Browse All Posts (121)

GitHub recounts how AI-assisted security researchers linked two unconfirmed PHP bugs to remote code execution, prompting an urgent FBI advisory. The video follows the response to leaked exploit details, including an emergency pull request for the official PHP Docker image and a discussion drawing more than 500 comments.

Emanuele Bartolesi shares technical takeaways from GitHub Universe 2026, covering Copilot model orchestration, persistent agent sessions, stacked pull requests, Azure-backed Git storage, and security previews. His recap includes CLI examples, enterprise deployment lessons, and practical limitations around billing, verification, permissions, and reported performance results.

Nick Brady reviews October’s Microsoft Foundry updates, including the managed Copilot harness preview, browser automation and hosted-agent resilience, and typed decisions with GPT-6 Luna. Alongside a Python routing example, he explains deployment constraints, checkpointing, browser safety, and SDK compatibility checks to make before upgrading.

sapanpatel outlines a supervisor-and-worker architecture for multi-agent systems in Microsoft Copilot Studio. The guide covers delegation patterns, structured context passing, asynchronous orchestration, and human escalation, alongside Power Platform environment isolation and audit logging to keep specialized agents bounded and preserve context during handoffs.

Fabrizio-Ferri demonstrates how Azure Route Server route maps summarize 100 spoke network prefixes into three aggregate routes for ExpressRoute. The walkthrough covers outbound policy configuration and route verification, while explaining how independent inbound and outbound policies support filtering, path selection, and network isolation across hybrid connections.

aonelakeuser announces general availability of OneLake security support for Eventhouse external tables in Microsoft Fabric. KQL queries now enforce centrally defined table, column, and row access controls automatically, allowing users to analyze Lakehouse data through Eventhouse without copying it or creating a separate path around existing permissions.

Yassine El Ouardi and Rafia Aqil demonstrate how to bill eligible Fabric Copilot usage to a separate capacity without moving the underlying workspace. Their lab assigns users to an F2 Copilot capacity while retaining analytics workloads on F16, then checks model assistance and capacity consumption.

Allison details CodeQL 2.27.4, which improves security query accuracy, reduces false positives, and corrects CWE metadata across several languages. The release also strengthens pack publishing checks and keeps JVM diagnostics separate from machine-readable CLI output, with automatic deployment to GitHub code scanning and inclusion in GitHub Enterprise Server 3.24.

Mariyaali explains how Microsoft Fabric’s Monitor in warehouse action connects capacity spikes to warehouse query history. The generally available capability carries the selected time window into Warehouse Monitor, where CPU time, data scanned, caching indicators, and workload metadata help identify expensive queries. Skills for Fabric offers an AI-assisted correlation option.

MattH_OneLake introduces Iceberg writes through OneLake table APIs and a preview API for table and column descriptions and tags. Compatible engines can create tables, commit snapshots, and obtain short-lived storage credentials, while shared metadata gives readers consistent context across tables and internal OneLake shortcuts.

Allison announces the public preview of GitHub Copilot in Xcode’s built-in coding assistant, starting with Xcode 27.3 beta. The integration supports all Copilot plans and requires macOS Tahoe 26.6 or later. The existing GitHub Copilot for Xcode extension remains available as a separate option.

alexathompson and Abhishek Narain explain how Microsoft Fabric administrators can inventory connections, restore ownership, and enforce authentication-method and Microsoft Entra tenant allowlists. The guidance includes REST API examples, deployment requirements for a read-only inventory app, and precautions to avoid disrupting existing connections and dependent workloads.

John Savill examines Microsoft eXecution Containers (MXC) as a way to contain AI agents and their processes on a local machine. The video covers operating system elements, containment, granular policies, a demonstration, learn modes, and flagging agentic processes.

John Savill provides a quick overview of Microsoft Execution Containers (MXC), focusing on their role in containing AI agents and the code those agents create. The video introduces containment as an aspect of running AI-generated code.

mk_sunitha explains how Fabric Apps connectors bring lakehouse, warehouse, SQL database, and Power BI semantic model data into focused applications. The article outlines supported operations, illustrates network operations and retail scenarios, and shows a CLI example for adding a read-only warehouse connector with permissions scoped to the application's needs.

shiv_narayanan introduces the preview of Salesforce Data 360 integration with OneLake. Fabric teams can query shared customer data alongside enterprise tables without maintaining a separate ingestion copy. The article covers shortcut setup, authentication choices, analytics and ML use cases, and the object types supported during preview.

Mark Russinovich explains how Azure Failure Prediction and Detection identifies hardware problems, moves eligible workloads, and coordinates repairs. Using SAP workloads as an example, he outlines notification channels and operational steps for mission-critical services, including live migration limitations, Scheduled Events handling, and Pacemaker configuration for high-availability clusters.

GitHub explains how GitHub Copilot verified proficiency assesses everyday tool usage without a separate exam. The video covers opting in, understanding the five proficiency tiers, reviewing results, sharing capability statements on LinkedIn, and claiming a Microsoft Pro Badge through Credly when eligible.

Eduard van Valkenburg and Wes Steyn introduce installable Microsoft Agent Framework builder skills for .NET and Python. They explain installation, pairing with Microsoft Learn MCP, and prompting coding agents to choose suitable patterns, preserve existing behavior, and validate changes. These skills guide development rather than run inside deployed applications.

Emanuele Bartolesi describes how he carries product requirements through AI-assisted implementation, review, testing, and deployment. His workflow combines GitHub Copilot CLI with reusable .NET skills, Playwright checks, and GitHub Actions releases to Azure, while keeping acceptance criteria, merge decisions, and checks after deployment under human review.

John Edward outlines a practical Windows developer workstation setup using WSL 2, Windows Terminal, Winget, PowerShell, and Docker Desktop. The guide covers filesystem choices, repeatable tool installation, editor configuration, local containers, and security settings that help reduce everyday setup friction while protecting development credentials and files.

The Visual Studio Code Team introduces the 1.146 Insiders release notes, which will evolve as features are added. The page links to the commit log and closed issues for tracking progress, but currently contains a placeholder rather than descriptions of specific changes.

Prakhar-Sharma demonstrates how to add a local MCP operations layer to an Azure VM image pipeline. The walkthrough covers querying Azure DevOps runs, checking release evidence, testing guarded pipeline queueing, and exposing the tools through GitHub Copilot Chat in Visual Studio Code while preserving existing approvals and permissions.

semod92 outlines a proposed Azure Container Registry cleanup approach combining soft delete, untagged-manifest retention, and scheduled image purges. The post includes CLI examples and dry-run checks, alongside considerations for protecting rollback images, incorporating vulnerability findings, monitoring deletions, and aligning cleanup rules with deployed workloads.

subhojitbasak explains how Arc-registered SQL Server instances can now migrate databases through the Azure portal without installing the Arc agent. The announcement covers supported Azure targets, migration methods, and limitations, including agent-dependent readiness assessments, manual database inventory updates, and the absence of Managed Instance link support.

Shivani650 walks through deploying Traefik on Azure Kubernetes Service using Helm, then connecting a private ingress endpoint to Azure Front Door Premium. The guide covers hostname and path routing, URL rewriting, TLS certificates, Private Link configuration, and checks for validating each layer of the application traffic flow.

Allison explains how the GitHub Copilot usage metrics API now separates local BYOK and enterprise custom model usage from Copilot-hosted models. The update adds user-level and aggregate reporting without changing existing totals, with important limitations around client telemetry, token counts, model names, and historical data.

Azure-Hacker examines ten Azure Local deployment mistakes, from unsupported hardware and permanent DNS choices to inherited policies, insufficient storage reserves and untested updates. The guide includes commands and configuration checks for architects and platform teams, with practical advice on access control, monitoring, backup and disaster recovery before production deployment.

GitHub introduces using GitHub Copilot through JetBrains Air, connecting an existing account to run coding and testing tasks locally with project context or in isolated cloud environments with Air Teams. The video covers plugin configuration and explains how this option differs from the native GitHub Copilot plugin.

Jason Farmer from Microsoft's digital operations team explains how an operational technology virtual coworker runs on site and keeps working when its cloud connection drops. The video highlights local execution as part of physical AI architecture, focusing on continuity when cloud connectivity is unavailable.

Allison announces Dependabot version update support for Agent Package Manager, helping keep AI agent skills, prompts, and instructions current through pull requests. The update handles supported Git dependencies and existing lockfiles, is available across all GitHub plans, and covers version updates rather than security updates.

soniagupta outlines three Azure Files public previews for data protection and recovery: vaulted backup for NFS shares, in-place snapshot rollback for SMB shares, and cross-region copy and synchronization through Azure Storage Mover. These capabilities address operational mistakes, regional disruptions, and recovery from ransomware using separate, protected backup points.

NaveenGopalakrishna explains how to deploy Claude Code through Microsoft Foundry and Azure API Management with Entra-based access, private networking and per-person budgets. The guide covers staged deployment, client configuration and acceptance tests, while documenting limits in regional recovery, model quotas, usage accounting and Desktop compatibility.

shilmenon describes expanded Microsoft Fabric mirrored catalog support for Azure Databricks materialized views and streaming tables. Fabric provides read-only access without copying data, while Databricks retains ownership of source processing. The update covers external-access prerequisites, catalog synchronization, SQL querying, and Power BI reporting through Direct Lake.

Richard Lander explains how NuGet packages can carry version-matched agent skills alongside library binaries. He demonstrates tools for discovering and copying those skills into repositories, compares package delivery with Git-based approaches, and discusses model evaluation, token costs, and the security implications of agents consuming package documentation.

bytecascade explains how Reference Data Join in Microsoft Fabric Eventstream enriches live telemetry with context from OneLake Delta tables. A cold-chain example demonstrates visual and SQL joins, reference-data refresh intervals, and downstream routing, while outlining the trade-offs between enriching events in motion and joining them after ingestion.

tompeplow demonstrates how OneLake supports agent-driven data engineering through hosted MCP servers, endpoint discovery, and shortcuts. Using GitHub Copilot, the examples explore mirrored databases, document permissions, combine external data in a Lakehouse, and run SQL and Spark analytics while working within existing access permissions.

vamsiboya_microsoft proposes a private-access pattern for Azure Databricks workspaces with overlapping compute CIDRs: place uniquely addressed front-end private endpoints in a Transit VNet while keeping compute networks isolated. The article explains DNS, browser authentication, and security dependencies, and explicitly distinguishes this architectural proposal from a tested reference architecture.

Microsoft Security Research and Jiacen Xu examine a fake-download campaign that bundles proxyware, distributed-computing software, and unwanted utilities through trojanized installers. They explain its staging and callback patterns, scheduled-task persistence, and Microsoft Defender hunting queries, with guidance for correlating evidence across an entire installer session rather than individual payload attempts.

SudhanshuKhera introduces Microsoft's ransomware recovery guidance for Azure SQL Database and Azure SQL Managed Instance. The seven-step framework covers preparation, containment, restoration, validation, and hardening, emphasizing clean restore points, preserving backups and evidence, and testing recovery procedures before an attacker compromises your data.