This week’s security landscape has been anything but quiet, with a critical ServiceNow vulnerability under active exploitation grabbing the spotlight. The flaw, tracked as CVE-2026-6875, is a pre-authentication remote code execution (RCE) bug in the ServiceNow AI Platform, and attackers are already using it in the wild. That makes this a ServiceNow RCE exploited scenario you need to take seriously right now. Alongside this urgent threat, the week brought a mix of evolving dangers: AI agents continue to log in as humans (most large companies now run more than one AI platform, increasing the attack surface), the Chaos ransomware variant msaRAT hides its command-and-control traffic inside a legitimate browser process using Chrome DevTools Protocol and WebRTC, and the new PR3TACK knowledgebase aims to close the gap between attacker invention and defender detection. Meanwhile, multi-patch vulnerability fixes in open source software remain a risk, as some fixes arrive as multiple commits where the first one leaves the flaw in place. This weekly security roundup covers the key developments you need to know about in cybersecurity news right now, from active exploitation of a pre-authentication RCE to the stealthy tactics shaping the threat landscape.
ServiceNow CVE-2026-6875: Critical Pre-Auth RCE Under Active Exploitation
Among the most urgent stories this week is the active exploitation of a critical vulnerability in ServiceNow’s AI Platform. This flaw, tracked as CVE-2026-6875, allows attackers to execute remote code without any authentication. That means anyone with network access to a vulnerable instance could potentially take full control—no login credentials required. The fact that this Servicenow rce exploited in the wild has security teams scrambling to understand the risk and find a way to protect their systems.
What Is CVE-2026-6875?
CVE-2026-6875 is a pre-authentication remote code execution vulnerability in the ServiceNow AI Platform. “Pre-auth” means the attack can happen before a user logs in, making it especially dangerous. The vulnerability is being actively exploited, but the specific attack vectors or methods used have not been publicly disclosed. This lack of detail leaves many organizations uncertain about which systems are at risk and how an attacker might gain entry.
How Is It Being Exploited?
While the exact exploitation technique remains under wraps, the impact is clear: attackers can run arbitrary code on vulnerable ServiceNow instances. This could lead to data theft, lateral movement within a network, or full compromise of the platform. Because the flaw is in the AI Platform, organizations that use ServiceNow’s AI features are particularly exposed. The active exploitation confirms that proof-of-concept code or exploit methods are already circulating in the wild, even if not publicly available.
Detection and Mitigation Challenges
As of now, no official ServiceNow patch has been confirmed for CVE-2026-6875, and mitigation guidance is absent. This leaves you in a difficult position. However, a useful tool has emerged: the Snowpick scanner, an open-source utility published by Bishop Fox. Snowpick is designed to help identify ServiceNow exposure on your network. By running this scanner, you can quickly check if your ServiceNow instances are visible and potentially vulnerable. While Snowpick won’t patch the flaw, it gives you a clearer picture of your attack surface—a critical first step while you wait for official remediation. Keep an eye on ServiceNow’s security advisories and consider implementing additional network-level controls, like restricting access to the AI Platform endpoints, until a patch arrives.
AI Agents: Human Impersonation, Breaches, and Code Vulnerabilities
While you wait for a ServiceNow patch, it’s worth stepping back to look at the bigger picture. The ServiceNow RCE exploited in recent attacks is a stark reminder of how quickly vulnerabilities can be weaponized. But the threat landscape is evolving in other ways too. AI agents are now a growing concern, as they continue to bypass security measures by logging in as humans. Most large companies run more than one AI platform simultaneously, and these agents can impersonate legitimate users across multiple systems. This makes it harder for traditional security tools to spot malicious activity, because the traffic looks like it comes from real employees.
Hugging Face Breach by Autonomous AI
One recent example underscores the risks. Hugging Face, a popular platform for machine learning models, disclosed a security breach carried out by an autonomous AI agent system. While details are sparse, the incident shows that AI-driven attacks are no longer theoretical. An autonomous agent executed the breach without direct human control, highlighting how AI agent security must become a priority for any organization using these tools.
Vulnerabilities in AI-Generated Code
AI isn’t just attacking systems — it’s also building them. Security firm Theori built 28 apps using AI coding agents and then scanned them through its pentesting platform. The results revealed vulnerabilities, though specific flaws were not disclosed. This suggests that code generated by AI can introduce the same weaknesses as human-written code, if not more. As you integrate AI into development workflows, you need to test that code just as rigorously.
- Human impersonation by AI agents is a real and present threat across multiple platforms.
- The Hugging Face breach by autonomous AI proves that attacks can happen without human intervention.
- Theori pentesting of AI-generated code found security gaps, emphasizing the need for thorough review.
These incidents all point to one thing: the ServiceNow RCE exploited is part of a broader trend where AI is both a target and a tool. Staying ahead means watching not just for traditional vulnerabilities, but for how AI agents might be used to exploit them.
msaRAT and Chaos Ransomware: Hiding C2 Traffic Inside Browser Processes
While the ServiceNow RCE exploited vulnerability has drawn attention, attackers are simultaneously refining how they hide their operations. A notable example is the Chaos ransomware family, which uses a component called msaRAT to keep its command-and-control (C2) traffic out of sight. Instead of relying on traditional covert channels, msaRAT tucks its C2 communication inside a legitimate browser process using Chrome DevTools Protocol and WebRTC. This technique makes the malicious traffic blend in with normal browser activity, evading many network monitoring tools.
How msaRAT Works
msaRAT is a remote access trojan that operates as part of Chaos ransomware. Its key innovation is hiding the C2 channel inside a browser process that you might already trust. By leveraging Chrome DevTools Protocol and WebRTC, msaRAT can send and receive commands through a browser’s own communication streams. Since browsers are constantly exchanging data, this traffic looks ordinary to security systems that scan for unusual patterns. The msaRAT component effectively turns your browser into a proxy for the attacker’s commands.
Chrome DevTools Protocol and WebRTC
Chrome DevTools Protocol is a standard interface for debugging and controlling Chrome-based browsers. WebRTC is a technology for real-time audio and video communication. Together, they allow msaRAT to establish a persistent, low-profile connection to its C2 server. The attacker can issue instructions, exfiltrate data, or deploy ransomware without triggering alerts that would flag a separate, suspicious process. This C2 hiding technique represents a significant advance in evasion, even though no real-world success rate data is available yet.
Implications for Network Detection
For network defenders, this means traditional monitoring that looks for unusual ports or outbound connections may miss the threat. You need to inspect browser traffic more closely, perhaps by analyzing WebRTC sessions or DevTools usage patterns. The Chaos ransomware example shows that attackers are constantly finding new ways to exploit trusted processes. Staying ahead requires understanding how tools like msaRAT operate and updating your detection strategies accordingly.
PR3TACK and OT Security: Preemptive Defense Against Novel Attacks
When attackers exploit vulnerabilities like the ServiceNow RCE, the gap between their invention of a new technique and your ability to detect it can be dangerously wide. That’s where the PR3TACK framework comes into play. PR3TACK stands for Preemptive Tactics and Countermeasures Knowledgebase, and its goal is straightforward: close that gap. Instead of waiting for a new attack to hit the news, PR3TACK helps you anticipate the moves attackers might make next. It provides defenders with proactive strategies to counter emerging threats before they become widespread problems.
The PR3TACK Framework
Think of PR3TACK as a forward-looking playbook. It shifts your focus from reacting to known attacks to preparing for unknown ones. By mapping out potential attacker behaviors, this knowledgebase gives you concrete steps to harden your systems. For example, if you understand how a novel exploit might chain together common misconfigurations, you can patch those gaps ahead of time. This preemptive defense is especially valuable in environments where even a brief downtime is unacceptable.
The Air Gap Myth in OT Security
When it comes to operational technology (OT) security—think industrial control systems, power grids, and manufacturing lines—many organizations still rely on an old assumption: the air gap. The idea is that if your critical systems aren’t connected to the internet, they’re safe. But Benjamin Bachmann argues that the air gap is mostly a myth in OT security. Modern OT environments often have hidden connections, such as remote maintenance ports, VPNs, or even USB drives carried between networks. Attackers are increasingly finding ways to breach these supposed barriers. Recognizing this myth is the first step toward a more realistic, layered defense strategy. Combining frameworks like PR3TACK with a clear-eyed view of OT vulnerabilities helps you build a security posture that doesn’t rely on outdated assumptions.
Multi-Patch Vulnerabilities: A Hidden Risk in Open Source Software
While layered defense helps, the reality of patch management is often more complex — especially in open source software. Recent high-profile vulnerabilities, such as the ServiceNow RCE exploited incidents, underscore how critical it is to understand not just whether a patch exists, but whether it actually works. New research from the University of Texas at Dallas reveals a troubling pattern: many vulnerability fixes arrive as multiple commits, and the first commit often leaves the flaw still in place.
The Study: Multi-Patch CVEs
Researchers analyzed 1,646 open source CVEs from 1999 to 2025 that required more than one patch. They found that a significant number of these fixes were incomplete at the first commit. This means that even after the initial patch is released, the software remains exposed until all subsequent patches are applied. For you, this creates a hidden window of opportunity for attackers — and a real headache for patch management.
Risks for Open Source Users
If you rely on open source components, understanding the multi-patch nature of many CVEs is crucial. The ServiceNow RCE exploited vulnerability is a reminder that even reputable platforms can have complex patch histories. Here are key takeaways:
- Don’t assume the first patch is enough. Always check for follow-up commits or additional fixes listed in the CVE record.
- Monitor commit logs for the projects you use. A single patch commit may not fully resolve the issue.
- Build a layered update process that includes verification of fix completeness, not just installation.
Incomplete fixes are a hidden risk that complicates open source security. By staying aware of multi-patch vulnerabilities, you can close the gap between patch release and true protection — and avoid the pitfalls that make the ServiceNow RCE exploited scenario so dangerous.
Frequently Asked Questions
How can organizations detect or prevent exploitation of ServiceNow vulnerabilities?
You should start by monitoring your network logs for unusual traffic patterns targeting ServiceNow endpoints. Deploy a web application firewall with rules tailored to block common exploit payloads. Keep your ServiceNow instance updated with the latest security patches as soon as they are available.
Is the ServiceNow pre-auth RCE exploit different from typical vulnerabilities in enterprise software?
Yes, this Servicenow rce exploited flaw stands out because it requires no authentication, meaning an attacker can trigger it without valid credentials. Most enterprise vulnerabilities need some level of access or user interaction. This makes it a higher priority to patch immediately.
What practical steps should I take if my organization uses ServiceNow and this vulnerability is unpatched?
First, apply any available security patch from ServiceNow without delay. If a patch is not yet released, implement temporary mitigations such as restricting network access to the ServiceNow instance and enabling additional logging. Review your incident response plan to ensure you can quickly address any signs of exploitation.
