For years, Visual Studio Code has stood as the undisputed favorite among developers: open source, cross-platform, and endlessly extensible. Its architecture has even inspired a wave of custom forks and specialized editions. But the way we code is evolving. AI is no longer just a helpful sidebar; it’s becoming an agentic partner embedded directly into our workflows. While editors like Cursor and Windsurf have already embraced this shift, VS Code is now stepping up with native AI agent capabilities powered by GitHub Copilot. In this post, we’ll explore how Microsoft is transforming VS Code from a traditional editor into an AI-driven development environment.
Let’s see how we can leverage the new feature
To be able to use it make sure you are on the correct VS Code version :
Once you are on this version, go to the top right corner of the VS Code and you will see Open Agent
When you click for the first time you will see a login screen, you can use Github account to login and you should see the something similar to below screen.
Top Section (Sessions):
- Session History: Shows your ongoing AI agent conversations and tasks (e.g., “Create test script for the userlist LWC component,” “Downloaded metadata”)
- New Session Button: Start fresh agent conversations with
Ctrl+N - This acts as your agent’s memory and task tracker
Bottom Section :
- Agents (3): Configure different AI agent personalities/roles
- Skills (12): Pre-built capabilities the agent can use
- Instructions: Custom guidelines for how agents should behave
- Hooks: Automated triggers for agent actions
- MCP Servers (2): Model Context Protocol servers for extended functionality
- Plugins: Additional agent extensions
Center – Active Agent Workspace
- Current Session: “New session in ITSMOrg with Copilot CLI”
- Input Box: Natural language interface (“What’s next on your roadmap?”)
- Model Selection: Choose between different AI models (Agent, Claude Haiku, etc.)
- Approval Settings: “Default Approvals” controls what the agent can do autonomously
- Branch Context: Working in “main” branch with Worktree integration
Right Panel – Project Context
- Files Tab: Shows your entire project structure (ITSMOrg Salesforce project)
- Changes Tab: Tracks modifications the agent makes
- The agent has full visibility into your codebase for context-aware assistance
This interface represents the shift from code completion to code collaboration. Instead of just suggesting snippets, VS Code’s agent can:
- ✅ Understand your entire project structure
- ✅ Execute multi-step tasks autonomously
- ✅ Maintain conversation history across sessions
- ✅ Be customized with specific skills and instructions
- ✅ Work with different AI models
- ✅ Integrate with your version control workflow
This is VS Code’s answer to AI-native editors like Cursor—bringing agentic capabilities directly into the world’s most popular code editor.
Conclusion
The evolution of VS Code into an agentic development environment marks more than just a feature update—it’s a fundamental shift in how we write, debug, and ship software. What began as a lightweight, extensible editor has matured into a collaborative workspace where AI doesn’t just suggest code; it understands context, executes multi-step tasks, and adapts to your workflow. While AI-native editors like Cursor demonstrated what’s possible, Microsoft’s decision to embed agent capabilities directly into VS Code ensures millions of developers can access next-generation tooling without leaving the ecosystem they already trust. As AI agents become more autonomous and deeply integrated, the developer’s role will continue to shift from writing every line to orchestrating intelligent workflows. The question is no longer whether AI belongs in our IDEs, but how we’ll leverage it to build faster, smarter, and more creatively. Have you tried VS Code’s new agent mode yet? Share your experience or favorite workflows in the comments below. Untill next post keep reading and sharing…
Leave a Reply