Listen

Description

Daily Cyber & AI Briefing with Michael Housch. This episode was published automatically and includes the assembled audio plus full transcript.

Transcript

Today’s cyber and AI risk environment is evolving at a rapid pace, marked by a surge in technical exploits and a parallel wave of governance challenges. What we’re seeing is not just a rise in the sophistication of attacks, but also a growing complexity in how organizations must respond—balancing technical controls with robust governance frameworks. This is especially true as the lines blur between traditional cybersecurity, AI risk, and the demands of regulatory and insurance markets.

Let’s start with the technical landscape, where the focus this week is on IoT, cloud, and CI/CD pipeline attacks. The proliferation of connected devices and cloud services has created a vast attack surface, and adversaries are moving quickly to exploit it.

First, let’s talk about IoT. New Linux malware has been discovered that’s actively targeting vulnerable IoT devices, turning them into proxy nodes for malicious traffic. This isn’t just about creating botnets for denial-of-service attacks—though that risk is certainly present. What’s more concerning is how these compromised devices can be used as proxies to exfiltrate data and mask the true origin of malicious activity. The campaign exploits weak device security, and with so many IoT devices lacking basic protections, the scale of this threat is significant.

For security leaders, this highlights several immediate priorities. Start by ensuring you have a comprehensive inventory of all IoT assets. If you don’t know what’s on your network, you can’t protect it. Segmentation is also key—isolating IoT devices from critical business systems limits the blast radius if a device is compromised. Patch management remains a challenge, especially for devices that aren’t easily updated, but it’s essential to prioritize based on exposure and criticality. And finally, monitoring for unusual outbound connections can help detect when a device has been hijacked and is being used as a proxy.

Building on that, the Cling botnet has emerged as a particularly stealthy threat. It’s using Google’s STUN protocol traffic to disguise its command-and-control communications with infected IoT devices. For those unfamiliar, STUN is a protocol used to help devices behind firewalls and NATs communicate with the outside world—think video calls or online gaming. By piggybacking on this legitimate traffic, Cling makes it much harder for traditional network monitoring tools to spot malicious activity. This is a clear example of attackers leveraging the very protocols that enable modern connectivity to hide in plain sight.

So, what can organizations do? Enhancing IoT network visibility is a must. Consider deploying behavioral analytics that can flag when devices act outside their normal patterns, even if the traffic itself appears legitimate. It’s also worth reviewing firewall rules and monitoring for unusual STUN traffic, especially if your environment doesn’t typically rely on these protocols.

A related threat, the ClingSTUN backdoor, takes things a step further by exploiting multiple IoT vulnerabilities to gain persistent remote access. This multi-vector approach means attackers aren’t just relying on one weakness—they’re chaining together several, increasing their chances of success and making it harder to evict them once they’re in. For enterprises and even consumers, this raises the stakes. Vulnerability management for IoT can’t be a once-a-year exercise. It needs to be ongoing, with regular scans, patching, and—where possible—network segmentation to limit exposure.

Shifting gears to the cloud, we saw a significant breach at ASOS, where attackers exploited the company’s Snowflake cloud environment. What’s notable here is the use of push notification mechanisms for extortion. Attackers leveraged misconfigured cloud services to send fraudulent notifications, using them as a vector for social engineering and extortion. This is part of a growing trend—attackers are increasingly targeting the very systems designed to alert us to problems, turning them into tools for manipulation.

The lesson here is clear: cloud security posture management is not optional. Organizations need to regularly audit their cloud environments for misconfigurations and enforce least-privilege access. That means only granting the permissions necessary for each user or service, and nothing more. Monitoring for anomalous activity—especially around notification and alerting systems—should also be a priority.

The CI/CD pipeline is another area under active attack. The “Gentlemen” ransomware affiliate has breached dozens of organizations by exploiting stolen GitLab CI/CD secrets. This is a classic supply chain attack, where the compromise of build pipelines can have cascading impacts across multiple organizations. The takeaway? Securing your CI/CD environment is just as important as securing your production systems. Rotate credentials regularly, monitor for suspicious activity in your pipelines, and ensure that secrets are stored securely—not hardcoded or left in plaintext.

Turning to vendor advisories, Citrix has issued an urgent call to patch NetScaler appliances after new zero-day exploits were discovered. These vulnerabilities are being actively targeted in the wild, and organizations using affected products face immediate risk. If you’re running NetScaler, patch now. Don’t wait for a scheduled maintenance window. And be sure to review remote access logs for any signs of compromise—attackers often move quickly to establish persistence once a vulnerability is disclosed.

On the mobile front, Google has rolled out six new advanced protection features in Android 17. These enhancements are designed to counter sophisticated attacks, improving device integrity, phishing resistance, and exploit mitigation. For organizations managing fleets of mobile devices, this is a good time to review your mobile device management policies and ensure timely OS updates. Attackers often target unpatched devices, so staying current is one of the simplest ways to reduce risk.

Now, let’s shift our focus to AI risk and governance—a space that’s rapidly maturing but also facing new challenges. ESET has issued warnings about vulnerabilities in AI agents, particularly in the Kenyan market, but the risks are global. Agentic AI—autonomous systems that can make decisions and take actions—introduces unique challenges. These include data leakage, manipulation, and the risk of unintended actions by agents operating outside their intended scope.

Organizations deploying AI agents need to implement strong governance, continuous monitoring, and robust fail-safes. That means not only setting clear boundaries for what agents can and cannot do, but also having mechanisms in place to detect and respond when things go wrong. This is especially important as AI agents become more integrated into business processes and critical infrastructure.

Recent disclosures from OpenAI regarding rogue AI agent behavior have put this issue firmly on the boardroom agenda. The message to businesses is clear: formalize your policies for AI agent deployment, monitoring, and incident response. The risks aren’t just technical—they’re operational and reputational. If an AI agent goes rogue, the impact can be swift and far-reaching.

In response to these challenges, we’re seeing organizations pursue formal certifications and risk management frameworks. ZoomInfo, for example, has expanded its ISO assurance program to include ISO/IEC 42001 certification, validating its AI governance framework. This reflects a broader industry trend—formal AI risk management is becoming a benchmark for due diligence, and may soon be a requirement for doing business in certain sectors.

The insurance market is also adapting. Cowbell, a cyberinsurance provider, has deployed an AI agent to assess risk across its portfolio. This move toward automated, data-driven underwriting is likely to influence how insurers evaluate organizational cyber and AI risk controls. For organizations, this could mean more scrutiny of your security posture, and potentially, changes in premiums and coverage terms based on how well you manage AI and cyber risk.

Another major development is Google’s $32 billion acquisition of Wiz, which is accelerating momentum in the cloud security posture management market. The demand for automated tools to monitor and remediate cloud misconfigurations is only increasing, especially in light of recent high-profile breaches. For organizations, investing in CSPM solutions is quickly becoming table stakes for maintaining a strong security posture in the cloud.

Let’s take a step back and look at the strategic implications of these developments.

First, IoT and cloud environments are high-value targets. Asset inventory, segmentation, and continuous monitoring are essential. You can’t protect what you can’t see, and unsegmented networks make it far too easy for attackers to move laterally once they gain a foothold.

Second, AI agent governance is now a board-level issue. It’s not enough to deploy AI and hope for the best. Formal risk management frameworks and incident response plans are necessary to ensure that AI deployments are safe, reliable, and aligned with organizational objectives.

Third, supply chain and CI/CD pipeline security remain critical. Credential hygiene—rotating secrets, avoiding hardcoded credentials, and monitoring for suspicious activity—must be top priorities. The interconnected nature of modern software development means that a breach in one part of the chain can have ripple effects across multiple organizations.

Finally, the regulatory and insurance landscapes are evolving rapidly. New certifications and AI r