cs.AINov 17, 2026

When Agent Skills Go Stale: Runtime Behavior and Selective Maintenance in Evolving Environments

Authors: Ömer Faruk Akgül, Shuli Jiang, Yuting Zhang, Yi Zhang, Xiaohan Fei, Wei Xia, Stefano Soatto

Organizations: University of Southern California · AWS Agentic AI

Abstract

Agent skills use past experience and current observations to guide decisions and actions on related tasks. This experience is retained in skill state, whose reuse can amortize inference-time reasoning and environment interaction. When environment changes, parts of this state can become outdated while others remain useful. Across four task settings and two models, we compare skill reuse before and after environment changes. Previously beneficial skills can reduce task performance by over 10 percentage points or increase execution cost by nearly 40% relative to no skill. Task success can also remain unchanged while execution savings decline. These effects depend on which environment-specific details become outdated and the agent's ability to compensate using current observations. We introduce selective maintenance, which replays checks from successful executions without additional model calls. It verifies whether observed differences conflict with skill state, repairs affected content, and validates updates, rebuilding when local repair is insufficient. No description of the change is supplied. Relative to reusing stale state, maintenance improves task accuracy by up to 67pp and reduces execution steps by up to 37%. These results suggest that preserving a skill's inference time value as environments evolve requires selectively maintaining its state.

Figures & tables

Appendix figures & tables12 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Skill1: Unified Evolution of Skill-Augmented Agents via Reinforcement Learning

    May 7, 2026Yaorui Shi, Yuxin Chen, Zhengxi Lu +6LLM Agent Skill LearningLLM Agent Skill Retrieval

  2. SkillNEO: Non-Elitist Evolutionary Optimization of Agent Skills

    Oct 14, 2026Enze Ma, Chunyu Miao, Yankai Chen +14LLM Agent Skill LearningLLM Agent Self-Improvement

  3. EvoGuard: Mitigating Degradation in Online Skill Evolution

    Nov 4, 2026Yanli Wang, Zichen Bai, Tianyue Jiang +12LLM Agent Skill LearningSelf-Evolving Agents