Type: url
Original: https://arxiv.org/pdf/2605.07358
Status: generated
Summary
This survey from The Chinese University of Hong Kong, Shenzhen defines agent skills as reusable procedural artifacts that coordinate tools, memory and runtime context under task-specific constraints, formalized as a root instruction document, auxiliary resources and applicability conditions, and argues that they bridge a “procedural gap” between raw tool access and robust task execution. It organizes the literature around a skill lifecycle of representation, acquisition (human-, experience-, task- and corpus-derived), retrieval and selection, and evolution (revision, validation, policy coupling, repository evolution and runtime governance), and closes with open challenges and research directions.