A simple Markdown file is apparently enough to boost GPT-5.5 by more than 20 points on procedural tasks. That's the promise of SkillOpt, a method from Microsoft and three Chinese universities that trains instruction documents for AI agents the same way model weights get trained.
These kinds of instruction documents, known as "skills," are already common in commercial products. Anthropic, for example, added a modular skill system to Claude last year that automatically loads topic-specific instructions, scripts, and resources depending on the task.
Skills typically bundle procedures, tool-use rules, output formats, and known failure patterns, and they've become a standard approach. Until now, according to the Microsoft team's paper, they were either written by hand, generated in a single pass by a language model, or loosely self-revised. None of these approaches behaves like a real optimizer, and none guarantees the skill actually improves.
SkillOpt trains the skill document like model weights, only keeping changes that measurably pay off. | Image: Yang et al.
The skill document becomes a trainable state
