Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

If you manage vendors, supplier drift looks familiar. A delivery arrives late. A vendor misses a service-level agreement (SLA) target, and nobody flags it. Tickets sit unresolved, and quality dips just enough that it never makes the weekly stand-up. The problem is timing. Most teams find out a vendor missed its uptime or response-time targets at renewal, months after anyone could’ve fixed it.

The tricky thing about AI compliance is that most organizations are going to experience it from both sides. You need to be able to explain how AI is being used inside your own organization, what it can access, and how you're managing the risk. At the same time, you need to understand how your vendors are using AI and whether that introduces new risk into your environment.

The European Commission has published its first official guidance to help businesses implement the Cyber Resilience Act (CRA)—the EU regulation establishing cybersecurity requirements for products with digital elements. The timeline is critical: reporting obligations for actively exploited vulnerabilities and severe incidents take effect on September 11, 2026, while the core CRA requirements will become mandatory from December 11, 2027.

Somewhere in your vendor list, right now, there is a door you have never checked. You didn’t build it, you don’t hold the key, and yet if someone walks through it, the breach notification goes out on your letterhead. The numbers too aren’t subtle. Verizon’s 2026 Data Breach Investigations Report found that 48% of breaches now involve a third party, up from 30% just a year earlier, the sharpest rise the report has ever recorded. Your vendors are your attack surface now.

CVE-2026-44756, tracked by SAP under the name OVERPASS, is a memory corruption vulnerability in the SAP kernel code that processes the Extended Passport (EPP), a standard SAP tracing structure used to follow call sequences across distributed system landscapes. The flaw stems from missing boundary validation during deserialization of EPP data, which causes unsafe memory behavior when the kernel processes externally supplied length fields. The vulnerability carries a CVSS v3.1 base score of 10.0 (Critical).

External threat models estimate how often somebody gets in and what they reach afterward. The susceptibility term does most of the work, weighing what an attacker can do against what the controls prevent. ‍ An insider is already inside. The credentials are valid, the access is entitled and the workflow is familiar. None of that makes the model harder to run, it changes which side of it breaks, and the break is on frequency rather than on magnitude. ‍

Three reads from surfaces most organizations already own produce a usable AI usage register in a morning. Entitlement, from the identity provider, showing who is licensed for what. Activity, from network or gateway logs, showing who reached which destination and how much. Identity, from the directory, showing who those people are and which scopes they sit in. ‍ The register answers more questions than people expect.

Approval works at the tool layer and it works well. A platform is assessed, terms are reviewed, a data processing agreement is signed, the tool enters the register, and named identities are entitled to it. Everything about that maps cleanly. ‍ Then somebody uses the approved platform to assemble an agent that acts on their behalf, with its own reach and its own credentials. The approval covered the application.

1.1.1.1 now validates DNSSEC signatures made with ML-DSA-44, a post-quantum signature algorithm standardized by the National Institute of Standards and Technology (NIST). This is a first step toward preparing DNSSEC for a future in which today’s signature algorithms are no longer secure. Cloudflare plans to achieve full post-quantum security by 2029. Much of the work so far has focused on TLS, but public-key cryptography is used in many other systems, including DNSSEC.

A browser agent in your cluster opens a supplier portal, screenshots it, and clicks somewhere the task never called for. The page looks exactly like the page the supplier serves, and to the person who checks it later, it still does. The classifier in front of the agent returned nothing, because the instruction that produced the click does not exist.