SkillOpt: Agent skills as trainable parameters

Summary

SkillOpt treats an agent's skill file as a trainable parameter outside a frozen target model, converting skill writing from one-shot prompting into a controlled optimization process. Across six benchmarks, seven target models, and three execution modes, SkillOpt achieves best or tied-best performance in all 52 evaluation cells, improving performance without updating model weights. The method keeps skills compact and auditable through bounded text edits, validation gating, rejected-edit feedback, and slow/meta updates, preventing uncontrolled prompt drift. Optimized skills transfer across model scales, agent harnesses, and related tasks, suggesting they capture reusable workflow knowledge rather than benchmark-specific instructions. This addresses the problem of manual skill editing that lacks guarantees of improvement.