Agent-Native Apps
//AGENT-NATIVE APPLICATION VOCABULARY

Mount

To mount an agent-native app is to load its published package into a compatible Host so the Host runs with the app's identity, constitution, workflows, and durable state. Mounting is gated: the Host checks the user's signed entitlement and the package's constitution hash before the app runs. Once mounted, the app draws intelligence from the Host while its governed actions, results, and model identity are recorded.
via agentnativeapps.com

Mounting is the third step of the category's core loop: publish, inspect, mount, run. It's deliberately a checked step rather than an install: the point at which a package's claims (who it is, what it may do, whose data it touches) are verified against what the user actually agreed to run.

Unmounting is equally defined. The user can update, diff, roll back, export, or fork within the rules, and state ownership doesn't change when an app is unmounted: the data, corrections, and accumulated context remain the user's.

Follow the vocabulary as it grows.