AI
Progressive disclosure
Loading information into context in stages rather than all at once - metadata always, instructions on trigger, resources on demand - so an agent pays the token cost only for what a task actually uses.
The design principle behind Agent Skills: a SKILL.md's frontmatter (~100 tokens) is always present, its body is read only when the description matches the task, and bundled files are reached only when the instructions point at them. It is what lets a hundred installed skills coexist without a hundred skills' worth of prompt.