Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Within Netwrix Security Research, we were helping a customer with an Entra ID assessment and knew they'd already done AD tiering on-prem. We also knew they had domain controllers virtualized in Azure, but it was hard to tell which Windows Server was actually a DC. We figured out that Azure Run Command lets you run commands as SYSTEM, so we used that to do reconnaissance and identify the DCs.

MITRE ATT&CK's detection analytic for adversaries mining SharePoint describes bulk access to files and metadata in a short window by privileged or rarely used accounts. That is also a description of Microsoft 365 Copilot answering a question. The technique hasn't changed, but the baseline has, and the exposure now happens with no vulnerability, no compromised credential, and no malicious intent anywhere in the chain.

Scammers are using AI to generate phishing pages that impersonate antivirus products, according to Malwarebytes.

An AI agent is software that can act on your behalf. It may read documents, emails, and web pages, then use the systems you connected to complete a task. Amazon Bedrock AgentCore makes these agents faster to build and deploy. But securing them requires more than limiting which systems an agent can access. You also need to consider what the agent can read once it gets there. A document, email, or web page may contain hidden or malicious instructions designed to influence an agent’s next action.

A tool gateway reads every call that crosses it and enforces policy on each one. It is blind to whatever never traverses it, which is established and not the interesting part. ‍ The useful number is what fraction of an agent's total action surface the gateway covers. Blindness of unknown size and blindness of known size are different problems, and only the second lets you state what a gateway-based claim is worth. ‍

Security controls change constantly. Firewall rules are modified, endpoint exclusions are added, identity policies are adjusted, and temporary exceptions appear as teams respond to new business and operational requirements. Some of those changes are expected. Some turn out to be harmless. Others quietly weaken the defenses organizations depend on. Detecting that a configuration changed is really important for security teams, but it’s just the beginning of the investigation.

If you're comparing vendors, our guide to the top Exposure Assessment Platforms in 2026 looks at how leading platforms approach discovery, prioritization, and remediation.

On September 4, 2026, Oren Yomtov, a security researcher from Accomplish, responsibly reported a vulnerability affecting Cloudflare Containers and Cloudflare Sandboxes (which is built on Containers), through Cloudflare’s bug bounty program. Cloudflare has fully remediated the vulnerability, and we have no evidence that customer data has been compromised.

When Mythos launched, headlines warning of its potential dangers were prolific. For security risk leaders, it felt like a watershed moment, one that signaled that AI had arrived, but simultaneously raised widespread concern about risk. Many wanted to understand how real the risk was for their organization and what should be done about it.

Docker Sandboxes isolate AI coding agents from the host. The GitGuardian Mixin Kit adds ggshield and AI hooks to scan prompts, tool actions, and outputs for secrets, giving developers a safer, repeatable path for agentic coding.