AI coding agents operate directly on users' filesystems and regularly corrupt data, delete files, and leak secrets. We conduct the first systematic study of agent filesystem misuse, analyzing 290~public reports. Our study reveals two fundamental gaps: users and agents have limited information about filesystem effects and insufficient control over them. To close these gaps, we propose shifting information and control from agents to filesystems. We introduce agent-native filesystems and identify three primitives they should provide: introspect effects, undo mutations, and gate accesses. Together, these primitives enable agents to work autonomously and correct their own mistakes, while reserving user interaction for sensitive accesses and final review. We build YoloFS, an agent-native filesystem, with three mechanisms: staging isolates mutations until the user commits them; snapshots let agents restore earlier states; and progressive permission lets users enforce and adapt rules during execution. To evaluate YoloFS, we introduce a new methodology that captures the interactions among user, agent, and filesystem. On 11 tasks with hidden side effects, YoloFS enables agents to self-correct in 8 and keeps all mutations staged for user review. On 112 routine tasks, YoloFS reduces user interactions and matches the baseline success rate. YoloFS is open-sourced at https://github.com/YoloFS/YoloFS.

翻译：暂无翻译