How we’re making GitHub Copilot smarter with fewer tools

GitHub Copilot Chat in VS Code has implemented two new systems—embedding-guided tool routing and adaptive tool clustering—to enhance performance by reducing the default toolset from 40 to 13 core tools, thereby improving success rates and lowering response latency. By utilizing the Model Context Protocol (MCP) and embedding-based selection, Copilot optimizes tool usage by grouping similar tools into "virtual tools" and pre-selecting the most semantically relevant ones for a given task, resulting in a 94.5% Tool Use Coverage rate. These changes ensure more efficient tool selection and usage, leading to faster and more accurate responses without sacrificing user experience. Future directions involve developing context-aware agents capable of long-context reasoning to further improve tool selection and usage.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.