如今的代理技能是手工制作的, 一次性生成的, 或通过松散控制的自我修订, 演变而来，其中没有一个表现得像技能, 的深度学习优化器，也没有一个在反馈下可靠地改进了其起点。我们认为该技能应该作为冻结代理, 的外部状态进行训练，并采用与权重空间优化可重复的相同规则。据我们所知，SkillOpt 是, 第一个用于座席技能的系统可控文本空间优化器: 一个单独的优化器模型将计分推出转化为有界添加/delete/replace 对单个技能文档的编辑,，并且仅当编辑严格提高保留的验证分数时才会接受编辑。文本学习率预算,拒绝编辑缓冲区,和分阶段缓慢/元更新使技能训练稳定，同时在部署时添加零推理时间模型调用。跨六个基准, 七个目标模型, 和三个执行工具(直接聊天, Codex, Claude Code), SkillOpt 在所有 52 个评估的(model, 基准, 工具) 个单元中是最好的或并列，并且击败了人类, 一次性 LLM, Trace2Skill, TextGrad, GEPA, 和 EvoSkill 中的每一个单元竞争对手技能。在 GPT-5.5 上，它将直接聊天, 中的平均无技能准确率提高了 +23.5 点，在 Codex 代理循环, 中提高了 +24.8，在 Claude 代码中提高了 +19.1。转移实验进一步表明，当在 Codex 和 Claude Code 执行环境, 之间跨模型规模, 转移时，优化的技能工件保留了价值，并且无需进一步优化即可转移到附近的数学基准。此 https URL 代码:

Agent skills today are hand-crafted, generated one-shot, or evolved through loosely controlled self-revision, none of which behaves like a deep-learning optimizer for the skill, and none of which reliably improves over its starting point under feedback. We argue the skill should instead be trained as the external state of a frozen agent, with the same discipline that makes weight-space optimization reproducible. SkillOpt is, to our knowledge, the first systematic controllable text-space optimizer for agent skills: a separate optimizer model turns scored rollouts into bounded add/delete/replace edits on a single skill document, and an edit is accepted only when it strictly improves a held-out validation score. A textual learning-rate budget, rejected-edit buffer, and epoch-wise slow/meta update make skill training stable while adding zero inference-time model calls at deployment. Across six benchmarks, seven target models, and three execution harnesses (direct chat, Codex, Claude Code), SkillOpt is best or tied on all 52 evaluated (model, benchmark, harness) cells and beats every per-cell competitor among human, one-shot LLM, Trace2Skill, TextGrad, GEPA, and EvoSkill skills. On GPT-5.5 it lifts the average no-skill accuracy by +23.5 points in direct chat, by +24.8 inside the Codex agentic loop, and by +19.1 inside Claude Code. Transfer experiments further show that optimized skill artifacts retain value when moved across model scales, between Codex and Claude Code execution environments, and to a nearby math benchmark without further optimization. Code: this https URL

科目: 人工智能 (cs.AI); 计算和语言 (cs.CL)

Subjects: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)