Coding assistants powered by AI are revolutionizing software development through the automation of coding, code reviewing, and project management, while at the same time allowing for coordination between multiple repositories.
The GitHub team has developed a suite of AI software development tools that are aimed at automating SDLC processes including planning, writing code, testing, and code review. These products include GitHub Copilot app, Copilot Fusion, Copilot Projects, Copilot Automations and Copilot Code Review Max. These tools should assist developers to orchestrate different AI agents, run projects and increase the efficiency of development process with control over implementation.
GitHub Copilot app creates a developer environment for working with AI agents and integration of them with the development tools. This tool allows performing such operations as specification creation, design participation, telemetry analysis and development of code. Copilot Fusion brings multi-model orchestration to the product, which allows analysing tasks for their reasonability, code generation, debugging and required tools before selection of appropriate AI models.
Some of the important characteristics include:
- Orchestration of various AI models.
- Parallel execution of agents across multiple repositories.
- Agent sessions on the cloud.
- AI automation through events and scheduling.
- Severe software defects detection.
With Copilot Projects, developers can set goals for AI agents that operate in multiple repositories using parallel, managed sessions. Developers can monitor the progress and then take decisions and control the next course of actions from one dashboard. Session persistence also ensures that the user can begin and manage the same cloud-based agent session from the GitHub application, Visual Studio Code (VS Code), CLI and the Web platform without having to restart their session.
When it comes to workflow automation, Copilot Automations provides AI-driven automation that can be initiated manually, at scheduled times, and triggered by repository events. Such workflows will be able to orchestrate AI agents for tasks such as triaging issues and reviewing code. The platform can even analyze the completed results and provide recommendations for future automations, pending the approval of the developer.
Copilot Code Review Max is yet another feature of GitHub that will be used to detect critical software defects, especially high severity bugs. While Lite and Balanced modes will handle review processes, Max will conduct more thorough analysis of the code changes. The features come with no additional costs to GitHub Enterprise customers since there is no need to subscribe to Copilot separately. Users only have to pay for the AI credits that they use.
