Pinned
From graph-based context to build-your-own Microsoft Security
Copilot agents, we're introducing a new AI-ready security platform
and innovations to empower defenders in the era of agentic AI:
msft.it/6010sQNtI
#MSSecure
00:00
Microsoft Security
22.9K posts
Microsoft Security
@msftsecurity
Be first to know about AI, threats, and new tools. Quick hits, expert tips, and real-time security news—follow for smarter, safer ops.
- A few months ago, we introduced codename MDASH, our new multi-model agentic scanning system, now in expanded private preview. Built to discover, validate, and prove software vulnerabilities at AI speed, the system is helping us explore new benchmarks and insights as we apply it
- The vision: bring people and AI together while a security platform continuously shields against and disrupts attacks. We’re honored that @forrester recognizes this vision, and named Microsoft a Leader in The Forrester Wave™: Extended Detection and Response Platforms, Q2 2026.
- We don’t just say it works—we show it, quarter over quarter. New email security benchmarking data is out now—built on real-world telemetry, not synthetic tests. The TL;DR: ➡️ Microsoft Defender consistently misses fewer high-severity cyberthreats than any SEG vendor evaluated
- 100+ specialized AI agents, one goal: find exploitable vulnerabilities before attackers do. Meet MDASH, Microsoft Security’s multi-model agentic scanning harness. Watch Sarah Young show it in action at #MicrosoftBuild.
00:00 - Just announced at #MicrosoftBuild: NEW security capabilities that deliver real-time guidance, scale with complexity, and give teams full-lifecycle visibility. 🧵Replying to @msftsecurity🔐 Purview data risk signals: Embedded in the Foundry Control Plane to detect sensitive data exposure and unsafe agent behavior. 🧵4/5🛡️ Defender AI model scanning: Find and fix vulnerabilities earlier with integrated security insights. Read all announcements here: msft.it/6011vb7oW🧵5/5
- 500M+ monthly players across Xbox consoles, PC, handheld, and cloud gaming. 🎮 That’s a lot to protect, but gaming goes far beyond accounts and devices. Gaming culture runs deep, spanning players, developers, platforms, and the teams behind it all—and each brings its own
- The Forrester Wave™: Workforce Identity Security Platforms, Q2 2026 highlights just how critical strong identity foundations are, especially as identity continues to be a top target for attackers. As organizations look for identity platforms that can keep up with evolving
- May (drop) Day is here 🙌 This month’s "In the Loop" updates are all about simplifying—across data security, identity, and AI. View the replies to see what’s NEW ⤵️Replying to @msftsecurity🧵 3/4 Microsoft Entra ID Account Recovery: enables users to regain access to their organizational accounts when they've lost access to all registered authentication methods.🧵 4/4 Windows 365 for Agents delivers a secure AI agent execution environment: Windows 365 for Agents and Agent 365 work together to provide a consistent, secure environment to run and govern agents. Agent 365 determines the work an agent is authorized to do and Windows 365 for
- AI safety should be continuous, not a checkpoint. Announcing two new open-source tools to help engineers build safety in from the start: ➡️ RAMPART: turn red-team techniques into repeatable, CI-ready tests ➡️ Clarity: validate what you should build before writing code Bringing
- DDoS attacks are on the rise. Solving them starts with stepping back—and going deep on defense. What that looks like: ➡️ Make abnormal traffic easier to spot and harder to exploit with rate limiting, geo-fencing, and anomaly detection ➡️ Design for graceful degradation:
- Project MDASH reaching #1 on the CyberGym leaderboard points to a shift in AI-driven security. A multi-model system designed and built from the ground up. It’s a strong signal of where autonomous code security is headed. msft.it/6014vuaV8
