Independently authored skills can impose incompatible requirements, competing preferences, and redundant instructions when jointly loaded by an LLM agent. We propose SkillFusion, an offline compiler that reconciles these interactions into a reusable merged skill. Its central idea is to make composition decisions explicit through typed, scoped constraints linked to their source directives. A type-scope gate selects candidate interactions, while selective fallback revisits pairs potentially missed by structural pruning. A classifier identifies interaction types, and an adversarial verifier re-examines uncertain judgments. The resolver applies conditional routing or explicit priorities to hard conflicts, verified abstraction to soft conflicts, and deduplication to redundant instructions. Source-linked checks guide synthesis and compression, and metadata synthesis supports discovery of the merged skill. The resulting artifact retains associated scripts and resources and can be reused across tasks and executor models. SkillFusion achieves the highest mean Success among evaluated non-oracle methods on all three agent benchmarks and on both transfer benchmarks across three executor models. At the largest tested instruction configurations, it achieves over 2.5 times the Success of unresolved multi-skill loading, highlighting the growing value of reconciliation as composition becomes more demanding. The code will be open-sourced after the review process concludes.
SkillFuzz: Fuzzing Skill Composition for Implicit Intents Discovery in Open Skill Marketplaces
School of Computer Science and Informatics, University of Liverpool(利物浦大学计算机科学与信息学学院);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
Towards Compositional Generalization of LLMs via Skill Taxonomy Guided Data Synthesis
State Key Laboratory of Complex & Critical Software Environment, Beihang University(复杂与关键软件环境国家重点实验室,北京航空航天大学);Beijing Academy of Artificial Intelligence(北京人工智能研究院);Beijing Institute of Technology(北京理工大学);Institute of Computing Technology, CAS(中国科学院计算技术研究所)
Shattered Compositionality: Counterintuitive Learning Dynamics of Transformers for Arithmetic
University of Wisconsin–Madison(威斯康星大学麦迪逊分校);Department of Computer Sciences, University of Wisconsin–Madison(威斯康星大学麦迪逊分校计算机科学系);Department of Statistics, University of Wisconsin–Madison(威斯康星大学麦迪逊分校统计学系)
LEGO: An LLM Skill-Based Front-End Design Generation Platform
School of IC, Peking University(北京大学集成电路学院);School of EECS, Peking University(北京大学电子信息技术学院);School of Microelectronics, Xidian University(西安电子科技大学微电子学院);Institute of EDA, Peking University(北京大学EDA研究院);Beijing Advanced Innovation Center for IC(北京集成电路先进创新中心)
DIRO & Institut Courtois, Université de Montréal(DIRO与Courtois研究所,蒙特利尔大学);Mila – Québec AI Institute(魁北克人工智能研究所);Microsoft Research(微软研究院);Canada CIFAR AI Chair(加拿大CIFAR人工智能主席)
Compositional Generalization from Learned Skills via CoT Training: A Theoretical and Structural Analysis for Reasoning
Gaoling School of Artificial Intelligence, Renmin University of China(中国人民大学人工智能学院);Beijing Key Laboratory of Research on Large Models and Intelligent Governance(北京大型模型与智能治理研究重点实验室);Engineering Research Center of Next-Generation Intelligent Search and Recommendation, MOE(下一代智能搜索与推荐工程研究中心,教育部);Tianjin University of Science and Technology(天津科技大学);Beijing Institute of Technology(北京理工大学)
Agentic Proposing: Enhancing Large Language Model Reasoning via Compositional Skill Synthesis
AI DATA, Alibaba Group Holding Limited(阿里数据,阿里巴巴集团控股有限公司);EPIC Lab, Shanghai Jiao Tong University(上海交通大学EPIC实验室);Shanghai University of Finance(上海财经大学);Wuhan University(武汉大学)
SkillEvolBench: Benchmarking the Evolution from Episodic Experience to Procedural Skills
The Ohio State University(俄亥俄州立大学);The University of Chicago(芝加哥大学);University College London(伦敦大学学院);University of Michigan(密歇根大学);The Chinese University of Hong Kong(香港中文大学);Case Western Reserve University(凯斯西储大学);Amazon(亚马逊)
Teaching Algorithmic Reasoning via In-context Learning
Mila(米拉研究所);Université de Montreal(蒙特利尔大学);Google Research(谷歌研究院)
大语言模型(LLM)通过扩展模型和数据大小,显示出越来越多的上下文学习能力。尽管取得了这些进展,LLM仍然无法解决算法推理问题。虽然为最终答案提供了一个基本原理,导致了多步推理问题的进一步改进,但Anil et al. 2022表明,即使是简单的算法推理任务,如奇偶校验,也远未解决。在这项工作中,我们确定并研究了成功教授算法推理的四个关键阶段:(1)将算法制定为技能,(2)同时教授多种技能(技能积累),(3)教授如何组合技能(技能组合)和(4)教授如何使用技能作为工具。我们表明,它是可以教算法推理LLMs通过上下文学习,我们称之为算法提示。我们评估我们的方法在各种算术和定量推理任务,并表现出显着的提升性能比现有的提示技术。特别是,对于长奇偶校验、加法、乘法和减法,与最佳可用基线相比,我们分别实现了约10倍、9倍、5倍和2倍的误差减少。
When Safe Skills Collide: Measuring Compositional Risk in Agent Skill Ecosystems
Carnegie Mellon University(卡内基梅隆大学);Georgia Institute of Technology(佐治亚理工学院);University of Glasgow(格拉斯哥大学);Independent Researcher(独立研究员);Corespeed Inc.(Corespeed公司)
Skill Expansion and Composition in Parameter Space
National University of Defense Technology(国防科技大学);Tsinghua University(清华大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室);Beijing Academy of Artificial Intelligence(北京人工智能研究院)
CLASP: Language-Driven Robot Skill Selection and Composition using Task-Parameterized Learning
German Aerospace Center (DLR), Institute of Robotics and Mechatronics (RMC)(德国航空航天中心(DLR),机器人与机电一体化研究所(RMC));Technical University of Munich (TUM)(慕尼黑工业大学(TUM))
RoboCoach: World Models as Active Coaches for Compositional Robot Skills
Renmin University of China(中国人民大学);Tsinghua University(清华大学);University of Nottingham(诺丁汉大学);Beijing Academy of Artificial Intelligence(北京人工智能研究院);Shanghai Qizhi Institute(上海期智研究院)