An Empirical Study of Downstream Adaptation for Agent Skills
2026-07-03 • Software Engineering
Software Engineering
AI summaryⓘ
The authors studied how developers change and adapt "skills"—pre-made parts of AI software—to fit their own needs. They looked at over a thousand examples and found that even though these skills are meant to be reused easily, people often have to rewrite a lot and fix problems to make them work properly. The authors also noticed that changes in one part of a skill often require fixing other parts too, and many adaptations added new security risks. Their work suggests that better tools and standards could help make adapting these skills easier and safer.
Large Language Modelssoftware reuseskillsforksadaptation patternssoftware abstractioninterface standardssecurity risksautomationsoftware orchestration
Authors
Xinjian Wu, Jingzhi Gong, Gunel Jahangirova, Zhenpeng Chen, Jie M. Zhang
Abstract
As Large Language Model (LLM) agents become integral to modern software systems, ``skills'' have emerged as a novel unit of software reuse, enabling developers to package workflows, decision procedures, and prompt-based policies. While skills are intended for reuse, downstream developers frequently modify published skills to fit local contexts, yet little is known about the nature of such adaptations. This paper presents the first empirical study of downstream skill adaptation in public forks, to understand how published skills are adapted, and to provide implications for researchers and engineers on improving skill design, evolution, and orchestration. Specifically, we analyze 1,126 skill-adaptation instances from six widely adopted skill repositories and develop a taxonomy comprising 46 adaptation patterns organized into 13 families. Our key findings reveal a reuse paradox: although skills are intended to be easily imported and reused, developers spend a lot of effort rewriting what the skills do, fixing skill discoverability, and translating them for different tools and languages, indicating a need for better abstractions, standardized interfaces, and automated support for skill adaptation. Furthermore, adaptations are highly interdependent, with changes in one component often requiring coordinated updates elsewhere, motivating automated support for detecting inconsistent modifications. We also find that nearly one-fifth of adaptations introduce security-sensitive content within the same instruction text that governs behavior.