AI RESEARCH
You Live More Than Once: Towards Hierarchical Skill Meta-Evolving
arXiv CS.AI
•
ArXi:2605.28390v1 Announce Type: new Test-time skill evolving is regarded as a new paradigm for enhancing deployed agentic systems. Existing works mainly focus on hard-coded skill evolving strategies or parametric learning that rely on expensive parameter updates in the underlying LLMs. In this paper, we nstrate that test-time refinement of the skill evolving framework itself is necessary for continuous improvement of the agent systems in different downstream scenarios, and lightweight algorithmic adaptation is feasible.