Arm and GitHub bring Copilot CLI agents into native Arm development

What the workflow offers

In a blog post by Shantu Roy, VP of Developer Relations and Customer Engagement, Arm describes a workflow that joins GitHub Copilot, native Arm compute, and GitHub's review and validation tools. The idea is to keep development tasks on the target architecture, which Arm says can give developers earlier feedback on dependencies, test results and runtime behavior.

Arm lists three parts of the setup:

• Work from your preferred device: GitHub Copilot CLI, a command-line coding agent, runs on organization-controlled Arm compute. Remote Control lets developers steer sessions from another device while commands keep executing natively on Arm.
• Develop and iterate natively: Developers use Copilot CLI to edit code, build, test and investigate failures in the Arm-native environment, reducing the need to switch between architectures.
• Review and validate changes: Pull requests go through GitHub for code review, while CI on Arm64 runners automates native builds and tests.

Arm says the approach targets developers working on Arm-based PCs, cloud infrastructure and connected devices.

Why native execution matters

Arm argues that teams often write and compile code in an x86 environment, submit changes for review, and only later test on Arm. That can delay the discovery of architecture-specific issues. Software that builds on x86 may hit incompatible dependencies, architecture-specific assumptions or differences in runtime behavior on Arm, and may miss platform-specific capabilities and performance characteristics.

Arm names three benefits of running Copilot CLI on native Arm compute:

• Optimization with real measurements: Developers can measure CPU utilization, execution time, latency, memory behavior and throughput on Arm hardware to find bottlenecks. For independent software vendors, Arm says this creates a chance to build products designed for Arm from the outset.
• Earlier detection of issues: Native builds and tests can expose problems with dependencies, compiler settings, architecture-specific code and runtime behavior. GitHub Actions adds automated builds and tests on Arm64 runners for repeatable validation.
• Faster iteration: Developers can modify code, run project commands, run tests and investigate failures in one environment, then feed results into further changes.

Arm also notes that developers remain responsible for reviewing agent-generated changes and validating their correctness.

How the pieces fit together

Arm assigns each technology a role. Arm provides the compute architecture and native execution environment. GitHub provides source control, pull requests, code review and automated validation, with GitHub Actions supporting Arm64 runners. GitHub Copilot CLI provides AI-assisted development on native Arm compute, directing agents to make code changes, run commands and investigate test failures in the target environment.

The post gives an example of a developer updating a media application. The developer asks Copilot CLI to modify a dependency and run the project's tests. Because the session runs on Arm compute, the commands execute natively, which Arm says helps expose dependency issues and test failures early. The developer then submits a pull request, and a GitHub Actions workflow validates it on an Arm64 runner as a separate, repeatable stage. After review and successful validation, the application can be deployed to its intended Arm environment, where developers can profile runtime behavior and use the measurements to guide optimization.

Arm adds that keeping tasks on organization-controlled Arm compute lets teams apply their own security policies, access controls and tooling.

What it means for developers

Arm frames the goal as going beyond checking whether software works on Arm. According to the company, native feedback and real measurements can help developers and ISVs build software that takes full advantage of the Arm platform from the start. The post points readers to Arm resources on running CI/CD workflows on Arm and building and testing with GitHub Actions runners. The source gives no pricing, availability dates or performance figures.