Google Expands Android Developer Tooling with CLI Device Streaming and New AI Skills

The landscape of Android development is evolving rapidly, driven by an expanding ecosystem of agents, large language models, tools, and command-line interfaces. In response to these shifts, Google has announced major updates to its developer tooling designed to streamline how high-quality applications are built, regardless of the preferred development environment. The latest updates center on enhancements to the Android CLI command-line tooling—notably the integration of remote physical hardware access through Android Device Streaming—alongside an expanding repository of structured Android skills aimed at grounding artificial intelligence agents with up-to-date, platform-specific standards.

As developers increasingly rely on AI-powered tools and automated workflows, bridging the gap between an LLM’s static training data and the fast-moving reality of platform development has become a critical challenge. Google’s latest releases are engineered to address this friction directly by giving agents secure, direct access to the environment they need to create, test, and manage Android projects efficiently from the terminal.

Android CLI for Command-Line Development

At the core of these updates is the Android CLI, a command-line interface tool engineered to simplify and accelerate Android development workflows. Built to support virtually any AI agent or external tool, the CLI functions as a central bridge for managing the entire development lifecycle. Working in tandem with the dedicated android-cli skill, modern coding agents can leverage the CLI to initialize projects, build applications, run comprehensive test suites, and handle environment setups with minimal human intervention.

Furthermore, the tooling empowers agents to create and run local emulators and execute rigorous test runs directly from the command line. This capability shifts traditional workflows toward a more automated paradigm, allowing developers to delegate repetitive and mechanical tasks to their preferred AI assistants while retaining high-level creative and architectural control over the application.

Android Device Streaming Now Available in CLI

One of the most persistent hurdles in mobile application development is ensuring seamless performance across a fragmented hardware landscape. While emulators are invaluable for early-stage development, catching hardware- and operating-system-specific issues ultimately requires testing on real physical devices. However, developers and their automated agents frequently lack physical access to every target device needed for thorough validation.

To solve this, Google has integrated Android Device Streaming directly into the Android CLI, bringing remote physical hardware access straight to the terminal. Through this integration, an AI agent can connect to real physical devices remotely anywhere it operates. The agent interacts with these physical devices as if they were physically plugged into the development machine via USB, communicating through a secure ADB over SSL connection.

This secure connection unlocks powerful remote capabilities for automated workflows. Agents can now spin up physical devices, deploy fresh builds, collect detailed logs and traces, and even capture screenshots headlessly—all from a terminal window. To begin utilizing the feature, developers simply need to link their project, instruct their agent to list available remote devices, and select the target hardware required for the task. This integration bridges the gap between automated code generation and genuine hardware validation, ensuring that AI-driven development does not sacrifice real-world testing fidelity.

Grounding Agents with Android Skills

While command-line tools provide the operational hands for AI agents, structured knowledge provides the direction. To bridge the gap between the default training of large language models and official platform standards, Google continues to expand its comprehensive Android skills repository.

Android skills are structured instructions encapsulated within standardized SKILL.md files. These files ground artificial intelligence agents using official guidance sourced directly from the official Android developer documentation. Rather than relying on a model’s static training cutoff date, these skills inject precise, fresh data, accurate API references, up-to-date code samples, and verified architectural patterns directly into the agent’s context window.

With more than twenty distinct skills currently available in the repository, developers can equip their AI agents to tackle increasingly complex and specialized tasks. The scope of these skills ranges from foundational framework setups to intricate UI paradigms. Because these skills are thoroughly evaluated and tested against strict criteria, they help mitigate common hallucinations or outdated coding practices often generated by raw models.

The architecture behind these skills is intentionally environment-agnostic. Whether a developer is writing code inside Android Studio, utilizing specialized environments like Antigravity, or pairing with third-party coding agents such as Claude and Codex, the Android skills function consistently across the entire setup. Managing these skills across different projects and individual agent directories is kept straightforward through the native capabilities of the Android CLI, allowing teams to maintain uniform coding standards regardless of their chosen editor or AI assistant.

Skill Spotlight: Wear Compose Material 3

Building applications for Wear OS presents unique design and engineering challenges that diverge significantly from standard smartphone development. Developers must account for round viewports, rotary input mechanics, ambient display modes, strict power consumption limits, specialized layout containers like TransformingLazyColumn, and dedicated structural components such as AppScaffold and ScreenScaffolds.

Without explicit, platform-specific guidance, standard large language models frequently lack the nuanced understanding required to implement these distinct patterns correctly, often resulting in suboptimal user experiences or broken layouts. To combat this limitation, Google introduced the specialized Wear Compose Material 3 skill, designated under the identifier wear/wear-compose-m3.

Early adopters across the industry have already reported significant productivity gains and dramatic codebase simplifications after integrating the new skill into their workflows. For instance, the engineering team at FotMob utilized the Wear Compose Material 3 skill to tackle a massive modernization effort for their existing Wear M3-based application. Their migration tasks included shifting multiple complex lists to TransformingLazyColumn structures, applying proper ScreenScaffold content padding, incorporating ListHeader titles, implementing SurfaceTransformation on interactive cards and buttons, aligning theme typography, and generating accurate Wear previews.

The results of this AI-assisted migration were striking. The generated code compiled successfully on the first attempt and was thoroughly verified on emulators for scrolling performance, rotary input handling, edge morphing, and right-to-left layout support. Crucially, the integration allowed the team to eliminate their legacy wrapper code entirely, alongside a mountain of custom rotary and focus boilerplate.

Furthermore, the skill successfully caught subtle architectural mistakes that the underlying AI model had missed on its own. These included failing to forward the ScreenScaffold content padding correctly into the list components and inadvertently relying on hardcoded text sizes instead of standardized theme typography. Reflecting on the impact of the tool, Roy Solberg, Android Tech Lead at FotMob, noted that a single skill deployed over the course of a single afternoon successfully enabled the migration of eight distinct lists while wiping out an accumulation of custom rotary code.

As Google continues to refine both its command-line toolchain and its growing ecosystem of verified AI skills, the focus remains firmly on empowering developers to build reliable, high-performance applications efficiently. By combining remote hardware access through Android Device Streaming with rigorously evaluated, context-aware AI guidance, the developer tooling ecosystem is adapting to meet the sophisticated demands of modern software engineering.