AI
news

27stories

Stories

MCP C# SDK 2.0
Stateless MCP servers
IQInfoQ .NET Edin Kapić Oct 26

MCP C# SDK 2.0 Makes HTTP Stateless by Default

Microsoft released version 2.0 of the official MCP C# SDK, implementing the 2026-07-28 Model Context Protocol revision. Streamable HTTP servers are now stateless by default, and Multi Round-Trip Requests (MRTR) let interactive tools ask clients for input. Releases 2.1 to 2.3 added hybrid session serving, a subscriptions/listen handler and several OAuth and discovery fixes.

Agent Framework
Stop when guards fail
DVDEV Community: #dotnet MICHAEL-MAURICE Oct 25

Making .NET AI Agent Guardrails Fail Closed When They Break

The article examines what a .NET AI agent does when its own guardrails fail, such as a throwing guard, a down classifier, a missing caller or miswired middleware. The author deliberately breaks each guard layer and shows how to make the request blocked instead of silently allowed. It ends with a practical checklist for developers building guarded agents.

Temporal predict
Where is the journey going?
MLMostly Lucid (Scott Galloway) Scott Galloway Oct 24

Predicting journey outcomes from timed behaviour in C#

Part 3 of Scott Galloway's series builds a temporal predictor in C#. It tracks a journey's centroid, velocity and acceleration in real time, then forecasts whether it ends in purchase, cancellation or departure within 60 seconds. The engine learns only from actual endpoints and abstains when evidence is thin, which makes it a concrete example for teams building SQLite-backed behavioural or service-health models.

Work IQ
Agents read Microsoft 365
KCKunal Chowdhury [email protected] (Kunal Chowdhury) Oct 23

Microsoft Work IQ: giving AI agents Microsoft 365 context via MCP and A2A

Microsoft has introduced Work IQ, a workplace intelligence layer that lets AI agents read and act on live Microsoft 365 and Dynamics 365 data. Developers can reach it through an MCP server, an A2A endpoint or a REST API instead of wiring up many Graph endpoints. Every request runs under the signed-in user's Entra ID identity.

ai cloud mcp 8 min
Foundry Agent
MCP in Foundry agents
BTTaiseer Joudeh (Bit of Tech) Taiseer Joudeh Oct 20

Connecting an MCP Server to Microsoft Foundry Agent Service

Part 9 of Taiseer Joudeh's MCP series shows how to attach an existing ASP.NET Core MCP server to a hosted Foundry agent without changing its code. The setup covers a Foundry project, an Entra app registration, a custom MCP connection using OAuth identity passthrough, and a toolbox. It matters for .NET teams that want the agent loop in Azure while user-based authorization still works.

Copilot SDK 1.0
Bring your own model
NB.NET Blog Bruno Capuano Oct 19

Bring Your Own Model to the GitHub Copilot SDK with C#

The GitHub Copilot SDK lets you keep GitHub's agent runtime while routing model inference to your own Microsoft Foundry or Azure OpenAI deployment through BYOK. The article walks through a .NET 10 console sample that configures ProviderConfig, authenticates with Microsoft Entra ID or an API key, and sends one prompt.

MCP · ASP.NET Core
Your API as MCP tools
DVDEV Community: #aspnetcore Rahmat Afridi Oct 19

Exposing Existing ASP.NET Core APIs to AI Agents with a Secure MCP Server

A practical guide to adding a Model Context Protocol server to an existing ASP.NET Core app using the official ModelContextProtocol C# SDK. It covers wrapping existing services as read-only tools, locking the endpoint down with JWT authentication and per-tool authorization, and verifying the surface with an MCP client. Useful for teams that want AI agents to use their current APIs without a rewrite.

Graph + RAG
RAG over Microsoft Graph
DPdotnetpro (de) Dr. Veikko Krypczyk Oct 19

Building a RAG assistant with Microsoft Graph and .NET

Microsoft Graph supplies calendar, mail and document data that can be indexed with Azure AI Search and passed to a language model to produce sourced answers. The article walks through a .NET retrieval-augmented generation flow with code for loading context, chunking documents and calling the model. It stresses least-privilege scopes and permission-aware retrieval, which matter for any enterprise assistant.

Feedback

Found a bug or have an idea? Your message lands straight in my inbox.

Type of message