The arrival of Claude Mythos, Anthropic’s frontier AI model with powerful offensive and defensive capabilities, has loudly reverberated in the cybersecurity community. Response has been swift.
In a recent interview “Proof of Concept: Mythos Clouds the Future of Cyber Defense” hosted by BankInfoSecurity, Gadi Evron, CISO-in-Residence for AI at the Cloud Security Alliance (CSA), and Rob T. Lee, Chief AI Officer and Chief of Research at SANS Institute outline exactly what security leaders must do right now. These two leading figures in the industry recently published a best-practices briefing from the CSA right after the Anthropic news.
The message is an important one: the security playbook needs a major update. Here’s why.
The Threat in Numbers
By all accounts, Mythos has scale. In internal tests, it produced 181 working Firefox exploits which was up from two in Anthropic’s prior best model under the same conditions. It hit a 72% success rate and can chain multiple vulnerabilities into a single attack path without human help.
Forescout’s Vedere Labs latest research shows that even off-the-shelf AI models can find vulnerabilities that human researchers missed — and top models (Claude Opus 4.6, Kimi K2.5) can now find and exploit flaws with minimal prompting.
Long term, this could push fixes upstream before release. Near term, it compresses timelines and raises the bar for already-stretched security teams.
The broader trend is equally alarming. According to the Zero Day Clock, the mean time from vulnerability disclosure to confirmed exploitation has collapsed from 2.3 years in 2019 to less than one day in 2026.
AI-assisted attack chains, once the exclusive province of nation-states, are now accessible to a far wider range of adversaries. The window security teams once relied on for patching and response is effectively gone.
Anthropic’s companion initiative, Project Glasswing, has provided early access to Mythos Preview to approximately 40 technology and cybersecurity vendors, giving the industry a brief head start. But the Cloud Security Alliance (CSA) and SANS Institute are clear: that window will not last.
What the CSA Recommends You Do Now
The full briefing is organized across three time horizons. For cybersecurity leaders, the key actions break down as follows:
| WHEN | WHAT TO DO |
|---|---|
| This week | Deploy LLM-powered agents against your own codebase and applications to identify exploitable vulnerabilities before attackers do.Audit AI agents already running in your environment — their prompts, tools, and pipelines.Enable strict egress filtering for all AI agent traffic.Revisit your organization’s current patch prioritization criteria given the compressed exploitation window. |
| Within 45 days | Automate triage and remediation pipelines to absorb the expected volume of incoming patches from Glasswing-participating vendors.Strengthen dependency management to reduce exposure from third-party and open-source components.Update your board briefing and risk register to reflect the new exploitation timeline realities. |
| Within 12 months | Build a standing VulnOps function — staffed, automated, and integrated with AI-driven discovery capabilities.Introduce AI agents broadly across the cyber workforce to give defenders the speed necessary to match adversaries.Re-evaluate risk tolerance for operational downtime caused by faster remediation cycles.Update governance structures to allow faster deployment of new AI-based defenses. |
The briefing also includes a 13-item risk register mapped to four industry frameworks – OWASP LLM Top 10 2025, OWASP Agentic Top 10 2026, MITRE ATLAS, and NIST CSF 2.0 –along with 10 diagnostic questions CISOs can use immediately to triage where their programs are most exposed.
Here’s a workbook of the CSA Risk Register.
Evron: “Mythos Is the First Wave”
Gadi Evron, CEO of Knostic and CISO-in-Residence for AI at the CSA, co-led the paper “The AI Vulnerability Storm: Building a Mythos-Ready Security Program” a strategy briefing drafted over a single weekend by more than 60 contributors and reviewed by over 250 CISOs.
The urgency in that timeline is intentional. As Evron has stated publicly: “We built this in three days because CISOs needed it now, not when it was perfect. Mythos is the first wave. The organizations that build the muscle now — the processes, the tooling, and a culture willing to adopt AI as a core part of how security gets done — will be the ones that meet the next wave on their own terms.”
Evron is also direct about the collective action problem that defenders face: “Attackers already operate as syndicates, crowdsourcing, sharing tools, moving as a collective. Defenders have to do the same.”
The briefing he co-authored identifies three strategic dimensions that security leaders must grapple with simultaneously. Operationally, teams should expect a surge in patch volume as Glasswing vendors begin releasing fixes — potentially mirroring the intensity of multiple simultaneous supply chain incidents compressed into weeks.
On risk, the CISO’s ability to manage exposure is becoming more constrained as business stakeholders are forced to revisit their tolerance levels. Strategically, the guidance calls for a longer-term gap analysis and an overhaul of governance processes to allow faster onboarding of AI-driven security controls.
Go deeper: See our initial take: “Claude Mythos: When Zero-Day Vulnerabilities Outpace Defenses”.
Lee: “This Is a Permanent Acceleration”
Rob T. Lee, Chief AI Officer and Chief of Research at SANS Institute, helped shape what may be the most actionable section of the briefing: a framework for how organizations must restructure their vulnerability operations to keep pace with machine-speed threats.
Lee’s view on what Mythos signals is unambiguous: “The window between vulnerability discovery and weaponization has collapsed into hours. What Mythos shows us is a permanent acceleration. This document gives CISOs something the commentary doesn’t: a risk register, priority actions with start dates, and a board briefing they can use this week.”
The emphasis on “permanent” matters. Lee and his co-authors are not describing a temporary spike in threat activity. They are describing a structural change to the offensive and defensive landscape — one that requires an equally structural response from defenders.
Go deeper: Watch our conversation “Security Without Certainty: Defending in the Age of Mythos”:
“The most important thing you can do as a security practitioner right now is use the runway you’ve been given,” says Rik Ferguson, VP of Security Intelligence, at Forescout. “We know this technology exists; there is no outcome from it right now. We’re not facing a tsunami of vulnerability reports and an avalanche of patches, so that gives you time to prepare and to build up the foundations. It is about making sure you have a complete understanding of what is connected to your network and how your business processes function with relation to your network architecture — and understanding the context.”
VulnOps: The Organizational Change That Can’t Wait
Perhaps the most important – and most underappreciated – concept in the CSA briefing is what Rob Lee frames as Vulnerability Operations, or VulnOps: a permanent organizational capability dedicated to AI-driven vulnerability discovery, triage, and remediation.
This is not a new project or a temporary task force. The briefing’s authors position VulnOps as the vulnerability-side analogue to DevOps — a continuous, staffed, automated function that becomes a standing pillar of the security organization.
Why does this matter so much? Because the security industry’s existing model is built for a world that no longer exists. Quarterly penetration tests, weeks-long patch cycles, and CVE-based threat intelligence pipelines were designed for a threat environment where organizations had months to respond to newly disclosed vulnerabilities. Lee has been direct in his writing: “Your quarterly pen test is adorable. The attackers run continuously.”
The CSA briefing makes VulnOps the longest-horizon priority action — something to build toward completing within 12 months, but something that organizations must begin architecting immediately. The function should be staffed and automated for continuous AI-driven discovery across the entire software estate. AI has now made this economically viable for organizations that previously could not afford it. The cost floor for sophisticated vulnerability discovery has dropped, meaning defenders can now access the same class of tools attackers are already deploying.
The briefing’s very first priority action reflects this logic directly: before addressing governance, before updating vendor contracts, before any other item — point AI agents at your own code this week. The goal is to find what attackers are already looking for before they find it first.
How Forescout Can Help
The network continues to be the immutable source of truth. The Forescout 4D Platform™ is the purveyor of first-party data about what’s on the network by:
- Continuously identifying and monitoring all connected devices, operating systems, and software.
- Automatically assessing exposure to emerging risks, including new vulnerabilities.
- Enforcing Universal Zero Trust Network Access with contextual and dynamic evaluation.
- Identifying logical network segments and implementing network-based segmentation enforcement.
Our capabilities map directly to three crucial areas classified as “HIGH” risk in the Mythos-ready Security Program Risk Register:
Incomplete Asset and Exposure Inventory
This is squarely in Forescout’s wheelhouse. We have highly differentiated capabilities around the visibility of devices, communications, vulnerabilities and exposures. By combining all those pieces of data, we enable decisions around remediation, patching, isolation, monitoring, containment, or segmentation.
Network Architecture Insufficient for Lateral Movement Containment
Building on the visibility of device-to-device communications, Forescout can baseline and monitor where lateral movement risks exist and overlay real-time device and landscape risk.
Continuous Vulnerability Management Maturity Gap
We understand the vulnerabilities, which devices matter most by criticality, and how exposed they are — and we can trigger the right response based on changing risk.
The Bottom Line for Security Leaders
Mythos is a near-future threat. Its capabilities are real, documented, and already in the hands of a limited set of organizations. The time to harden defenses before equivalent capabilities spread more broadly has begun.
The CSA and SANS are not calling for panic. They are calling for structural change. Security programs built around periodic review cycles, human-speed response, and reactive patching are, as the briefing puts it, “not designed for this speed.” The leaders who treat the coming months as a window to build muscle – in VulnOps, in AI-assisted defense, in governance agility – will be positioned to weather what comes next. Those who wait will not.
The real risk isn’t just finding vulnerabilities—it’s not knowing what you have, what’s exposed, and where lateral movement can spread once something is exploited.