The Next Generation of Developer Productivity with GitHub Copilot & Visual Studio
summary
Shows Visual Studio’s GitHub Copilot integration actually building features: uses agent mode plus MCP (Model Context Protocol) via an MCP.json server to pull GitHub issues, and a repo-level copilot-instructions.md to enforce style. Demo reproduces an XML formatter and a URL encoder/decoder inside DevToys (30k-star repo), runs LLMs Claude 4 and GPT‑4.1, creates checkpoints, attaches build logs, and produces PRs; demos mostly succeed though earlier runs sometimes fail without sufficient context.
AI-written summary · about
chapters (12)
- 0:00AI-powered developer productivity
- 1:51AI coding assistance evolution
- 3:12DevToys and demo challenge
- 5:58Copilot instructions guide AI
- 10:38Fixing build errors in VS
- 14:49Agent mode and MCP
- 17:38XML formatter with agent mode
- 20:27Spec-driven development with MCP
- 33:29Copilot agent on issues
- 48:00Copilot agent outside VS
- 56:07Reviewing AI changes
- 1:00:22Future of AI in workflows
related talks
01h 00mTips & Tricks: Get the Most Out of GitHub CoPilot
42mWhat’s new in GitHub Copilot and Visual Studio
28mMaking unfamiliar C++ code familiar with GitHub Copilot
01h 20mReal World Development with GitHub Copilot and VS Code
29mBoost your coding productivity with Github Copilot Agent
46mPragmatic techniques to get the most out of GitHub Copilot
47mUnlocking Productivity with GITHUB Copilot in 45 Minutes
40mLevel Up with Copilot