SkillEvo: Self-Renewing Evolution Gradients from Multi-Turn Interaction Feedback

2026-08-13

AI Summary

Current agent skills often stagnate in evolution due to their reliance on single-turn feedback, which fails to expose deeper deficiencies. This work proposes a continual evolution framework grounded in multi-turn user simulation, which hierarchically reveals skill flaws and constructs trustworthy evolutionary gradients. An independent governance mechanism is introduced to proactively correct factual errors and structural redundancies, rather than merely rejecting degraded candidates. The approach achieves the first effective translation of multi-turn interactive feedback into optimizable gradients, significantly outperforming self-reflection methods (+23.0 points) and single-turn QA-driven approaches (+15.4 points) across six cloud service categories, nine production skills, and 98 reference documents. By overcoming the limitations of single-turn evaluation, this framework enables sustainable self-evolution of agent capabilities.

Abstract

Agent Skills are today either hand-authored or produced in a single LLM generation pass, and consequently possess no closed loop through which they might improve from the interaction failures they actually cause. Recent work does close this loop, but derives its feedback from single-turn question-answering evaluation. The consequence is a sharp asymmetry: once the first round has patched the gaps that a single exchange can reveal, the evolution gradient decays, the defects that surface only across multiple turns remain invisible, and evolution stalls. Governance in these systems is likewise driven by an end-to-end verification score, a scalar gate that can reject a degraded candidate but can neither localize nor repair its structural cause. We argue that the binding constraint on sustained skill evolution is neither editing capability nor the number of iterations, but whether the evaluation feedback keeps supplying trustworthy evolution gradients. We introduce SkillEvo, in which trustworthy feedback generates the gradient and controllable governance constrains its direction. The first component recasts multi-turn user simulation from an evaluation endpoint into a feedback generator: follow-up questions expose defects layer by layer, so that every round of revision both consumes feedback and produces new feedback. The second replaces the passive rejection of a scalar gate with an independent governance layer that actively repairs factual degradation and structural bloat, preventing the gradient from drifting as degradation accumulates. Across six categories of cloud services, 9 production Skills, and 98 skill-reference files, SkillEvo surpasses self-reflection-based evolution by 23.0 points and single- turn-QA-driven evolution by 15.4 points.

Problem

• skill evolution
• multi-turn interaction
• feedback gradient
• agent skills
• governance

Innovation

• multi-turn interaction
• evolution gradients
• feedback generation
• controllable governance
• skill evolution