Android CLI is designed to make command-line Android development easier and support AI agents and other tools working on Android projects. The Android Developers Blog post describing the tools is by Simona Milanovic, a Developer Relations Engineer.
With the android-cli skill, agents can create, build, test and manage Android projects, help configure a development environment, create and run emulators, and execute test runs. The tooling connects project-level work with device access through the command line.
Android Device Streaming provides remote access to physical devices. Through Android CLI, agents can use that service over an ADB-over-SSL connection that the blog describes as working as if the devices were plugged in over USB. Supported tasks include starting devices, deploying builds, collecting logs and traces, and capturing screenshots headlessly from the terminal.
The combination brings physical-device testing into an agent-accessible terminal workflow. The blog describes the connection as secure ADB over SSL; the supplied information does not detail other security controls, so the claim should be read narrowly.
Android skills are structured instructions stored in SKILL.md files and grounded in official guidance from developer.android.com. The blog says the skills bring API references, samples, architectural patterns and fresher information into an agent’s context. More than 20 skills are available; they are described as thoroughly evaluated and designed to work across different environments.
One example is the Wear Compose Material 3 skill, released by the Android team. The FotMob engineering team used it for work that included modernizing its existing Wear M3-based app. The blog says the resulting changes compiled successfully and were checked on an emulator for scrolling, rotary input, edge morphing and right-to-left layout.
Together, device streaming and reusable skills address two parts of agent-assisted Android development: access to real hardware and guidance tailored to Android APIs and patterns. The available information describes Google’s tooling and one developer example, but does not provide a basis for comparing it with competing platforms or judging wider market adoption.
