插件

人脑式上下文压缩逻辑，大幅减少长任务上下文占用以及 token 消耗

Incremental file mounting with read dedupe: mounted line ranges are never re-sent to the model, on-disk changes invalidate and remount, with a Mounted Files tab and token-savings accounting.

四层省 Token，不牺牲模型智能：handoff、rtk、headroom、caveman 四个省 token 的 skill，每个也是独立可装插件。