By now, you’ve heard about the latest frontier AI models that are remarkably good at finding vulnerabilities in code and creating potential exploits. So good, in fact, that these models have been significantly limited from general use in an attempt to give defenders time to find and fix vulnerabilities before attackers find and exploit them.

For context, on April 7, 2026, we began testing Anthropic’s Claude Mythos model as a launch partner for Project Glasswing. Our conclusion was clear: The latest models are extraordinarily capable at finding vulnerabilities and changing them into critical exploit paths in near-real-time. In Defender's Guide to the Frontier AI Impact on Cybersecurity, I shared our early findings and recommendations.

Since then, we’ve continued testing the latest frontier AI models, including Anthropic’s Mythos and Claude Opus 4.7 and OpenAI’s GPT-5.5-Cyber as part of the Trusted Access for Cyber program. The big question just a few weeks ago was: “Are we overstating the model capabilities?” With more testing, I can confidently say we weren’t. In fact, these models are likely even better at finding vulnerabilities than we initially realized. Today, we’re providing an update on our ongoing research, our learnings uncovered in the process, and the approach we’re taking to protect our customers.

Find and Fix Before Attackers Find and Exploit

Today, we released our May “Patch Wednesday” security advisories, our monthly cadence of transparent vulnerability disclosure and remediation. This is the first time where the majority of findings were the result of frontier AI models scanning our code.

• These are the results of the full, initial scan of over 130 products across all three platforms.
• As of today, we’ve patched all important vulnerabilities in our SaaS delivered products, and all customer-operated products now have patches available.
• Today’s advisory covers 26 CVEs (representing 75 issues) versus our usual volume (typically less than 5 CVEs in a month); none of which are being exploited in the wild. Note, this excludes CyberArk vulnerabilities, which are disclosed in their normal process.

It's important to understand this isn’t a one-and-done situation. We’re now rescanning, applying all our learnings about how to provide the right context and threat intelligence to the models. We intend to fix every vulnerability we find before advanced AI capabilities become widely available to adversaries.

While incredibly powerful, AI models aren’t simply magic. To achieve high-fidelity results, you need to build AI scanning harnesses, leverage context, guardrails and threat intelligence. We’ve also discovered a variance across models, due to variations in their training. A multimodel approach is required to identify the superset of vulnerabilities. And finally, while the immediate priority is finding and fixing the vulnerabilities that organizations currently have, the longer-term shift is incorporating these models directly into the software development lifecycle. This is the light at the end of the tunnel: A future where software is secure by design.

Four Steps Every Organization Needs to Take Immediately

Regardless of the current restricted access, we believe these capabilities will flow more broadly to other models. We now estimate a narrow three-to-five-month window for organizations to outpace the adversary before AI-driven exploits start to become the new norm. This impending vulnerability deluge demands urgency. Organizations that haven’t put appropriate safeguards in place will face an entirely new class of risk. Here’s what we recommend:

1. Find and Fix Vulnerabilities In Your Applications, Products and Code
   Find and fix before attackers find and exploit.
   • Leverage AI models to identify vulnerabilities across all codebase.
   • Apply the same AI scanning to your open-source supply chain, and remediate or mitigate findings.
   • Run accelerated patching tightly coordinated with product and development teams.
2. Assess, Reduce and Remediate Your Exposure
   Reduce what is reachable by attackers, secure what must be accessible, such as customer-facing applications.
   • Attack surface management products, like Cortex Xpanse^®, have never been more critical for finding and reducing exposure.
   • The latest frontier AI models are very adept (with the right AI scanning harness) at evaluating exposures, understanding security misconfigurations and prioritizing attack-path reachability.
   • Audit your supply chain, including AI infrastructure, runtime environments and model dependencies.
3. Ensure Attack Protections
   Vulnerability exploits are typically just one step of a multi-step attack lifecycle. Ensuring best-in-class protections is now even more important for preventing breaches.
   • Map current sensor coverage to identify critical blind spots in detection, prevention and telemetry.
   • Deploy best-in-class XDR everywhere with an emphasis on real-time ML-based detection and prevention of attacks with all hosts on-premises and cloud included.
   • Deploy Agentic Endpoint Security to secure wide-scale adoption of vibe coding and AI security across the enterprise (e.g. Prisma AIRS^® and our recent acquisition of Koi are now a necessity for securing the agentic endpoint).
   • Secure enterprise browsers with AI-based security are a must have for securing where users now do their work.
   • Zero trust and Identity Security are foundational to securing every user and connection, extending to internal segmentation and outbound application connections.
4. Deploy Real-Time Security Operations
   Autonomous AI-driven attacks will drive attack lifecycles to minutes requiring every SOC to achieve single-digit mean time to detect (MTTD) and mean time to respond (MTTR).
   • Attack detections must be AI/ML-driven to detect even frequently changing and novel attacks at scale.
   • These AI detections must operate against a wide range of first party and third party data sources. A best in class AI SOC must operate on ALL relevant data sources.
   • Automation, both natively integrated and throughout the SOC lifecycle, is necessary to achieve single-digit MTTR. This automation will increasingly be agentic.
   • This must be delivered as a platform to remove seams and gaps created by point solutions.
   • Assess and act as quickly as possible.

Fighting AI with AI — AI Frontier Security Innovations Coming Soon

So far, frontier AI models only find new attacks, not new attack techniques. This means that with the right innovations, we can expand our use of AI to solve the security challenges that organizations are facing, and deliver what our customers need to stay ahead of the ever-evolving threat landscape, including:

• Reimagining virtual patching with proactive, high-fidelity content updates across network, endpoint and cloud security – We expect that across open source and technology suppliers there will be a deluge of patches, and virtual patching will provide a mitigation layer necessary to give your teams time to update. We expect to roll out the first phase of capabilities very soon.
• Enhanced attack preventions, including cyber-LLM trained ML and small language models (SML) and behavior protections – Early testing with Cortex XDR^® and our network security security services, such as WildFire^® malware prevention, indicate high protection coverage from the types of attacks created using these new frontier AI models.
• Using these models to scan our code, applications and even security configurations – Our intention is to productize these capabilities and incorporate them into our platforms.

Unit 42 — We’re Here to Help

We recognize that not everyone has the capacity and/or expertise to action all of the recommendations to effectively counter frontier AI-driven risks in the short timeframe mandated by AI innovation. Our Unit 42 Frontier AI Defense service is designed to discover and remediate your current exposure before attackers do, strengthen controls that reduce exposure and contain impact and modernize security operations so teams can detect and respond at machine speed.

This is a pivotal moment for our industry. While the scale of the challenge presented is real, I’m confident in our ability to solve it. We’re here to help our customers navigate this transition and ensure that as the landscape continues to evolve, the advantage remains with the defender.

Forward-Looking Statements

This blog contains forward-looking statements that involve risks, uncertainties and assumptions, including, without limitation, statements regarding the benefits, impact, or performance or potential benefits, impact or performance of our products and technologies or future products and technologies. These forward-looking statements are not guarantees of future performance, and there are a significant number of factors that could cause actual results to differ materially from statements made in this blog. We identify certain important risks and uncertainties that could affect our results and performance in our most recent Annual Report on Form 10-K, our most recent Quarterly Report on Form 10-Q, and our other filings with the U.S. Securities and Exchange Commission from time-to-time, each of which are available on our website at investors.paloaltonetworks.com and on the SEC's website at www.sec.gov. All forward-looking statements in this blog are based on information available to us as of the date hereof, and we do not assume any obligation to update the forward-looking statements provided to reflect events that occur or circumstances that exist after the date on which they were made.

The post Defender's Guide to the Frontier AI Impact on Cybersecurity: May 2026 Update appeared first on Palo Alto Networks Blog.

Take the quickest path from detection to protection

Remove unwanted alerts from your environment using plain language

Red Canary brings critical Zscaler context into investigations to enable targeted threat response and save you time

Red Canary brings its AI-powered investigations to Wiz’s cloud-native application protection

Attackers often go after outdated and unused test accounts, or stumble upon publicly accessible cloud storage containing critical data that’s a bit dusty. Sometimes an attack exploits a vulnerability in an app component that was actually patched, say, two years ago. As you read these breach reports, a common theme emerges: the attacks leveraged something outdated: a service, a server, a user account… Pieces of corporate IT infrastructure that sometimes fall off the radar of IT and security teams. They become, in essence, unmanaged, useless, and simply forgotten. These IT zombies create risks for information security, regulatory compliance, and lead to unnecessary operational costs. This is generally an element of shadow IT — with one key difference: nobody wants, knows about, or benefits from these assets.

In this post, we try to identify which assets demand immediate attention, how to identify them, and what a response should look like.

Physical and virtual servers

Priority: high. Vulnerable servers are entry points for cyberattacks, and they continue consuming resources while creating regulatory compliance risks.

Prevalence: high. Physical and virtual servers are commonly orphaned in large infrastructures following migration projects, or after mergers and acquisitions. Test servers no longer used after IT projects go live, as well as web servers for outdated projects running without a domain, are also frequently forgotten. The scale of the problem is illustrated by Lets Encrypt statistics: in 2024, half of domain renewal requests came from devices no longer associated with the requested domain. And there are roughly a million of these devices in the world.

Detection: the IT department needs to implement an Automated Discovery and Reconciliation (AD&R) process that combines the results of network scanning and cloud inventory with data from the Configuration Management Database (CMDB). It enables the timely identification of outdated or conflicting information about IT assets, and helps locate the forgotten assets themselves.

This data should be supplemented by external vulnerability scans that cover all of the organization’s public IPs.

Response: establish a formal, documented process for decommissioning/retiring servers. This process needs to include verification of complete data migration, and verified subsequent destruction of data on the server. Following these steps, the server can be powered down, recycled, or repurposed. Until all procedures are complete, the server needs to be moved to a quarantined, isolated subnet.

To mitigate this issue for test environments, implement an automated process for their creation and decommission. A test environment should be created at the start of a project, and dismantled after a set period or following a certain duration of inactivity. Strengthen the security of test environments by enforcing their strict isolation from the primary (production) environment, and by prohibiting the use of real, non-anonymized business data in testing.

Forgotten user, service, and device accounts

Priority: critical. Inactive and privileged accounts are prime targets for attackers seeking to establish network persistence or expand their access within the infrastructure.

Prevalence: very high. Technical service accounts, contractor accounts, and non-personalized accounts are among the most commonly forgotten.

Detection: conduct regular analysis of the user directory (Active Directory in most organizations) to identify all types of accounts that have seen no activity over a defined period (a month, quarter, or year). Concurrently, it’s advisable to review the permissions assigned to each account, and remove any that are excessive or unnecessary.

Response: after checking with the relevant service owner on the business side or employee supervisor, outdated accounts should be simply deactivated or deleted. A comprehensive Identity and Access Management system (IAM) offers a scalable solution to this problem. In this system, the creation, deletion, and permission assignment for accounts are tightly integrated with HR processes.

For service accounts, it’s also essential to routinely review both the strength of passwords, and the expiration dates for access tokens — rotating them as necessary.

Forgotten data stores

Priority: critical. Poorly controlled data in externally accessible databases, cloud storage and recycle bins, and corporate file-sharing services — even “secure” ones — has been a key source of major breaches in 2024–2025. The data exposed in these leaks often includes document scans, medical records, and personal information. Consequently, these security incidents also lead to penalties for non-compliance with regulations such as HIPAA, GDPR, and other data-protection frameworks governing the handling of personal and confidential data.

Prevalence: high. Archive data, data copies held by contractors, legacy database versions from previous system migrations — all of these often remain unaccounted for and accessible for years (even decades) in many organizations.

Detection: given the vast variety of data types and storage methods, a combination of tools is essential for discovery:

• Native audit subsystems within major vendor platforms, such as AWS Macie, and Microsoft Purview
• Specialized Data Discovery and Data Security Posture Management solutions
• Automated analysis of inventory logs, such as S3 Inventory

Unfortunately, these tools are of limited use if a contractor creates a data store within its own infrastructure. Controlling that situation requires contractual stipulations granting the organization’s security team access to the relevant contractor storage, supplemented by threat intelligence services capable of detecting any publicly exposed or stolen datasets associated with the company’s brand.

Response: analyze access logs and integrate the discovered storage into your DLP and CASB tools to monitor its usage — or to confirm it’s truly abandoned. Use available tools to securely isolate access to the storage. If necessary, create a secure backup, then delete the data. At the organizational policy level, it’s crucial to establish retention periods for different data types, mandating their automatic archiving and deletion upon expiry. Policies must also define procedures for registering new storage systems, and explicitly prohibit the existence of ownerless data that’s accessible without restrictions, passwords, or encryption.

Unused applications and services on servers

Priority: medium. Vulnerabilities in these services increase the risk of successful cyberattacks, complicate patching efforts, and waste resources.

Prevalence: very high. services are often enabled by default during server installation, remain after testing and configuration work, and continue to run long after the business process they supported has become obsolete.

Detection: through regular audits of software configurations. For effective auditing, servers should adhere to a role-based access model, with each server role having a corresponding list of required software. In addition to the CMDB, a broad spectrum of tools helps with this audit: tools like OpenSCAP and Lynis — focused on policy compliance and system hardening; multi-purpose tools like OSQuery; vulnerability scanners such as OpenVAS; and network traffic analyzers.

Response: conduct a scheduled review of server functions with their business owners. Any unnecessary applications or services found running should be disabled. To minimize such occurrences, implement the principle of least privilege organization-wide and deploy hardened base images or server templates for standard server builds. This ensures no superfluous software is installed or enabled by default.

Outdated APIs

Priority: high. APIs are frequently exploited by attackers to exfiltrate large volumes of sensitive data, and to gain initial access into the organization. In 2024, the number of API-related attacks increased by 41%, with attackers specifically targeting outdated APIs, as these often provide data with fewer checks and restrictions. This was exemplified by the leak of 200 million records from X/Twitter.

Prevalence: high. When a service transitions to a new API version, the old one often remains operational for an extended period, particularly if it’s still used by customers or partners. These deprecated versions are typically no longer maintained, so security flaws and vulnerabilities in their components go unpatched.

Detection: at the WAF or NGFW level, it’s essential to monitor traffic to specific APIs. This helps detect anomalies that may indicate exploitation or data exfiltration, and also identify APIs that get minimal traffic.

Response: for the identified low-activity APIs, collaborate with business stakeholders to develop a decommissioning plan, and migrate any remaining users to newer versions.

For organizations with a large pool of services, this challenge is best addressed with an API management platform in conjunction with a formally approved API lifecycle policy. This policy should include well-defined criteria for deprecating and retiring outdated software interfaces.

Software with outdated dependencies and libraries

Priority: high. This is where large-scale, critical vulnerabilities like Log4Shell hide, leading to organizational compromise and regulatory compliance issues.

Prevalence: Very high, especially in large-scale enterprise management systems, industrial automation systems, and custom-built software.

Detection: use a combination of vulnerability management (VM/CTEM) systems and software composition analysis (SCA) tools. For in-house development, it’s mandatory to use scanners and comprehensive security systems integrated into the CI/CD pipeline to prevent software from being built with outdated components.

Response: company policies must require IT and development teams to systematically update software dependencies. When building internal software, dependency analysis should be part of the code review process. For third-party software, it’s crucial to regularly audit the status and age of dependencies.

For external software vendors, updating dependencies should be a contractual requirement affecting support timelines and project budgets. To make these requirements feasible, it’s essential to maintain an up-to-date software bill of materials (SBOM).

You can read more about timely and effective vulnerability remediation in a separate blog post.

Forgotten websites

Priority: medium. Forgotten web assets can be exploited by attackers for phishing, hosting malware, or running scams under the organization’s brand, damaging its reputation. In more serious cases, they can lead to data breaches, or serve as a launchpad for attacks against the given company. A specific subset of this problem involves forgotten domains that were used for one-time activities, expired, and weren’t renewed — making them available for purchase by anyone.

Prevalence: high — especially for sites launched for short-term campaigns or one-off internal activities.

Detection: the IT department must maintain a central registry of all public websites and domains, and verify the status of each with its owners on a monthly or quarterly basis. Additionally, scanners or DNS monitoring can be utilized to track domains associated with the company’s IT infrastructure. Another layer of protection is provided by threat intelligence services, which can independently detect any websites associated with the organization’s brand.

Response: establish a policy for scheduled website shutdown after a fixed period following the end of its active use. Implement an automated DNS registration and renewal system to prevent the loss of control over the company’s domains.

Unused network devices

Priority: high. Routers, firewalls, surveillance cameras, and network storage devices that are connected but left unmanaged and unpatched make for the perfect attack launchpad. These forgotten devices often harbor vulnerabilities, and almost never have proper monitoring — no EDR or SIEM integration — yet they hold a privileged position in the network, giving hackers an easy gateway to escalate attacks on servers and workstations.

Prevalence: medium. Devices get left behind during office moves, network infrastructure upgrades, or temporary workspace setups.

Detection: use the same network inventory tools mentioned in the forgotten servers section, as well as regular physical audits to compare network scans against what’s actually plugged in. Active network scanning can uncover entire untracked network segments and unexpected external connections.

Response: ownerless devices can usually be pulled offline immediately. But beware: cleaning them up requires the same care as scrubbing servers — to prevent leaks of network settings, passwords, office video footage, and so on.

David Fletcher// My wife and I recently purchased a 2016 Ford Flex to replace an aging version of the same make and model that met an untimely fate. During the […]

The post Internet of Things Exploration: 2016 Ford Flex appeared first on Black Hills Information Security, Inc..

Beau Bullock // I’ve added in a few modules to MailSniper that will assist in remote attacks against organizations that are hosting an externally facing Exchange server (OWA or EWS). Specifically, […]

The post Attacking Exchange with MailSniper appeared first on Black Hills Information Security, Inc..