使用基线显著提升了任务性能，在实例级别降低了技能诱导的退化，并降低了总推理成本。
英文摘要
General-purpose skills promise reusable procedural knowledge for language agents, yet semantic relevance does not guarantee execution utility: a retrieved skill may encode assumptions that conflict with the current task, execution environment, or other retrieved skills. We formalize this problem as the skill--execution misfit. To address it, we propose SkillAligner, a training-free execution-time skill adaptation framework that treats retrieved skills as adaptable drafts rather than fixed instructions. Before execution, SkillAligner performs a one-time joint adaptation that specializes useful skill fragments to task requirements, aligns their procedural assumptions with the available execution interface, and composes the resulting guidance by resolving dependencies, conflicts, and redundancy across skills. The adapted content is consolidated into a compact execution guide and reused throughout the subsequent trajectory. Extensive experiments across diverse agent benchmarks and model backbones show that SkillAligner substantially improves task performance over existing skill-use baselines, reduces skill-induced regressions at the instance level, and lowers total inference cost.
发表机构
- Zhejiang University（浙江大学）
机构由 AI 辅助整理，请以论文原文为准。
