microsoft.com

Every one of the 38 is a mark on the map. Named now: 1 to 26, newest first.

1. Microsoft researchers expose backdoors in coding agent trainingmicrosoft.com12 Nov
2. Microsoft finds sleeper agent backdoors in AI model neuronsmicrosoft.com11 Nov
3. Microsoft finds Rust memory safety bugs using AI researchmicrosoft.com9 Nov
4. Microsoft tracks AI agents abusing exposed secrets in Azuremicrosoft.com6 Nov
5. Microsoft adds task based controls to govern AI agent accessmicrosoft.com5 Nov
6. Microsoft reports 7.3 billion phishing threats in Q3 2026microsoft.com5 Nov
7. Microsoft details fake installer campaign delivering monetisation payloadsmicrosoft.com3 Nov
8. Microsoft named Gartner Leader for AI agent management platformsmicrosoft.com3 Nov
9. Microsoft finds just 2,000 CVEs linked to real world attacksmicrosoft.com2 Nov
10. Malicious Claude Code Extension Plants Persistent SSH Backdoormicrosoft.com31 Oct
11. Microsoft details help desk impersonation and AD CS attacksmicrosoft.com30 Oct
12. IDC names Microsoft a leader in enterprise endpoint securitymicrosoft.com29 Oct
13. Microsoft urges CISOs to govern AI agents as adoption growsmicrosoft.com28 Oct
14. Compromised LiteLLM Release Steals Credentials on Python Startupmicrosoft.com23 Oct
15. Microsoft outlines controls to secure AI agents as adoption acceleratesmicrosoft.com19 Oct
16. Microsoft and NVIDIA Set Out Enterprise Safety Guide for Open AI Modelsmicrosoft.com16 Oct
17. Microsoft launches PQC TLS pilot to test post-quantum certificatesmicrosoft.com8 Oct
18. Microsoft Urges CISOs to Patch Critical Systems Within 24 Hoursmicrosoft.com6 Oct
19. Microsoft warns AI agents are widening the attack surface for enterprisesmicrosoft.com1 Oct
20. Microsoft Ignite 2026 Puts Agentic AI at the Heart of Securitymicrosoft.com30 Sep
21. Fake MSP360 RMM Installers Let Attackers Deploy ScreenConnectmicrosoft.com29 Sep
22. Microsoft Adds Network Controls to Block Sensitive Data From AI Agentsmicrosoft.com24 Sep
23. Microsoft Unveils Defender Security Hub for AI Era Attacksmicrosoft.com23 Sep
24. Microsoft Defender Missed 221 High-Severity Email Threats per 1,000 Usersmicrosoft.com17 Sep
25. AI-Assisted Invoice Scam Impersonates CEOs to Steal $50,000 Paymentsmicrosoft.com10 Sep
26. Microsoft Warns AI Brands Are Fueling a New Phishing Wavemicrosoft.com10 Sep

names a CVE that is exploited in the wildnames a CVE

Zoom in on the map and the dots under the middle of it are named where they are.