MCP grows up: Chrome DevTools control, C# SDK 1.0, and early WebMCP
MCP tooling is rapidly maturing with a C# SDK 1.0, a Chrome DevTools MCP server for reliable browser automation, and early WebMCP experiments for agent-ready sites. The open-source [chrome-devtools-mcp](https://github.com/ChromeDevTools/chrome-devtools-mcp) lets agents like Gemini, Claude, Cursor, and Copilot drive a live Chrome, capture DevTools traces for performance insights, and automate flows via Puppeteer; note it exposes browser content to MCP clients. Microsoft’s C# SDK 1.0 adds full support for the 2025-11-25 MCP spec, including stronger auth server discovery, icon metadata, Client ID Metadata Documents over DCR, tool-in-sampling, and better long-running HTTP/SSE handling, per [InfoWorld](https://www.infoworld.com/article/4141959/mcp-c-sharp-sdk-1-0-arrives-with-improved-authorization-server-discovery.html). On the web surface, a practitioner guide outlines “WebMCP” patterns in Chrome Beta (navigator.modelContext, OAuth 2.1, and testing strategies) for making sites agent-ready ([guide](https://atalupadhyay.wordpress.com/2026/03/06/%f0%9f%8c%90webmcp-the-complete-guide-to-making-your-website-ai-agent-ready/)). Counterpoints argue for simpler agent workflows over Markdown/CLIs rather than MCP servers, echoing the CLI-first case in this [video](https://www.youtube.com/watch?v=-QznXY_pJvw&pp=ygUYQUkgY29kaW5nIGFnZW50IHdvcmtmbG93) and a [TNS piece](https://thenewstack.io/skills-vs-mcp-agent-architecture/), so choose based on control vs. complexity trade-offs.