Public Preview: AI Toolkit for GitHub Copilot Brings Prompt-First Agent Development to VS Code
Lee_Stott presents the public preview of the AI Toolkit for Visual Studio Code, unveiling prompt-first agent development powered by GitHub Copilot and Microsoft Agent Framework, designed for fast and seamless AI agent creation in VS Code.
Public Preview: AI Toolkit for GitHub Copilot Brings Prompt-First Agent Development to VS Code
Author: Lee_Stott
Overview
The AI Toolkit for Visual Studio Code now introduces prompt-first agent development with GitHub Copilot, aiming to simplify and accelerate the creation of AI agents. By embedding agent workflows directly into VS Code, developers can transform natural language prompts into robust agent code, leverage model recommendations, and streamline production-readiness.
Key Features
1. Model Selection and Recommendations
- Model-wise agent building: Easily browse, select, and experiment with both local and cloud-based AI models directly from VS Code.
- Smart recommendations: GitHub Copilot analyses your agent scenarios to suggest suitable models, reducing guesswork and helping you quickly make technical choices.
2. Build & Orchestrate Agents with Microsoft Agent Framework
- Prompt-first agent creation: Initiate new agents (single or multi-agent) simply by describing your goals in natural language.
- Orchestration: Leverage the Microsoft Agent Framework for agent lifecycle management and workflow visualization.
- Multi-agent workflows: Design and visualize collaborative agent systems for complex requirements.
3. Trace Agent Behaviors
- Integrated tracing: Collect localized traces and inspect detailed agent-call actions, all in the context of your VS Code workspace.
- Debugging and transparency: Gain clear insights into agent decision processes and execution flows for troubleshooting and optimization.
4. Evaluate Agent Response Quality
- Guided evaluation: Copilot offers structured evaluation approaches, metric suggestions, and test data generation to help you assess agent performance.
- CI/CD integration: Bring evaluation steps into your build and deployment pipelines to ensure ongoing quality assurance as agents evolve.
How to Get Started
- Install the AI Toolkit for Visual Studio Code.
- Start new agent projects or integrate agent development into existing projects directly in VS Code with Copilot assistance.
- Share feedback or request features on the GitHub repo.
Release Details
- Version: 4.0 (Public Preview)
- Announced at GitHub Universe, October 29, 2025
Additional Resources
Try the new toolkit, streamline your agent development, and help shape its future with your feedback!
This post appeared first on “Microsoft Tech Community”. Read the entire article here