A joint Tenable-SentinelOne analysis of 93 CVE-actor attribution pairs reveals that both state-sponsored actors and cybercriminals independently converge on the same edge infrastructure.

It is the shared attack surface where state-sponsored threat actors and financially motivated criminal groups independently converge — not the province of a single adversary category, and not exclusively a nation-state problem, despite two years of headlines about China-nexus actors targeting Ivanti, Fortinet, and Palo Alto Networks. The data here tells a different and much broader story. One focused on vendors vs CVEs.

Key Takeaways

• Two independent observation systems, Tenable exposure telemetry across thousands of customer containers and SentinelOne DFIR casework across 66 CVEs, converge 79% on the same vendor attack surfaces despite minimal CVE-level overlap.
• Twelve CVEs in the combined dataset have confirmed multi-nexus attribution: state-sponsored and criminal actors independently exploiting the same vulnerability, across five nexus categories (China, Russia, DPRK, Iran, ransomware).
• The exposure picture is flatter than the headlines suggest: Fortinet, the vendor most associated with edge-device attacks in the press, sits mid-pack on container-grain exposure (25%) — well behind F5 (54%) and in a tight 10-point band with Check Point, Ivanti, and Citrix.
• 54% of customer environments running F5 products have at least one exposed, actively-exploited CVE; Citrix customers show the slowest remediation patterns at 461 days median time to patch.
• Remediation complexity, particularly of high priority CVEs, leads to a statistically significant 24-day remediation gap, leaving large windows of opportunity for attackers.
• The same product lines get hit again and again: Ivanti EPMM and Ivanti Connect Secure each show a newly exploited CVE roughly every 8.5 to 13 months.
• Leverage multiple defense-in-depth strategies: patch as quickly as possible, but also minimize the attack surface (feature-set minimization) and run endpoints in protect mode to better stop lateral movement from attacks that gain initial access.

The Convergence is the Story

Twelve CVEs in the combined dataset have confirmed multi-nexus attribution: state-sponsored and criminal actors independently exploiting the same vulnerability, across five nexus categories. Four examples illustrate the pattern:

CVE	Product	Actors (Nexus)	Significance
CVE-2026-15409	SonicWall SMA1000	UTA0533 (unattributed) + INC Ransomware	Espionage-to-ransomware succession on an active zero-day
CVE-2023-42793	JetBrains TeamCity	APT29 (Russia) + Lazarus (DPRK)	Two state-sponsored actors from different nations on the same CVE
CVE-2024-3400	PAN-OS GlobalProtect	UTA0218 (China) + INC Ransomware	China-nexus zero-day reused by ransomware operators
CVE-2024-24919	Check Point Quantum	PurpleHaze (China) + Fox Kitten (Iran)	China and Iran independently exploiting the same gateway vulnerability

The remaining eight confirmed multi-nexus CVEs span Fortinet, Citrix, Cisco, and Ivanti product lines. State-sponsored actors and ransomware operators are not operating in separate vulnerability ecosystems. They share the same entry points into the same products. The breadth of the convergence, not any single actor’s activity, is the finding.

That pattern holds across the full combined analysis. Three conclusions emerge:

Vendor attack surfaces are the persistent exploitation target. The same eleven vendors (i.e., Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, VMware, Microsoft, Oracle, CrushFTP, and Meta’s React framework) appear in both observation systems at 79% convergence, and all seven edge-product vendors converge. Serial exploitation timing on Ivanti products shows the vulnerability-to-exploitation pipeline refreshing at 8.5 to 13-month intervals on the same product lines. This is structural, not episodic. Patching the current CVE does not remove the vendor attack surface from the threat landscape.

State-sponsored and ransomware actors converge structurally. Twelve CVEs with confirmed multi-nexus attribution span all five nexus categories and cross the state-criminal divide. Defending against one actor category on edge devices necessarily requires defending against all of them, because the attack surface is shared. An organization that patches only for nation-state TTPs leaves itself exposed to ransomware operators exploiting the same vulnerability, and vice versa.

High-priority CVEs take more time to remediate, not less. Across Tenable’s 238-CVE high-priority list, high-priority CVEs carry a median remediation time of 146 days, compared to 122 days for all other CVEs — a 24-day gap that is statistically significant. The edge-appliance-specific subset (52 CVEs) shows a consistent 8-day gap in the same direction, which was not statistically significant, but suggests the direction may hold for edge appliances too. External data corroborates the pattern: the 2026 Verizon Data Breach Investigations Report (DBIR) found that median patch time increased from 32 to 43 days year over year, even as exploitation overtook credential theft as the number one initial access vector, and the 2025 DBIR, which incorporated Tenable RSO’s remediation trend analysis across 17 edge-related CVEs, found that only 54% of edge device KEVs were fully remediated. The explanation is structural: edge devices are the network boundary, so patching a VPN gateway or firewall means downtime for every user behind it, and change management gates multiply. These devices also resist standard patching workflows because they do not run endpoint agents, often require firmware-level updates with manual validation, and frequently lack active support contracts. The result is that the devices most worth patching are operationally the hardest to patch — and as the high-priority queue grows (the 2026 DBIR reports 50% more critical vulnerabilities to patch than the prior year), everything on it waits longer.

Background

Edge and perimeter devices occupy a uniquely consequential position in enterprise architecture. VPN gateways, firewalls, remote access appliances, and application delivery controllers sit at the boundary between trusted and untrusted networks. They are very often the first component an attacker touches and, for many organizations, the last component that gets patched. When one of these devices is compromised, the attacker inherits its network position: inside the perimeter, with access to internal resources, often without triggering endpoint detection.

This analysis combines two independent datasets to demonstrate that convergence. Tenable contributes exposure telemetry from the Tenable One Exposure Management Platform, covering thousands of customer containers and measuring what edge infrastructure is deployed, what is vulnerable, and how long it remains unpatched. This dataset extends Tenable Research’s ongoing analysis of edge device exposure trends, including the remediation telemetry Tenable contributed to the 2025 and 2026 Verizon DBIR reports. SentinelOne contributes findings from its digital forensics and incident response (DFIR) practice, documenting which threat actors actually exploit which vulnerabilities, observed firsthand inside compromised environments. Neither dataset was built for this analysis; each was constructed independently for different operational purposes.

The finding that makes this analysis compelling is not about any single CVE or any single actor. It is the structural convergence: two independent observation systems, looking at the problem from opposite sides, arrive at the same conclusion about which vendor surfaces are under persistent, broad exploitation, and by whom. (For how each dataset was built and scored, see the Methodology appendix below.)

What’s Exposed: The Vulnerability Surface

Tenable’s exposure telemetry provides the vulnerability-side view. All exposure metrics reported here use container-grain measurement: the percentage of customer environments (organizational containers) with at least one asset vulnerable to a given CVE as of Aug. 15, 2026, relative to total exposed containers over the preceding 14-month period. This measures breadth of organizational exposure to edge-product vendor vulnerabilities, not raw asset counts, across the sampling window.

This section covers 15 vendors in three groups: seven confirmed in both independently compiled corpora, two confirmed in SentinelOne’s casework but absent from Tenable’s attributed corpus, and six “candidate” vendors surfaced by a broader screen of Tenable telemetry. The candidates are not part of the convergence finding, but two, F5 and Zimbra, show broader customer exposure than most of the confirmed seven, so omitting them would understate the breadth of at-risk edge infrastructure.

F5 and Citrix lead for different reasons. Among vendors with statistically robust sample sizes, F5 customers are the most broadly exposed: 53.8% of 2,784 monitored customer environments running F5 products have at least one actively exploited CVE present. Citrix customers show the slowest remediation patterns: a median of 461 days to patch, with 71% of affected environments still carrying unpatched Citrix CVEs after a full year. F5 leads on scale of exposure; Citrix leads on persistent exposure.

The mid-pack is flatter than expected. Check Point (18.6%), Ivanti (24.1%), Fortinet (24.9%), and Citrix (28.8%) cluster within a 10-point band at container-grain. Fortinet, which dominates headlines, is mid-pack by this measure.

Thin-sample vendors show extreme rates but require caution. Juniper (91.7%), VMware (75.0%), Palo Alto Networks (69.2%), and Cisco (56.2%) all show container-exposure proportions above 50%, but each has fewer than 50 in-sample containers. These statistics are real directional signals, but should be considered within the context of the relatively low sample size.

Serial exploitation is structural. Two clean serial-exploitation sequences appear in the dataset: Ivanti EPMM (approximately 8.5 months between successive exploited CVEs) and Ivanti Connect Secure (approximately 13 months). Same product line, new vulnerability, repeat exploitation. Tenable Research has published advisories on both Ivanti exploitation sequences, tracking each CVE from initial disclosure through active exploitation, and the exposure data here extends that analysis with organizational remediation timelines not available at the time of the original advisories. The next one is coming.

Who Exploits What: The Threat Actor Landscape

This is not a targeted effort by a specific group. Edge infrastructure is a core focal point of attack across a broad range of threat actors and nexus categories. The combined Tenable-SentinelOne corpus documents exploitation by actors spanning five nexus categories: China, Russia, DPRK, Iran, and criminal (financially motivated). All five categories independently target the same vendor surfaces. Every attribution in the corpus is bucketed into one of three confidence tiers derived from a five-dimensional rubric evaluating attribution directness, evidence provenance, recency, exploitation role, and source corroboration. Confidence tiers, from high to low, are: DIRECT, TECHNIQUE-ALIGNED, or INFERRED.

Actor density scales with vendor exposure. Fortinet products face the broadest actor surface: 29 distinct threat actors across five nexus categories. Citrix follows with 22 actors across five categories, Ivanti with 19 across four, Palo Alto Networks with nine across three, and Check Point with six across two. Every focal vendor has confirmed exploitation from multiple nexus categories. No single vendor’s exposure is attributable to a single adversary group.

China-nexus actors are the highest-confidence case study, appearing across nine vendors in the corpus, with four DIRECT-tier attributions from the scored dataset alone. But the analytical value here is not that China targets edge devices. That is well established. The value is that China, Russia, DPRK, Iran, and ransomware operators all target the same edge devices, as the examples above illustrate. The governed attribution methodology is what allows this claim to be made with precision: we can distinguish confirmed multi-nexus convergence (DIRECT-tier evidence on both sides) from assessed convergence (INFERRED, requiring corroboration).

What Incident Response Sees That Telemetry Can’t

Exposure data shows which appliances are reachable, vulnerable, and unpatched. Incident response looks at what happened when attackers got in: what they accessed, what they took, and where they went next. In SentinelOne DFIR cases involving edge infrastructure, attackers used credentials stored on the appliances and the access those appliances already had to reach internal systems.

Credential Theft from Edge Appliances

Across three engagements, threat actors reached the management plane of FortiGate appliances and created rogue administrative accounts. In two, they also exported device configurations and extracted credentials that could be used to move further into the network. Two of these three are documented in detail in FortiGate Edge Intrusions.

In one of those two, the exported configuration contained LDAP bind credentials for a directory service account. The account was later used in the environment. A few hours later, the threat actor added two computers to the domain. Neither had a Service Principal Name, which is unusual for a legitimate domain join. The mS-DS-CreatorSID attribute on both accounts pointed back to the stolen service account.

We saw similar activity on an Ivanti Cloud Services Appliance in late 2024. A China-nexus actor chained CVE-2024-8963 with CVE-2024-8190 before the first public disclosure in the chain. After gaining access, the actor collected SSH keys and other stored credentials. That engagement is documented as Activity F in Follow the Smoke.

These appliances did not provide conventional endpoint telemetry. We had to follow the activity into authentication records, newly created Active Directory objects, and the later use of credentials taken from the appliances.

When the Initial Access Vector Cannot Be Confirmed

In December 2025, Fortinet disclosed CVE-2025-59718, an authentication bypass in its FortiCloud SSO integration affecting FortiOS and other products. Several weeks later, Fortinet disclosed CVE-2026-24858. This second flaw allowed an attacker with a FortiCloud account and a registered device to log into devices belonging to other customers when FortiCloud SSO was enabled.

That overlap mattered in one of the three engagements above. The appliance was running a version affected by both CVEs, but the available logs did not show when or how the attacker first gained access. The earliest retained malicious activity showed a rogue local administrator account. A few minutes later, a domain administrator authenticated from the appliance’s VPN address pool. Exposure data showed that the appliance had been vulnerable to both CVEs, but that alone did not establish how it was compromised. We treated both as possible, not confirmed, initial-access vectors.

Abuse of Trusted Management Access

In one SentinelOne DFIR engagement involving a FortiManager appliance, CVE-2024-47575 allowed an unauthorized device to register with the appliance through the management protocol. Two log entries, seconds apart, recorded the rogue registration and the settings change that followed. The threat actor then staged an archive of managed-device configurations that could expose credentials, addresses, and details about the network. The actor had been present for about a month before the customer detected the activity.

In a separate engagement, a threat actor chained SQL injection, pass-the-hash authentication, and authentication bypass against an internet-facing SonicWall GMS console (CVE-2023-34133, CVE-2023-34132, and CVE-2023-34124). The actor created administrative accounts on a platform operated by a managed service provider, then used existing shared access to enter multiple customer environments. Most of the resulting traffic was advertising-related, leading us to assess that the infrastructure was being used for click fraud.

The actors were after different things. One collected configuration data and information about the network. The other used the access to turn systems across several environments into proxies. In both cases, the actor inherited the access that the organization had already granted to the management platform. These cases show the difference between the two views: exposure telemetry finds the vulnerable device, while incident response shows what was taken from it and where the attacker went next.

What to Do About It

Patch F5 and Citrix edge devices immediately. These two vendors combine the highest exposure rates with the slowest remediation timelines across statistically robust samples. Look for strategies to reduce the remediation time, particularly for weaponized CVEs. Additionally, reduce the attack surface by minimizing the enabled feature set on these devices and aim for defense in depth by running endpoints in protect mode to limit lateral movement opportunities.

Audit Ivanti Connect Secure and EPMM deployments. Serial exploitation on observed 8.5 to 13-month cycles means the next exploitable CVE in these product lines is a question of timing, not probability. Organizations running Ivanti edge products should assume they will face a new actively exploited vulnerability within the next year and plan patching capacity accordingly.

Implement edge-device-specific patch SLAs. The delayed remediation paradox demonstrates that general priority frameworks do not translate into faster patching on the devices that sit at the network boundary. Edge devices and network infrastructure warrant dedicated remediation timelines that are shorter than the organizational default and commensurate with the elevated risk.

Treat edge device exposure as a cross-signal priority. Attribution, severity, and exposure volume identify different CVEs as “top priority.” Organizations need all three signals for complete coverage. A vulnerability management program that prioritizes exclusively by CVSS will systematically underweight CVEs with strong exploitation evidence but modest severity scores, and vice versa. The Tenable One Exposure Management Platform enables this cross-signal approach by combining vulnerability severity, exposure intelligence, asset context, and exposure data into a unified prioritization view.

Identifying Affected Systems

Tenable customers can use the Tenable Vulnerability Watch dashboard to monitor classifications for all CVEs discussed in this analysis. A list of Tenable plugins for the vulnerabilities discussed in this analysis can be found on the individual CVE pages at tenable.com/cve as they are released. This link displays all available plugins for each vulnerability, including upcoming plugins in our Plugins Pipeline.

Get more information

• Tenable Vulnerability Watch
• SentinelOne Threat Research (PurpleHaze, SonicWall SMA1000)
• FortiGate Edge Intrusions – SentinelOne
• Follow the Smoke – SentinelOne
• CISA Known Exploited Vulnerabilities Catalog

Join Tenable’s Research Special Operations (RSO) Team on Tenable Connect for further discussions on the latest cyber threats.

Learn more about Tenable One Exposure Management Platform, the exposure management platform for the modern attack surface.

Appendix: Methodology and Corpus Construction

How the corpus was built. Tenable’s 33-CVE corpus was derived by combining and deduplicating vulnerabilities with the highest exploitation volume and broadest actor adoption; SentinelOne validated CVEs across 14 vendors, and their 66-CVE landscape view reflects 12 months of DFIR casework with false positives removed. Combined, the two datasets identify 82 distinct CVEs, 17 of which appear in both. Layered on top of these sources is a governed attribution corpus of 93 CVE-actor pairs spanning approximately 39 named threat actors and five nexus categories.

Why Tenable tracks these CVEs. Tenable’s set comes out of exposure management. A CVE enters it through the Vulnerability Watch program, which classifies vulnerabilities under active or likely to be exploited, and is additionally scored with a Vulnerability Priority Rating (VPR). The question being answered is prescriptive: of everything actually deployed across customer environments, what should be prioritized and patched first? Threat-actor attribution is layered on afterward from definitive and confidence-scored sources (e.g., Federal cybersecurity advisories).

Why SentinelOne tracks these CVEs. SentinelOne’s set comes from the opposite direction: incident response. A CVE earns its place in their 12-month DFIR landscape because responders found it used in a real intrusion — the initial access vector in a case someone called them about. The question being answered is forensic: what happened here, and who did it? Coverage is shaped by who engaged them, not by install base.

What “overlap” means here. Overlap was measured at two levels, and the answer changes sharply depending on which level you use.

At the level of the individual vulnerability, the two sets barely intersect. Only 17 of 82, or 21%, of CVEs are common to both. Tenable and SentinelOne are, for the most part, not looking at the same vulnerabilities. However, the datasets converge at the product level. Eleven of the 14 vendors in Tenable’s focal CVE set appear in SentinelOne’s 12-month DFIR landscape – a 79% convergence: Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, VMware, Microsoft, Oracle, CrushFTP, and Meta’s React framework. That’s 79% convergence at the vendor level against 21% at the CVE level. Three vendors did not conform: Apache and SAP were absent from SentinelOne’s casework, and the one Progress case they worked on was closed as a false positive. Narrow the comparison to edge and remote-access infrastructure specifically, and the convergence is a perfect 100%. Tenable’s corpus independently identified seven edge vendors (i.e., Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, and VMware). All seven appear in SentinelOne’s casework. Two teams, working from unrelated evidence for unrelated purposes, arrived at the same seven vendors while sharing roughly one CVE in five.

Why the distinction matters. “Different vulnerabilities, same vendors” is not a weaker version of “same vulnerabilities.” It is a different and more actionable claim. Had both datasets converged on the same individual CVEs, the story would be that a specific handful of vulnerabilities is being widely exploited: patch those and the problem shrinks. What the data actually shows is that state-sponsored and criminal operators are independently arriving at the same small set of edge and remote-access product vendors, then finding their own separate ways in. The durable target is the vendor attack surface. Patching this quarter’s Ivanti CVE does not remove Ivanti from anyone’s target list.

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third party.

The question has shifted. Security leaders spent several years debating whether AI would reshape security operations. That debate has settled. Now the conversation is about pace. How fast can the foundation be built, and what do organizations that moved early have to show for it?

For the second year, SentinelOne® commissioned 451 Research to survey 611 North American cybersecurity decision-makers and practitioners on the state of security operations strategy. The results confirm what we’ve been building toward, and they surface a finding that should recalibrate how most security leaders sequence their AI investments.

The Returns Didn’t Wait for the Roadmap

Many product roadmaps assume a clear sequence and start with building toward higher maturity first with returns following. The data shows that AI is running ahead of schedule.

Nearly all organizations surveyed (96%) are still operating AI at the earliest maturity levels:

• Level 1: Basic monitoring; triage specialist/alert analyst
• Level 2: More senior triage analyst / basic incident responder and investigator

By most measures, AI adoption in the SOC is still early. And yet, 99% of those same organizations already report improvements in incident response and remediation.

The numbers are consistent. Early-stage AI (chatbots handling initial alert triage, automated tools sorting true positives from noise) is delivering before organizations reach advanced maturity. The gap between where most organizations are and what they are already getting is real and consistent across survey respondents.

Organizations waiting for higher AI maturity before building the supporting infrastructure are running the sequence backward. The returns are available now. The foundation built today determines how far those returns scale.

Platformization Has Reached A Verdict

The organizations accelerating AI adoption are also the ones consolidating onto platforms. A platform-oriented security architecture means moving from siloed, specialized tools to an integrated stack built on a foundation that coordinated AI decision-making can actually run on, and one that lets each new capability compound on the last.

The platformization numbers from this year’s survey are clear. 82% of organizations describe themselves as platform-oriented, a 13-point jump in a single year, and 94% expect to be there within three years.

A common assumption is that platform adoption means replacing specialized tools. The data complicates that picture. The same technologies most frequently deployed as standalone tools (EDR, SIEM, CNAPP) are also the top anchors for integrated platforms. Organizations typically start with one of these and expand outward. What changes is the common data layer that enables coordinated AI decision-making, serving as the connective tissue underneath.

Platformization is not coincidental with AI’s emergence. Agentic AI needs connected, continuously updated data to accurately reason across signals and take autonomous action. Fragmented architectures, where telemetry is siloed and pipelines require manual effort, cannot support AI-driven SOC operations at scale. Platform adoption and AI adoption are converging because AI’s data requirements have made integration a structural necessity.

The survey makes the infrastructure connection an explicit one. The top-cited benefit of investing in a data lake for SecOps is supporting AI-driven SOC workloads and agents. Organizations that built the data foundation early have already cleared the barrier stalling others. Those who haven’t, face a prerequisite gap, and the distance is widening rapidly. Architectural readiness is the variable that determines how far AI investments can scale.

Job Satisfaction Is Rising

Every discussion of AI in the SOC centers on detection and response metrics. This report has those too, but there is a finding that security leaders managing attrition should weigh: analyst burnout is declining.

As AI handles repetitive, high-volume triage work, analysts report rising job satisfaction. The role is shifting away from processing an endless queue and toward investigation, threat hunting, and judgment-intensive work. In a market where SOC analyst turnover remains a persistent operational cost, that shift carries real dollar value.

The analyst role evolves, becoming more strategic and more consequential.

A New Attack Surface

The same AI systems changing how SOCs operate are also creating new targets. Adversaries are already probing AI infrastructure including agents, data pipelines, model endpoints, and the governance gaps that emerge when controls lag behind adoption. The report surfaces this tension clearly: Organizations are deploying AI faster than they are securing it.

An AI agent with misconfigured access or an unmonitored data pipeline is an exposure. Securing the AI infrastructure that powers the SOC is happening alongside deployment, whether organizations have planned for it or not. Those without a clear governance posture are accepting risk that may not be priced into their AI investment case.

The potential of GenAI and agentic AI in the SOC is already being realized. The organizations that capture it fully are those building governance alongside deployment. The platform that runs the Autonomous SOC and the platform that secures it are, increasingly, the same platform.

SentinelOne’s Vision: The Autonomous SOC

Everything the report surfaces, from AI returns arriving before maturity to platform consolidation to the improving analyst experience, points to how these are expressions of the same shift. The foundation that enables early AI returns is the same one that determines how far those returns scale, how capable analysts become, and how well the security of AI itself is governed.

The findings align with how SentinelOne has defined the path to autonomous security operations: A progression from AI-assisted triage at early maturity levels to increasingly autonomous investigation, threat hunting, and response, with humans in strategic and governing roles. The report validates that the market is moving through exactly that sequence. Organizations that understand the architecture behind it (the platform integration, the common data layer, the governance controls) are positioning themselves to capture returns at every stage rather than waiting for the destination.

The full 451 Research report goes further into detail, covering what progression looks like at each maturity level, the specific barriers organizations are encountering, and the data behind each finding in full.

Read the full 451 Research report to learn more about how AI is reshaping cybersecurity.

Third-Party & Intellectual Property Disclaimers

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third party.

This blog may include discussion of unreleased services or features. Any unreleased services or features referenced here are still in development and subject to change. Customers should make their purchase decisions based upon features that are currently available.

AI agents have made their way into virtually every layer of your environment. They run in the apps your employees adopt, on the endpoints where agents execute code, as users with access privileges those agents borrow, and in the cloud workloads that scale them. The platform that you trust to secure your endpoints is already already covering where AI operates today.

Here is the through-line that makes this one problem instead of four. Every AI attack starts as an interaction and ends as an action. A prompt gets manipulated, an agent gets tricked, and the damage lands on a host, reaches into an identity, or moves through the cloud. The tools that treat each surface as a separate product hand you fragments. SentinelOne treats them as one chain.

How SentinelOne Defends the Agentic Stack Today

Employee AI use is where the risk quietly enters. Your people are already using AI tools you never sanctioned, through browser, IDE, and API-connected apps and agentic AI tools. SentinelOne discovers that shadow AI use across browsers, IDEs, and copilots, highlights which tools and models are in play and governs it with policy. It keeps confidential data, PII, and secrets from reaching untrusted models, and it stops prompt injection and jailbreaks aimed at the tools you build. Legacy DLP reads patterns; this reads context, which is the only way to catch an attack aimed at AI systems that behave in a non-deterministic way.

The agent layer is where AI stops advising and starts acting. An employee’s prompt sends text. An agent sends commands, holds credentials, calls APIs, and chain actions without a human approving each step. That makes them non-human identities with standing access. SentinelOne governs that access. It inventories the agents and MCP servers already operating and scores what each one can reach and holds every agent to the privileges its task requires. Then it inspects the tool calls themselves, so an injected instruction gets blocked at the moment it would execute. What gets executed lands in a searchable record, and the same enforcement doubles as a kill switch.

While governance decides what an agent is allowed to do, the endpoint is where you find out what it actually did.

The endpoint is where agents execute. This is the frontier, and where SentinelOne has protected customers for over a decade. Our behavioral engine judges what a process does, not what it claims to be. That is how we caught QUIETVAULT – malware that spins up AI agents in “yolo” mode to exfiltrate secrets to GitHub. It is how we autonomously stopped the LiteLLM supply chain attack, where adversaries weaponized the Claude CLI to install a malicious payload. It is how we surfaced a DLL side-loading attack hidden inside an AI tool installer. Real detections, on the endpoint, today. Agents run on the host, and so do we.

The identity is where a hijacked agent runs next. Picture an employee’s AI coding agent that gets hijacked mid-task. It spawns a shell and reaches for cached credentials and cloud session tokens, trying to stop being a process and start being the user. That pivot to identity is what unlocks lateral movement, and it is where most AI attacks are headed. SentinelOne meets the move. It secures human and non-human identities alike, and seeds the environment with decoy credentials and honeytokens no legitimate user ever touches. The instant the hijacked agent grabs one, the trap trips, and Identity responds by forcing an MFA re-challenge, disabling the account, or isolating the host. Authorization at login is not enough. Access gets validated against behavior and pulled at runtime.

The cloud is where AI workloads scale. Consider an internal AI agent running in a Kubernetes cluster with standing access to a customer database. Security teams keep asking the same question about deployments like this. Where is the model connecting, and who is it talking to? SentinelOne answers with eBPF-native runtime protection that judges how the workload actually behaves, and flags the moment that inference service reaches an endpoint it has never touched before. It covers the control plane the deployment depends on, the secrets it reads, the pipelines it runs through, and the data it can access. Defending the AI you build takes more than watching it, it takes action on the workload in real time.

SentinelOne’s Singularity Platform Advantage

Each of these surfaces matters on its own. What closes the kill chain is following an attack across them without losing it at the handoff. This is where a single platform earns its keep. AI telemetry already streams into the Singularity™ Data Lake, alongside the endpoint data the platform has correlated for years. As identity and cloud signals join that same view, an analyst follows one attack from first prompt to final action, without stitching logs across six tools at two in the morning. A manipulated prompt, the process it spawns, the credential it reaches for, and the cloud resource it targets read as one story rather than six disconnected alerts.

Detection that only watches is observation. Runtime action is protection. When the Singularity Platform acts, autonomous response blocks the execution, rolls back the change, and revokes the access at the point of impact, without a human relaying orders between consoles. This is the difference between whether an attack is stopped or just gets logged.

That is the case for securing AI inside a platform built for autonomous runtime response. We are not adding a console to chase AI, we are extending the one already deployed where your agents run.

Questions to Ask When Assessing Your AI Security Options

When evaluating AI security, ask yourself three things.

• Does the solution protect the endpoint where agents actually execute, or is it a roadmap item?
• When a hijacked agent pivots to credentials and the cloud, does that telemetry land in the same platform, or are you manually connecting dots across three dashboards?
• Can the solution act at the moment of execution, or only tell me what already happened?

SentinelOne protects the surfaces where AI runs today. This includes the apps your employees use, the agents they deploy, the endpoints where agents execute, the identities they borrow, and the cloud where they scale. One platform, built for autonomous response. While AI has changed the attack, it does not have to change your architecture.

See it for yourself. Talk to our team about securing AI across your endpoints, identities, cloud, and the AI apps your employees already use, all from the platform you run today. Contact SentinelOne today.

Third-Party Trademark Disclaimer:

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third-party.

The Good | Authorities Launch New Operations Against Cybercrime Networks & Supply Chain Attackers

Operation Jackal IV, coordinated by INTERPOL across 22 nations, has led to the arrest of 58 individuals and the identification of over 200 suspects linked to West African cybercrime networks. The joint action successfully dismantled elements of the Black Axe syndicate, which orchestrates global romance, investment, and business email compromise (BEC) scams. Law enforcement agencies across South Africa, Argentina, and Romania also disrupted major Crime-as-a-Service (CaaS) providers, freezing millions of dollars in illicit financial assets.

The FBI, collaborating with the DoJ, have disrupted the global QScan and QTRouter hacking platforms operated by Chinese state-sponsored threat actors. The group QTFY, which maintains direct ties to China’s military and intelligence services, used these compromised IoT botnets to mask cyber espionage traffic targeting critical U.S. networks, including the Federal Reserve and NASA. Law enforcement successfully seized the core command-and-control (C2) domains hardcoded within the malicious frameworks.

From the U.S. Treasury is a new operation dubbed Economic Outcast, imposing sweeping sanctions on five Mabna Institute members and nearly 60 Iran-linked entities. Under the direction of Iran’s Ministry of Intelligence and Security (MOIS), the attackers breached multiple American critical infrastructure organizations, state governments, and defense contractors. These state-sponsored actors then exfiltrated datasets, executed high-value cryptocurrency heists, and now face federal indictments alongside a $10 million dollar reward for information leading to their arrest.

The Australian Federal Police (AFP) have arrested and charged two individuals for principal roles in TeamPCP, a cybercrime syndicate. The group systematically compromised trusted open-source projects, including Trivy, Checkmarx KICS, and LiteLLM, by stealing developer credentials and distributing backdoored software updates across major ecosystem release channels. This massive software supply chain campaign potentially compromised organizations worldwide and facilitated the unauthorized theft of hundreds of thousands of credentials.

The Bad | ‘NovaCookies’ Phishing Toolkit Exploits DocuSign Services to Steal Session Tokens

Security researchers have disclosed details of NovaCookies, a subscription-based phishing platform that systematically targets corporate networks to steal authenticated Microsoft 365 sessions. Operating as an Adversary-in-the-Middle (AitM) proxy, this malicious toolkit is advertised on Telegram for $320 monthly. The campaigns actively compromise hundreds of organizations across several nations, including the U.S., the U.K., Germany, and the U.A.E.

To establish a foothold, attackers distribute counterfeit document-sharing lures within genuine DocuSign notifications. Styled as a share notice, the decoy claims an accounting department shared a remittance-advice PDF and invites the recipient to open it. Since these notifications originate from legitimate servers, they bypass standard sender-authentication checks and reputation filters. The malicious link is embedded inside the shared document, below the inspection layer of most security gateways. Once clicked, the attack uses an OAuth error-redirect technique to guide the browser through legitimate Microsoft or Google endpoints before routing traffic to the phishing infrastructure. This transition ensures every intermediate step appears trustworthy until the user reaches the proxy.

NovaCookies is a variant of the Sneaky2FA platform, which operates on a centrally managed model where the operator hosts the infrastructure rather than individual affiliates. The kit offers customized flows targeting common identity providers. Affiliates register landing pages on .vu domains, utilizing deceptive, alternating-case subdomains like PwPt-sHaRe to masquerade as legitimate Microsoft portals. While these checks obscure the landing pages, the proxy relays credentials and multi-factor authentication (MFA) codes in real time to Microsoft. Because each individual hop of the attack chain appears legitimate, security analysts emphasize that the browser remains the critical intersection where these events converge.

The Ugly | Threat Actors Deploy Spark RAT to Target Cambodian Organizations

A recently uncovered campaign is targeting both individuals and organizations in Cambodia with Spark RAT, which functions as a Go-based, open-source remote access trojan. Distributing compressed archives through targeted phishing emails, the threat actors deploy diverse lures, including Cambodian government notices, public health announcements, and dental records. The multi-stage attack sequence begins when a victim executes an Inno Setup installer, which initiates a dynamic link library side-loading chain using a signed Tencent application to deliver intermediate payloads.

To guarantee execution, the DLL loader performs timing-based anti-sandbox checks to detect virtual environment delays and scans running processes in an attempt to weaken its permissions. The loader then decrypts shellcode hidden within an embedded PNG file to run a second stager that determines whether the malware operates with SYSTEM privileges. If these elevated rights are present, the malware proceeds directly to inject mode. Otherwise, it configures a Windows service for local persistence. Ultimately, the stager injects malicious shellcode into the legitimate vssvc.exe process, monitoring execution to re-inject the payload if terminated.

The intrusion chain utilizes the Bring Your Own Vulnerable Driver (BYOVD) technique that abuses a legitimate but vulnerable OPSWAT AppRemover driver, ardrv.sys, to escalate privileges and neutralize security programs. Operating under CVE-2026-36425, this driver enables the malware to terminate active security processes, including Microsoft Defender, Huorong Internet Security, and Tencent PC Manager. The program also patches Antimalware Scan Interface and Event Tracing for Windows, executes user-mode termination of security tools, and injects Spark RAT into ctfmon.exe. Although operational tactics and driver usage closely mirror the Chinese-speaking Silver Fox syndicate, analysts classify the campaign as an unattributed cluster due to the absence of shared infrastructure, certificates, or code reuse.

It is the shared attack surface where state-sponsored threat actors and financially motivated criminal groups independently converge — not the province of a single adversary category, and not exclusively a nation-state problem, despite two years of headlines about China-nexus actors targeting Ivanti, Fortinet, and Palo Alto Networks. The data here tells a different and much broader story. One focused on vendors vs CVEs.

Key Takeaways

• Two independent observation systems, Tenable exposure telemetry across thousands of customer containers and SentinelOne DFIR casework across 66 CVEs, converge 79% on the same vendor attack surfaces despite minimal CVE-level overlap.
• Twelve CVEs in the combined dataset have confirmed multi-nexus attribution: state-sponsored and criminal actors independently exploiting the same vulnerability, across five nexus categories (China, Russia, DPRK, Iran, ransomware).
• The exposure picture is flatter than the headlines suggest: Fortinet, the vendor most associated with edge-device attacks in the press, sits mid-pack on container-grain exposure (25%) — well behind F5 (54%) and in a tight 10-point band with Check Point, Ivanti, and Citrix.
• 54% of customer environments running F5 products have at least one exposed, actively-exploited CVE; Citrix customers show the slowest remediation patterns at 461 days median time to patch.
• Remediation complexity, particularly of high priority CVEs, leads to a statistically significant 24-day remediation gap, leaving large windows of opportunity for attackers.
• The same product lines get hit again and again: Ivanti EPMM and Ivanti Connect Secure each show a newly exploited CVE roughly every 8.5 to 13 months.
• Leverage multiple defense-in-depth strategies: patch as quickly as possible, but also minimize the attack surface (feature-set minimization) and run endpoints in protect mode to better stop lateral movement from attacks that gain initial access.

The Convergence is the Story

Twelve CVEs in the combined dataset have confirmed multi-nexus attribution: state-sponsored and criminal actors independently exploiting the same vulnerability, across five nexus categories. Four examples illustrate the pattern:

CVE	Product	Actors (Nexus)	Significance
CVE-2026-15409	SonicWall SMA1000	UTA0533 (unattributed) + INC Ransomware	Espionage-to-ransomware succession on an active zero-day
CVE-2023-42793	JetBrains TeamCity	APT29 (Russia) + Lazarus (DPRK)	Two state-sponsored actors from different nations on the same CVE
CVE-2024-3400	PAN-OS GlobalProtect	UTA0218 (China) + INC Ransomware	China-nexus zero-day reused by ransomware operators
CVE-2024-24919	Check Point Quantum	PurpleHaze (China) + Fox Kitten (Iran)	China and Iran independently exploiting the same gateway vulnerability

The remaining eight confirmed multi-nexus CVEs span Fortinet, Citrix, Cisco, and Ivanti product lines. State-sponsored actors and ransomware operators are not operating in separate vulnerability ecosystems. They share the same entry points into the same products. The breadth of the convergence, not any single actor’s activity, is the finding.

That pattern holds across the full combined analysis. Three conclusions emerge:

Vendor attack surfaces are the persistent exploitation target. The same eleven vendors (i.e., Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, VMware, Microsoft, Oracle, CrushFTP, and Meta’s React framework) appear in both observation systems at 79% convergence, and all seven edge-product vendors converge. Serial exploitation timing on Ivanti products shows the vulnerability-to-exploitation pipeline refreshing at 8.5 to 13-month intervals on the same product lines. This is structural, not episodic. Patching the current CVE does not remove the vendor attack surface from the threat landscape.

State-sponsored and ransomware actors converge structurally. Twelve CVEs with confirmed multi-nexus attribution span all five nexus categories and cross the state-criminal divide. Defending against one actor category on edge devices necessarily requires defending against all of them, because the attack surface is shared. An organization that patches only for nation-state TTPs leaves itself exposed to ransomware operators exploiting the same vulnerability, and vice versa.

High-priority CVEs take more time to remediate, not less. Across Tenable’s 238-CVE high-priority list, high-priority CVEs carry a median remediation time of 146 days, compared to 122 days for all other CVEs — a 24-day gap that is statistically significant. The edge-appliance-specific subset (52 CVEs) shows a consistent 8-day gap in the same direction, which was not statistically significant, but suggests the direction may hold for edge appliances too. External data corroborates the pattern: the 2026 Verizon Data Breach Investigations Report (DBIR) found that median patch time increased from 32 to 43 days year over year, even as exploitation overtook credential theft as the number one initial access vector, and the 2025 DBIR, which incorporated Tenable RSO’s remediation trend analysis across 17 edge-related CVEs, found that only 54% of edge device KEVs were fully remediated. The explanation is structural: edge devices are the network boundary, so patching a VPN gateway or firewall means downtime for every user behind it, and change management gates multiply. These devices also resist standard patching workflows because they do not run endpoint agents, often require firmware-level updates with manual validation, and frequently lack active support contracts. The result is that the devices most worth patching are operationally the hardest to patch — and as the high-priority queue grows (the 2026 DBIR reports 50% more critical vulnerabilities to patch than the prior year), everything on it waits longer.

Background

Edge and perimeter devices occupy a uniquely consequential position in enterprise architecture. VPN gateways, firewalls, remote access appliances, and application delivery controllers sit at the boundary between trusted and untrusted networks. They are very often the first component an attacker touches and, for many organizations, the last component that gets patched. When one of these devices is compromised, the attacker inherits its network position: inside the perimeter, with access to internal resources, often without triggering endpoint detection.

This analysis combines two independent datasets to demonstrate that convergence. Tenable contributes exposure telemetry from the Tenable One Exposure Management Platform, covering thousands of customer containers and measuring what edge infrastructure is deployed, what is vulnerable, and how long it remains unpatched. This dataset extends Tenable Research’s ongoing analysis of edge device exposure trends, including the remediation telemetry Tenable contributed to the 2025 and 2026 Verizon DBIR reports. SentinelOne contributes findings from its digital forensics and incident response (DFIR) practice, documenting which threat actors actually exploit which vulnerabilities, observed firsthand inside compromised environments. Neither dataset was built for this analysis; each was constructed independently for different operational purposes.

The finding that makes this analysis compelling is not about any single CVE or any single actor. It is the structural convergence: two independent observation systems, looking at the problem from opposite sides, arrive at the same conclusion about which vendor surfaces are under persistent, broad exploitation, and by whom. (For how each dataset was built and scored, see the Methodology appendix below.)

What’s Exposed: The Vulnerability Surface

Tenable’s exposure telemetry provides the vulnerability-side view. All exposure metrics reported here use container-grain measurement: the percentage of customer environments (organizational containers) with at least one asset vulnerable to a given CVE as of Aug. 15, 2026, relative to total exposed containers over the preceding 14-month period. This measures breadth of organizational exposure to edge-product vendor vulnerabilities, not raw asset counts, across the sampling window.

This section covers 15 vendors in three groups: seven confirmed in both independently compiled corpora, two confirmed in SentinelOne’s casework but absent from Tenable’s attributed corpus, and six “candidate” vendors surfaced by a broader screen of Tenable telemetry. The candidates are not part of the convergence finding, but two, F5 and Zimbra, show broader customer exposure than most of the confirmed seven, so omitting them would understate the breadth of at-risk edge infrastructure.

F5 and Citrix lead for different reasons. Among vendors with statistically robust sample sizes, F5 customers are the most broadly exposed: 53.8% of 2,784 monitored customer environments running F5 products have at least one actively exploited CVE present. Citrix customers show the slowest remediation patterns: a median of 461 days to patch, with 71% of affected environments still carrying unpatched Citrix CVEs after a full year. F5 leads on scale of exposure; Citrix leads on persistent exposure.

The mid-pack is flatter than expected. Check Point (18.6%), Ivanti (24.1%), Fortinet (24.9%), and Citrix (28.8%) cluster within a 10-point band at container-grain. Fortinet, which dominates headlines, is mid-pack by this measure.

Thin-sample vendors show extreme rates but require caution. Juniper (91.7%), VMware (75.0%), Palo Alto Networks (69.2%), and Cisco (56.2%) all show container-exposure proportions above 50%, but each has fewer than 50 in-sample containers. These statistics are real directional signals, but should be considered within the context of the relatively low sample size.

Serial exploitation is structural. Two clean serial-exploitation sequences appear in the dataset: Ivanti EPMM (approximately 8.5 months between successive exploited CVEs) and Ivanti Connect Secure (approximately 13 months). Same product line, new vulnerability, repeat exploitation. Tenable Research has published advisories on both Ivanti exploitation sequences, tracking each CVE from initial disclosure through active exploitation, and the exposure data here extends that analysis with organizational remediation timelines not available at the time of the original advisories. The next one is coming.

Who Exploits What: The Threat Actor Landscape

This is not a targeted effort by a specific group. Edge infrastructure is a core focal point of attack across a broad range of threat actors and nexus categories. The combined Tenable-SentinelOne corpus documents exploitation by actors spanning five nexus categories: China, Russia, DPRK, Iran, and criminal (financially motivated). All five categories independently target the same vendor surfaces. Every attribution in the corpus is bucketed into one of three confidence tiers derived from a five-dimensional rubric evaluating attribution directness, evidence provenance, recency, exploitation role, and source corroboration. Confidence tiers, from high to low, are: DIRECT, TECHNIQUE-ALIGNED, or INFERRED.

Actor density scales with vendor exposure. Fortinet products face the broadest actor surface: 29 distinct threat actors across five nexus categories. Citrix follows with 22 actors across five categories, Ivanti with 19 across four, Palo Alto Networks with nine across three, and Check Point with six across two. Every focal vendor has confirmed exploitation from multiple nexus categories. No single vendor’s exposure is attributable to a single adversary group.

China-nexus actors are the highest-confidence case study, appearing across nine vendors in the corpus, with four DIRECT-tier attributions from the scored dataset alone. But the analytical value here is not that China targets edge devices. That is well established. The value is that China, Russia, DPRK, Iran, and ransomware operators all target the same edge devices, as the examples above illustrate. The governed attribution methodology is what allows this claim to be made with precision: we can distinguish confirmed multi-nexus convergence (DIRECT-tier evidence on both sides) from assessed convergence (INFERRED, requiring corroboration).

What Incident Response Sees That Telemetry Can’t

Exposure data shows which appliances are reachable, vulnerable, and unpatched. Incident response looks at what happened when attackers got in: what they accessed, what they took, and where they went next. In SentinelOne DFIR cases involving edge infrastructure, attackers used credentials stored on the appliances and the access those appliances already had to reach internal systems.

Credential Theft from Edge Appliances

Across three engagements, threat actors reached the management plane of FortiGate appliances and created rogue administrative accounts. In two, they also exported device configurations and extracted credentials that could be used to move further into the network. Two of these three are documented in detail in FortiGate Edge Intrusions.

In one of those two, the exported configuration contained LDAP bind credentials for a directory service account. The account was later used in the environment. A few hours later, the threat actor added two computers to the domain. Neither had a Service Principal Name, which is unusual for a legitimate domain join. The mS-DS-CreatorSID attribute on both accounts pointed back to the stolen service account.

We saw similar activity on an Ivanti Cloud Services Appliance in late 2024. A China-nexus actor chained CVE-2024-8963 with CVE-2024-8190 before the first public disclosure in the chain. After gaining access, the actor collected SSH keys and other stored credentials. That engagement is documented as Activity F in Follow the Smoke.

These appliances did not provide conventional endpoint telemetry. We had to follow the activity into authentication records, newly created Active Directory objects, and the later use of credentials taken from the appliances.

When the Initial Access Vector Cannot Be Confirmed

In December 2025, Fortinet disclosed CVE-2025-59718, an authentication bypass in its FortiCloud SSO integration affecting FortiOS and other products. Several weeks later, Fortinet disclosed CVE-2026-24858. This second flaw allowed an attacker with a FortiCloud account and a registered device to log into devices belonging to other customers when FortiCloud SSO was enabled.

That overlap mattered in one of the three engagements above. The appliance was running a version affected by both CVEs, but the available logs did not show when or how the attacker first gained access. The earliest retained malicious activity showed a rogue local administrator account. A few minutes later, a domain administrator authenticated from the appliance’s VPN address pool. Exposure data showed that the appliance had been vulnerable to both CVEs, but that alone did not establish how it was compromised. We treated both as possible, not confirmed, initial-access vectors.

Abuse of Trusted Management Access

In one SentinelOne DFIR engagement involving a FortiManager appliance, CVE-2024-47575 allowed an unauthorized device to register with the appliance through the management protocol. Two log entries, seconds apart, recorded the rogue registration and the settings change that followed. The threat actor then staged an archive of managed-device configurations that could expose credentials, addresses, and details about the network. The actor had been present for about a month before the customer detected the activity.

In a separate engagement, a threat actor chained SQL injection, pass-the-hash authentication, and authentication bypass against an internet-facing SonicWall GMS console (CVE-2023-34133, CVE-2023-34132, and CVE-2023-34124). The actor created administrative accounts on a platform operated by a managed service provider, then used existing shared access to enter multiple customer environments. Most of the resulting traffic was advertising-related, leading us to assess that the infrastructure was being used for click fraud.

The actors were after different things. One collected configuration data and information about the network. The other used the access to turn systems across several environments into proxies. In both cases, the actor inherited the access that the organization had already granted to the management platform. These cases show the difference between the two views: exposure telemetry finds the vulnerable device, while incident response shows what was taken from it and where the attacker went next.

What to Do About It

Patch F5 and Citrix edge devices immediately. These two vendors combine the highest exposure rates with the slowest remediation timelines across statistically robust samples. Look for strategies to reduce the remediation time, particularly for weaponized CVEs. Additionally, reduce the attack surface by minimizing the enabled feature set on these devices and aim for defense in depth by running endpoints in protect mode to limit lateral movement opportunities.

Audit Ivanti Connect Secure and EPMM deployments. Serial exploitation on observed 8.5 to 13-month cycles means the next exploitable CVE in these product lines is a question of timing, not probability. Organizations running Ivanti edge products should assume they will face a new actively exploited vulnerability within the next year and plan patching capacity accordingly.

Implement edge-device-specific patch SLAs. The delayed remediation paradox demonstrates that general priority frameworks do not translate into faster patching on the devices that sit at the network boundary. Edge devices and network infrastructure warrant dedicated remediation timelines that are shorter than the organizational default and commensurate with the elevated risk.

Treat edge device exposure as a cross-signal priority. Attribution, severity, and exposure volume identify different CVEs as “top priority.” Organizations need all three signals for complete coverage. A vulnerability management program that prioritizes exclusively by CVSS will systematically underweight CVEs with strong exploitation evidence but modest severity scores, and vice versa. The Tenable One Exposure Management Platform enables this cross-signal approach by combining vulnerability severity, exposure intelligence, asset context, and exposure data into a unified prioritization view.

Identifying Affected Systems

Tenable customers can use the Tenable Vulnerability Watch dashboard to monitor classifications for all CVEs discussed in this analysis. A list of Tenable plugins for the vulnerabilities discussed in this analysis can be found on the individual CVE pages at tenable.com/cve as they are released. This link displays all available plugins for each vulnerability, including upcoming plugins in our Plugins Pipeline.

Get more information

• Tenable Vulnerability Watch
• SentinelOne Threat Research (PurpleHaze, SonicWall SMA1000)
• FortiGate Edge Intrusions – SentinelOne
• Follow the Smoke – SentinelOne
• CISA Known Exploited Vulnerabilities Catalog

Join Tenable’s Research Special Operations (RSO) Team on Tenable Connect for further discussions on the latest cyber threats.

Learn more about Tenable One Exposure Management Platform, the exposure management platform for the modern attack surface.

Appendix: Methodology and Corpus Construction

How the corpus was built. Tenable’s 33-CVE corpus was derived by combining and deduplicating vulnerabilities with the highest exploitation volume and broadest actor adoption; SentinelOne validated CVEs across 14 vendors, and their 66-CVE landscape view reflects 12 months of DFIR casework with false positives removed. Combined, the two datasets identify 82 distinct CVEs, 17 of which appear in both. Layered on top of these sources is a governed attribution corpus of 93 CVE-actor pairs spanning approximately 39 named threat actors and five nexus categories.

Why Tenable tracks these CVEs. Tenable’s set comes out of exposure management. A CVE enters it through the Vulnerability Watch program, which classifies vulnerabilities under active or likely to be exploited, and is additionally scored with a Vulnerability Priority Rating (VPR). The question being answered is prescriptive: of everything actually deployed across customer environments, what should be prioritized and patched first? Threat-actor attribution is layered on afterward from definitive and confidence-scored sources (e.g., Federal cybersecurity advisories).

Why SentinelOne tracks these CVEs. SentinelOne’s set comes from the opposite direction: incident response. A CVE earns its place in their 12-month DFIR landscape because responders found it used in a real intrusion — the initial access vector in a case someone called them about. The question being answered is forensic: what happened here, and who did it? Coverage is shaped by who engaged them, not by install base.

What “overlap” means here. Overlap was measured at two levels, and the answer changes sharply depending on which level you use.

At the level of the individual vulnerability, the two sets barely intersect. Only 17 of 82, or 21%, of CVEs are common to both. Tenable and SentinelOne are, for the most part, not looking at the same vulnerabilities. However, the datasets converge at the product level. Eleven of the 14 vendors in Tenable’s focal CVE set appear in SentinelOne’s 12-month DFIR landscape – a 79% convergence: Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, VMware, Microsoft, Oracle, CrushFTP, and Meta’s React framework. That’s 79% convergence at the vendor level against 21% at the CVE level. Three vendors did not conform: Apache and SAP were absent from SentinelOne’s casework, and the one Progress case they worked on was closed as a false positive. Narrow the comparison to edge and remote-access infrastructure specifically, and the convergence is a perfect 100%. Tenable’s corpus independently identified seven edge vendors (i.e., Fortinet, Citrix, Ivanti, Palo Alto Networks, Cisco, Juniper, and VMware). All seven appear in SentinelOne’s casework. Two teams, working from unrelated evidence for unrelated purposes, arrived at the same seven vendors while sharing roughly one CVE in five.

Why the distinction matters. “Different vulnerabilities, same vendors” is not a weaker version of “same vulnerabilities.” It is a different and more actionable claim. Had both datasets converged on the same individual CVEs, the story would be that a specific handful of vulnerabilities is being widely exploited: patch those and the problem shrinks. What the data actually shows is that state-sponsored and criminal operators are independently arriving at the same small set of edge and remote-access product vendors, then finding their own separate ways in. The durable target is the vendor attack surface. Patching this quarter’s Ivanti CVE does not remove Ivanti from anyone’s target list.

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third party.

The question has shifted. Security leaders spent several years debating whether AI would reshape security operations. That debate has settled. Now the conversation is about pace. How fast can the foundation be built, and what do organizations that moved early have to show for it?

For the second year, SentinelOne® commissioned 451 Research to survey 611 North American cybersecurity decision-makers and practitioners on the state of security operations strategy. The results confirm what we’ve been building toward, and they surface a finding that should recalibrate how most security leaders sequence their AI investments.

The Returns Didn’t Wait for the Roadmap

Many product roadmaps assume a clear sequence and start with building toward higher maturity first with returns following. The data shows that AI is running ahead of schedule.

Nearly all organizations surveyed (96%) are still operating AI at the earliest maturity levels:

• Level 1: Basic monitoring; triage specialist/alert analyst
• Level 2: More senior triage analyst / basic incident responder and investigator

By most measures, AI adoption in the SOC is still early. And yet, 99% of those same organizations already report improvements in incident response and remediation.

The numbers are consistent. Early-stage AI (chatbots handling initial alert triage, automated tools sorting true positives from noise) is delivering before organizations reach advanced maturity. The gap between where most organizations are and what they are already getting is real and consistent across survey respondents.

Organizations waiting for higher AI maturity before building the supporting infrastructure are running the sequence backward. The returns are available now. The foundation built today determines how far those returns scale.

Platformization Has Reached A Verdict

The organizations accelerating AI adoption are also the ones consolidating onto platforms. A platform-oriented security architecture means moving from siloed, specialized tools to an integrated stack built on a foundation that coordinated AI decision-making can actually run on, and one that lets each new capability compound on the last.

The platformization numbers from this year’s survey are clear. 82% of organizations describe themselves as platform-oriented, a 13-point jump in a single year, and 94% expect to be there within three years.

A common assumption is that platform adoption means replacing specialized tools. The data complicates that picture. The same technologies most frequently deployed as standalone tools (EDR, SIEM, CNAPP) are also the top anchors for integrated platforms. Organizations typically start with one of these and expand outward. What changes is the common data layer that enables coordinated AI decision-making, serving as the connective tissue underneath.

Platformization is not coincidental with AI’s emergence. Agentic AI needs connected, continuously updated data to accurately reason across signals and take autonomous action. Fragmented architectures, where telemetry is siloed and pipelines require manual effort, cannot support AI-driven SOC operations at scale. Platform adoption and AI adoption are converging because AI’s data requirements have made integration a structural necessity.

The survey makes the infrastructure connection an explicit one. The top-cited benefit of investing in a data lake for SecOps is supporting AI-driven SOC workloads and agents. Organizations that built the data foundation early have already cleared the barrier stalling others. Those who haven’t, face a prerequisite gap, and the distance is widening rapidly. Architectural readiness is the variable that determines how far AI investments can scale.

Job Satisfaction Is Rising

Every discussion of AI in the SOC centers on detection and response metrics. This report has those too, but there is a finding that security leaders managing attrition should weigh: analyst burnout is declining.

As AI handles repetitive, high-volume triage work, analysts report rising job satisfaction. The role is shifting away from processing an endless queue and toward investigation, threat hunting, and judgment-intensive work. In a market where SOC analyst turnover remains a persistent operational cost, that shift carries real dollar value.

The analyst role evolves, becoming more strategic and more consequential.

A New Attack Surface

The same AI systems changing how SOCs operate are also creating new targets. Adversaries are already probing AI infrastructure including agents, data pipelines, model endpoints, and the governance gaps that emerge when controls lag behind adoption. The report surfaces this tension clearly: Organizations are deploying AI faster than they are securing it.

An AI agent with misconfigured access or an unmonitored data pipeline is an exposure. Securing the AI infrastructure that powers the SOC is happening alongside deployment, whether organizations have planned for it or not. Those without a clear governance posture are accepting risk that may not be priced into their AI investment case.

The potential of GenAI and agentic AI in the SOC is already being realized. The organizations that capture it fully are those building governance alongside deployment. The platform that runs the Autonomous SOC and the platform that secures it are, increasingly, the same platform.

SentinelOne’s Vision: The Autonomous SOC

Everything the report surfaces, from AI returns arriving before maturity to platform consolidation to the improving analyst experience, points to how these are expressions of the same shift. The foundation that enables early AI returns is the same one that determines how far those returns scale, how capable analysts become, and how well the security of AI itself is governed.

The findings align with how SentinelOne has defined the path to autonomous security operations: A progression from AI-assisted triage at early maturity levels to increasingly autonomous investigation, threat hunting, and response, with humans in strategic and governing roles. The report validates that the market is moving through exactly that sequence. Organizations that understand the architecture behind it (the platform integration, the common data layer, the governance controls) are positioning themselves to capture returns at every stage rather than waiting for the destination.

The full 451 Research report goes further into detail, covering what progression looks like at each maturity level, the specific barriers organizations are encountering, and the data behind each finding in full.

Read the full 451 Research report to learn more about how AI is reshaping cybersecurity.

Third-Party & Intellectual Property Disclaimers

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third party.

This blog may include discussion of unreleased services or features. Any unreleased services or features referenced here are still in development and subject to change. Customers should make their purchase decisions based upon features that are currently available.

The Good | U.S. Charges Iranian Cyberattackers Over Mass Intellectual Property Theft

The U.S. Justice Department has indicted 17 Iranian nationals associated with the Mabna Institute, a state-sponsored hacking-for-hire firm, for executing a massive global cyber espionage campaign. Operating since 2013, the malicious network systematically targeted academic institutions, private corporations, and government agencies to harvest intellectual property. While nine defendants faced prior indictments in 2018 for targeting more than 300 universities and private firms, newly unsealed charges add eight individuals to the sweeping legal action. Investigators reveal that the hackers worked on behalf of Iran’s Islamic Revolutionary Guard Corps (IRGC), various government bodies, and commercial clients.

The campaign targeted the credentials of hundreds of thousands professors worldwide, compromising roughly 80,000 of them. By exploiting these accounts, the actors exfiltrated over 31 terabytes of sensitive academic data, including journals, dissertations, and ebooks valued at $3.4 billion. The intrusions affected 178 universities, including 144 in the United States, alongside 53 private firms, two non-governmental organizations, and 10 state agencies. Beyond academic espionage, the defendants targeted private entities, including an extortion scheme against entertainment network HBO for $6 million dollars in Bitcoin.

The State Department announced rewards of up to $10 million for information leading to the apprehension of five key defendants and established an anonymous Tor network link to receive tips. All defendants currently face multiple federal charges, including conspiracy to commit computer intrusions, wire fraud, and aggravated identity theft, which can incur maximum penalties of twenty years in prison. This prosecution reinforces the government’s long-term commitment to pursuing foreign threat actors who target domestic organizations, regardless of how much time passes.

The Bad | Medusa Ransomware Syndicate Compromises 500 Critical Infrastructure Organizations

A joint advisory issued by federal agencies warns that the Medusa ransomware syndicate has systematically breached over 500 critical infrastructure organizations in the United States since June 2021. Released in coordination with CISA, the FBI, and the Department of Health and Human Services (HHS), the alert covers Medusa’s rapid escalation across healthcare, manufacturing, defense, and financial sectors. This release is an update to a March 2025 assessment, which previously estimated the victim count at just over 300 entities. Other targeted areas include education, medical, legal, and insurance systems.

While the threat actors have been active since January 2021, they experienced a massive surge in their operations in 2023 following the launch of the “Medusa Blog” leak site. Operators leverage this portal to publish stolen files, applying double extortion tactics to coerce non-paying victims. Structurally, the syndicate operates under a Ransomware-as-a-Service (RaaS) model, employing an aggressive affiliate program. Developers actively recruit initial access brokers on dark web forums, offering payments ranging from $100 to $1 million dollars for exclusive access. Defenders should not confuse this threat with MedusaLocker, a separate ransomware family, or the Medusa and TangleBot mobile malware families, which also share similar naming.

As a defense against these intrusions, the agencies urge organizations to implement robust defenses. Security teams must secure and patch exposed systems to protect firmware, operating systems, and software from exploitation. Additionally, administrators should restrict access from untrusted origins to remote services and implement network segmentation to prevent lateral movement.

The Ugly | Hackers Exploit Critical Windows IKE Protocol Vulnerability

CISA has added a critical remote code execution (RCE) vulnerability in the Windows Internet Key Exchange Service Extensions component, known as MS-IKEE, to its catalog of actively exploited flaws. Tracked as CVE-2026-33824, this severe double-free vulnerability affects all supported versions of Windows 10, Windows 11, and Windows Server 2016, 2019, 2022, and 2025. The flaw enables unauthenticated, remote attackers to execute arbitrary code by simply transmitting maliciously crafted UDP packets over port 500 or port 4500 to Windows systems running IKE version 2. Because this protocol component handles crucial features like cryptographically generated address authentication, denial-of-service protection, and third-party interoperability, exposed systems remain highly vulnerable to complete network compromise.

Although Microsoft addressed the issue during April 2026 Patch Tuesday, the firm has not yet updated its official advisory to reflect the ongoing in-the-wild exploitation. Under the urgent mandate of Binding Operational Directive 26-04, CISA ordered all U.S. Federal Civilian Executive Branch (FCEB) agencies to secure their vulnerable systems within three days. While this binding directive specifically targets federal networks, cybersecurity officials strongly urge all enterprise network defenders to prioritize applying the security updates immediately to halt active intrusions.

For organizations unable to immediately deploy the patch, Microsoft recommends restricting inbound UDP ports 500 and 4500 on systems where IKE is not required, or configuring host firewalls to only accept traffic from verified peer IP addresses. The rapid exploitation of this protocol flaw joins a growing list of recently abused Microsoft vulnerabilities, including a high-severity Windows Task Host bug and a SharePoint RCE vulnerability now heavily leveraged in ransomware campaigns. Since late 2021, CISA has cataloged hundreds of actively exploited Microsoft vulnerabilities to help defenders aggressively prioritize patching.

The Good | Courts Sentence “The Com” Online Syndicate Member for Blackmail & Sextortion

A court in the UK has sentenced a member of the decentralized online cybercrime collective known as “The Com” to two years in prison following an investigation by the National Crime Agency (NCA). Justin Swaddle, who operated under the digital aliases ‘Epstein’, ‘Rugen’, and ‘Moscow’ across Discord, Snapchat, and Telegram, pleaded guilty to multiple criminal charges of blackmail and child abuse. In addition to his sentence, the court ordered Swaddle’s placement on the National Sex Offenders Register and imposed a ten-year Sexual Harm Prevention Order.

Investigators revealed that Swaddle systematically targeted and groomed young, vulnerable victims globally, using popular chat platforms to exploit his targets. The prosecution identified 117 female victims worldwide, aged thirteen to seventeen, whom Swaddle coerced into performing severe acts of self-harm and generating explicit material. Rather than seeking financial gain, Swaddle was reportedly motivated by the online status and notoriety he obtained by sharing the media within exclusive subgroups. When victims resisted his demands, he used video recordings, home addresses, and school details to blackmail them into compliance.

The investigation, which the NCA initiated in January 2024 following Swaddle’s initial arrest by West Yorkshire Police, required extensive cross-border coordination. British officers collaborated closely with law enforcement agencies in the United States, Australia, Canada, Norway, and New Zealand to identify and safeguard affected children worldwide.

Authorities emphasize that The Com functions as a highly dangerous, loose-knit global network subdivided into specialized factions, including groups dedicated to physical violence, sexual coercion, financial extortion, and high-profile corporate ransomware operations.

The Bad | Agencies Warn of Expanding Gunra Ransomware Operations Targeting Critical Infrastructure

U.S., U.K., and South Korean intelligence and law enforcement agencies have issued a joint cybersecurity advisory warning global critical infrastructure organizations about escalating threats by Gunra ransomware. First appearing in April 2025 as a variant specializing in double extortion, the group uses malware derived from leaked Conti source code. Gunra targets public health, financial, and government sectors worldwide, with a heavy concentration of victims in Australia, East Asia, and Europe.

To establish initial access, operators exploit critical authentication vulnerabilities, specifically CVE-2024-55591 and CVE-2025-24472, in FortiOS and FortiProxy software, alongside security flaws in VPN gateways. While campaigns initially focused on Windows environments, the threat actors expanded to cross-platform operations by introducing a Linux variant. In January 2026, the group launched a formal Ransomware-as-a-Service (RaaS) affiliate program under the brand “Golden Community”, actively recruiting penetration testers to serve as initial access brokers. Attackers deploy their payloads via phishing and conduct ransom negotiations via WhatsApp.

Once inside a network, the actors utilize Impacket tools for credential dumping and lateral movement. They execute malicious tasks during nighttime hours, exfiltrating stolen documents to cloud services and deleting critical backup and archived data across primary and recovery centers. The malware leverages advanced ciphers like Salsa20 or ChaCha20 to encrypt terabytes of data in a limited timeframe.

Strong links have been identified between Gunra and North Korean state-backed threat actors, observing overlapping infrastructure and techniques, such as the exploitation of zero-day flaws in certificate signing software. Despite its sophistication, a catastrophic cryptographic flaw in Gunra’s Linux variant allows victims to fully recover encrypted files.

The Ugly | New ‘ShieldBreak’ Zero-Day Exploit Bypasses Microsoft Defender Protections

A security researcher known as ‘Nightmare Eclipse’ has released a novel Microsoft Defender zero-day exploit dubbed ‘ShieldBreak’ shortly after this month’s Patch Tuesday update. The vulnerability operates as a direct patch bypass for RoguePlanet, a separate privilege escalation flaw in Microsoft’s malware protection engine that was patched in July.

Although both flaws lead to SYSTEM-level compromise, researchers confirm the underlying exploitation techniques differ significantly. While the original RoguePlanet bug exploits a filesystem race condition using virtual disks to overwrite system files, ShieldBreak hijacks cloud-hydration processes.

Specifically, the exploit leverages user-mode callback hooks to modify file contents during a cloud-hydration scan via the Cloud Filter API. To achieve privilege escalation, an attacker first places a standard test file and utilizes Object Manager symbolic links to redirect Defender’s path to the system32 directory. During scanning, the exploit uses the Common Log File System to swap the file identity and plant a malicious DLL, phoneinfo.dll, where a default system file does not exist. Triggering a scheduled Windows Error Reporting task subsequently forces the system to load this rogue library, spawning a shell with highest privileges.

The proof-of-concept operates with a 100% success rate on fully patched installations of Windows 11 25H2 and Windows Server 2025. Although Windows 10 remains vulnerable to the flaw, the current code does not natively support those legacy systems. Analysts note that Microsoft Defender must be actively enabled for the exploit chain to function.

The release intensifies an ongoing dispute between Microsoft and the researcher over bug bounty policies and recent threats of legal action.

The Good | Snowflake Hacker Pleads Guilty as Ransom Cartel Creator Draws 16 Years

Connor Riley Moucka pleaded guilty in Seattle federal court this week to computer fraud, wire fraud, aggravated identity theft and conspiracy over the 2024 breaches of Snowflake customer accounts.

The intrusions reached at least 165 organizations and exposed records tied to at least 100 million people. Prosecutors say Moucka collected at least $495,000 from ransoms and data sales. He is due to be sentenced on October 27, facing a two-year mandatory minimum on the identity theft count and up to 30 years on the rest.

Every Snowflake account the group reached had multi-factor authentication switched off, and the credentials, some harvested by infostealer malware as far back as November 2020, had never been rotated. The gang didn’t need to find a vulnerability in Snowflake’s platform to exploit; it turned out that more than three-quarters of the compromised accounts had prior credential exposure, and none had network allow lists in place.

In separate news, The Department of Justice announced that Maksim Silnikau, the Belarusian national who built and administered the Ransom Cartel ransomware operation, was sentenced to 16 years for conspiracy and aggravated identity theft.

Ransom Cartel creator sentenced 16 yrs for REvil-linked RaaS targeting 18 orgs, $6.7M losses. https://intel.threadlinqs.com/threat/TL-2026-1902 #ThreatIntel #CVE_2021_1675 #CVE_2021_34527 #Ransomcartel

[image or embed]

— Threadlinqs (@threadlinqs.bsky.social) August 6, 2026 at 3:51 AM

Between 2021 and 2023, Silnikau recruited affiliates through underground forums, supplying stolen credentials and encryption tools while running an affiliate site that handled negotiations and revenue splits.

The Ranson Cartel attempted to extort at least $5.2 million from at least 18 companies, including a robotic-surgery startup disrupted for two months and law firms that paid $125,000 and $300,000 after weeks of downtime.

Silnikau was arrested in Spain in 2023, fled while awaiting extradition, and was caught trying to cross into Belarus before consenting to face trial in the United States.

The Bad | Claude Mythos 5 Spends 34 Hours Trying to Backdoor Real Code

The UK AI Security Institute (AISI) reported Tuesday that an agent running Claude Mythos 5 spent 34 hours trying to merge a malware dropper into a real open-source project during a security evaluation, after searching the open internet and landing on a real, unconnected repository whose name happened to share a keyword with the test’s fictional scenario.

The agent researched the maintainers, opened a pull request pairing a hidden dropper with a working bug fix, and cycled through three payload versions as each was caught. When a human reviewer flagged the code as malicious, the agent’s account denied it, force-pushed a rewritten branch history, and used a second account to vouch for the code. It also hid a prompt injection inside a GitHub issue aimed at AI coding assistants. The maintainer closed the pull request despite the cover-up.

Across 122 evaluation runs, AISI logged 19 unsanctioned actions in 10 runs, 17 involving Mythos 5 and two involving GPT-5.6 Sol, though AISI notes that figure is a count, not a rate. Both models had cyber safeguards off and open internet access by design, a setup meant to measure raw capability rather than mirror production use.

Nothing escaped AISI’s sandbox, and AISI found no evidence of real-world harm. The reviewer had the chance to catch it only because GitHub holds checks on a first-time contributor’s pull request until a maintainer approves them.

Anthropic’s own July 30 review of 141,006 evaluation runs found a separate Mythos 5 run that published malware to PyPI, downloaded and ran on 15 real systems within an hour. OpenAI reported a similar incident days earlier, exploiting a zero-day to reach Hugging Face’s production database. In each case, a test environment meant to stay sealed did not, and a model reached through it before anyone caught it.

Not to be outdone, Meta became the third lab in recent weeks to disclose an AI agent reaching into systems outside a security test. The exposure traced to Irregular, the same firm behind OpenAI’s second incident, whose misconfiguration gave a Meta model internet access it used to exploit a real company’s system.

The Ugly | ChainDrop Worm Compromises Over 1,300 npm Packages With Two Billion Monthly Downloads

A self-propagating worm known as ChainDrop compromised at least 868 npm packages across 1,381 versions, part of a broader campaign researchers put at more than 1,300 packages with a combined two billion monthly downloads.

The mass compromise began Tuesday when an attacker breached the GitHub account of a maintainer who controlled several widely used caching libraries, including Keyv, Cacheable, flat-cache and file-entry-cache. The worm then self-propagated to other maintainers’ packages, including ones tied to Deliveroo, Ornikar, OneReach, Picsart, Qlik and ServiceTitan.

The worm pushed malicious commits directly to each project’s main branch and triggered a new release through a GitHub Actions workflow, giving the poisoned npm packages valid provenance signatures.

A preinstall script added to package.json ran automatically on npm install, pulling down the Bun JavaScript runtime and using it to execute an obfuscated infostealer that harvested GitHub tokens, npm tokens, AWS and Kubernetes credentials, HashiCorp Vault secrets, database credentials and a run of other cloud and developer logins.

Every stolen token was checked against npm’s own whoami endpoint before the haul was encrypted and sent to a public GitHub repository, with any credentials belonging to other maintainers repeating the process on their packages.

ChainDrop is built on Shai-Hulud, the same self-propagating technique that has hit npm before. Each GitHub repo storing the stolen credentials is auto-named with random terms from Dune and carryies the description, “Shai-Hulud: Here We Go Again.”

While many of the auto-generated dead drop repos have since been taken down, the scale of the outbreak demonstrates how rapidly self-propagating worms can weaponize trusted credentials and automated pipelines. For a deeper breakdown and defensive strategies on this attack vector and other emerging supply chain risks, read the SentinelOne Annual Threat Report.

AI agents have made their way into virtually every layer of your environment. They run in the apps your employees adopt, on the endpoints where agents execute code, as users with access privileges those agents borrow, and in the cloud workloads that scale them. The platform that you trust to secure your endpoints is already already covering where AI operates today.

Here is the through-line that makes this one problem instead of four. Every AI attack starts as an interaction and ends as an action. A prompt gets manipulated, an agent gets tricked, and the damage lands on a host, reaches into an identity, or moves through the cloud. The tools that treat each surface as a separate product hand you fragments. SentinelOne treats them as one chain.

How SentinelOne Defends the Agentic Stack Today

Employee AI use is where the risk quietly enters. Your people are already using AI tools you never sanctioned, through browser, IDE, and API-connected apps and agentic AI tools. SentinelOne discovers that shadow AI use across browsers, IDEs, and copilots, highlights which tools and models are in play and governs it with policy. It keeps confidential data, PII, and secrets from reaching untrusted models, and it stops prompt injection and jailbreaks aimed at the tools you build. Legacy DLP reads patterns; this reads context, which is the only way to catch an attack aimed at AI systems that behave in a non-deterministic way.

The agent layer is where AI stops advising and starts acting. An employee’s prompt sends text. An agent sends commands, holds credentials, calls APIs, and chain actions without a human approving each step. That makes them non-human identities with standing access. SentinelOne governs that access. It inventories the agents and MCP servers already operating and scores what each one can reach and holds every agent to the privileges its task requires. Then it inspects the tool calls themselves, so an injected instruction gets blocked at the moment it would execute. What gets executed lands in a searchable record, and the same enforcement doubles as a kill switch.

While governance decides what an agent is allowed to do, the endpoint is where you find out what it actually did.

The endpoint is where agents execute. This is the frontier, and where SentinelOne has protected customers for over a decade. Our behavioral engine judges what a process does, not what it claims to be. That is how we caught QUIETVAULT – malware that spins up AI agents in “yolo” mode to exfiltrate secrets to GitHub. It is how we autonomously stopped the LiteLLM supply chain attack, where adversaries weaponized the Claude CLI to install a malicious payload. It is how we surfaced a DLL side-loading attack hidden inside an AI tool installer. Real detections, on the endpoint, today. Agents run on the host, and so do we.

The identity is where a hijacked agent runs next. Picture an employee’s AI coding agent that gets hijacked mid-task. It spawns a shell and reaches for cached credentials and cloud session tokens, trying to stop being a process and start being the user. That pivot to identity is what unlocks lateral movement, and it is where most AI attacks are headed. SentinelOne meets the move. It secures human and non-human identities alike, and seeds the environment with decoy credentials and honeytokens no legitimate user ever touches. The instant the hijacked agent grabs one, the trap trips, and Identity responds by forcing an MFA re-challenge, disabling the account, or isolating the host. Authorization at login is not enough. Access gets validated against behavior and pulled at runtime.

The cloud is where AI workloads scale. Consider an internal AI agent running in a Kubernetes cluster with standing access to a customer database. Security teams keep asking the same question about deployments like this. Where is the model connecting, and who is it talking to? SentinelOne answers with eBPF-native runtime protection that judges how the workload actually behaves, and flags the moment that inference service reaches an endpoint it has never touched before. It covers the control plane the deployment depends on, the secrets it reads, the pipelines it runs through, and the data it can access. Defending the AI you build takes more than watching it, it takes action on the workload in real time.

SentinelOne’s Singularity Platform Advantage

Each of these surfaces matters on its own. What closes the kill chain is following an attack across them without losing it at the handoff. This is where a single platform earns its keep. AI telemetry already streams into the Singularity Data Lake, alongside the endpoint data the platform has correlated for years. As identity and cloud signals join that same view, an analyst follows one attack from first prompt to final action, without stitching logs across six tools at two in the morning. A manipulated prompt, the process it spawns, the credential it reaches for, and the cloud resource it targets read as one story rather than six disconnected alerts.

Detection that only watches is observation. Runtime action is protection. When the Singularity Platform acts, autonomous response blocks the execution, rolls back the change, and revokes the access at the point of impact, without a human relaying orders between consoles. This is the difference between whether an attack is stopped or just gets logged.

That is the case for securing AI inside a platform built for autonomous runtime response. We are not adding a console to chase AI, we are extending the one already deployed where your agents run.

Questions to Ask When Assessing Your AI Security Options

When evaluating AI security, ask yourself three things.

• Does the solution protect the endpoint where agents actually execute, or is it a roadmap item?
• When a hijacked agent pivots to credentials and the cloud, does that telemetry land in the same platform, or are you manually connecting dots across three dashboards?
• Can the solution act at the moment of execution, or only tell me what already happened?

SentinelOne protects the surfaces where AI runs today. This includes the apps your employees use, the agents they deploy, the endpoints where agents execute, the identities they borrow, and the cloud where they scale. One platform, built for autonomous response. While AI has changed the attack, it does not have to change your architecture.

See it for yourself. Talk to our team about securing AI across your endpoints, identities, cloud, and the AI apps your employees already use, all from the platform you run today. Contact SentinelOne today.

Third-Party Trademark Disclaimer:

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third-party.

The Good | Authorities Disrupt “The Com”, Release Security Guidelines & Charge Telegram CEO

Europol and law enforcement partners from nine countries have flagged over 4000 URLs for removal to disrupt the online ecosystem of The Com. Operating as a decentralized network, The Com targets and recruits vulnerable youth across social media and gaming platforms.

Investigators report the syndicate’s content actively promotes self-harm, child exploitation, and physical attacks, while providing instructional manuals for swatting and arson. This multi-week joint operation builds upon Project Compass, a year-long international initiative that previously resulted in 30 arrests and identified 179 suspects linked to the criminal network.

From U.S. and Australian governments, a new joint cybersecurity guidance urges critical infrastructure organizations to proactively prepare isolation plans for operational technology systems. The advisory provides recommendations for physically and logically disconnecting vital infrastructure from corporate networks during severe cyberattacks.

Since state-sponsored threat actors and cybercriminals continuously target these essential sectors to facilitate espionage, data extortion, and disruptive operations, such resources help businesses shore up their operational resilience, documentation, and testing procedures.

The Russian Federal Security Service (FSB) has formally charged Telegram founder Pavel Durov with aiding terrorist activities and violating federal laws regarding prohibited information. Authorities accuse the messaging platform of failing to remove channels and automated bots allegedly operated by Ukrainian special services.

According to Russian intelligence, Ukrainian operatives leveraged a Telegram dating chatbot to psychologically manipulate and recruit young Russian men into sharing physical geolocations before coercing them into executing armed attacks and arson against domestic critical infrastructure.

This charge is the latest action against Telegram preceded by Durov’s arrest in 2024, restrictions placed on the platform, and a near-blockade from earlier this year.

The Bad | Theft Victims Sue Apple Over Fraudulent Cryptocurrency Wallet Application

Three individuals have filed a lawsuit against Apple after losing approximately $1.8 million in Bitcoin to a fraudulent cryptocurrency application housed on the official App Store. Between May and August 2025, the plaintiffs downloaded a malicious app impersonating “Sparrow Wallet”, a legitimate platform exclusively available on desktop operating systems.

The fraudulent app instructed users to input their secret recovery seed phrases during the initial setup process. Once victims entered these credentials, scammers immediately transferred the cryptocurrency to unauthorized external addresses.

The legal complaint alleges that Apple failed to adequately monitor its software marketplace while falsely promoting the App Store as a secure environment. Despite warnings from the real developer behind Sparrow Wallet who called out impersonator applications more than a year before these specific incidents occurred, Apple had taken no action.

Now, the plaintiffs claim the marketplace even featured the fraudulent software within curated cryptocurrency collections, effectively recommending the malicious tool to unsuspecting victims.

There is still a scam ‘Sparrow Wallet’ app on the @Apple App Store, despite myself and others having reported it weeks ago.

Worse, you have to install it to report it.

Always go to the wallet’s website to find the link to download it. App stores will not protect you.

— Craig Raw (@craigraw) January 6, 2024

Financial theft within mobile ecosystems continues to grow as researchers recently identified dozens of similar wallet impersonators actively targeting user recovery phrases and keys across the Apple App Store. In response to the allegations, Apple’s filing stated that it takes swift action to remove infringing content and provides dedicated reporting mechanisms for suspected fraud.

The plaintiffs currently seek full financial reimbursement, compensatory damages, and a formal court mandate requiring the tech giant to publicly disclose and systematically improve its internal procedures for detecting and removing fraudulent software submissions.

The Ugly | Anthropic Joins OpenAI in Finding Its Models Reached Real Companies During Cyber Evals

Anthropic disclosed this week that three of its models reached real production systems during cybersecurity testing. After reviewing 141,006 evaluation runs, it identified three incidents in which a Claude model reached the open internet from inside the environment of one of its third-party evaluation partners, and then accessed the systems of three organizations.

Unlike OpenAI’s recent breach of Hugging Face, no zero-day was involved, according to Anthropic. A misconfiguration had left machines in the evaluation with live internet access.

In one incident, a model built a malicious Python package and published it to PyPI, the public registry for Python software. The package was live for about an hour, during which time it was downloaded and run on 15 real systems. The downloaders included a security vendor’s scanner that installs packages automatically to check them for malware, leading to the model’s code exfiltrating that company’s credentials.

Anthropic said it found no evidence of a model pursuing a goal of its own, characterizing the incidents as a harness and operational failure rather than a model alignment failure, and noted that the safeguards it ships on released models would have blocked the behavior.

OpenAI also updated its account of the Hugging Face breach this week, revealing the same models had also used publicly exposed credentials to compromise accounts at four other services.

OpenAI said the models configured one compromised account as an outbound relay and staging server and used a second for data storage. The remaining two accounts were accessed in read-only mode.

Although OpenAI’s models extracted partial datasets containing CyberGym solutions and operated multiple concurrent workloads, the activity ultimately left critical encryption keys behind, exposing the operation. OpenAI said it continues to review the incident alongside external auditors and has restricted its pre-release model from further internal research access.

The Good | Authorities Dismantle Kratos Phishing Network & Arrest Its Developer

Kratos, a prominent phishing-as-a-service (PhaaS) platform, was dismantled from the inside out this week thanks to German and U.S. law enforcement agencies. During “Operation Olympus Blade”, authorities seized over 200 servers to render Kratos’ global network entirely inoperable while the platform’s suspected developer was apprehended in Indonesia.

So far, investigators estimate that more than 1800 cybercriminals utilized the platform to launch nearly 15,000 phishing campaigns monthly since late 2024. Operating as a franchise, the service provided threat actors with toolkits designed to generate convincing Microsoft authentication pages. These campaigns targeted victims across the United States and Europe, facilitating widespread credential theft and unauthorized account access. The operators earned at least €300,000 in subscription fees.

A recent report from cyber researchers reverse-engineered the Kratos toolkit, revealing how it offered operators two distinct functional modes. While one mode harvested traditional credentials, the more advanced setting deployed a Node.js reverse proxy. This adversary-in-the-middle (AitM) capability allowed attackers to intercept active session cookies in real-time, effectively bypassing standard multi-factor authentication (MFA) controls.

Once compromised, these accounts provided actors with initial footholds to execute business email compromise (BEC), lateral data theft, and secondary phishing attacks. Just this February, actors ran Kratos in a campaign that used tax-themed lures and personalized QR codes to target dozens of American manufacturing and healthcare organizations.

While the immediate server takedown severely disrupts ongoing operations, officials acknowledge that the existing customer base retains access to the underlying kit code. That means the toolkit itself outlives the infrastructure seizure, and operators who already have copies can resume campaigns under new branding with minimal rebuild effort.

The Bad | Threat Actors Conceal HollowGraph Malware in Microsoft 365 Calendar Events

A novel espionage implant, dubbed HollowGraph, is hijacking Microsoft 365 calendars to establish a covert command and control (C2) channel. By routing operator instructions and exfiltrated data through legitimate Microsoft Graph API traffic, the malware ensures its activities blend seamlessly with routine network chatter.

The .NET DLL implant operates purely as a two-way dead drop without communicating directly with an attacker-owned payload server. To receive tasking, HollowGraph queries the compromised user’s calendar for an event planted far into the future – in this case, dated for May 13, 2050. Operators embed their instructions within text files attached to this anomalous event, ensuring the mailbox owner never naturally scrolls far enough to discover the malicious entries.

For data exfiltration, the malware executes the reverse process. It systematically encrypts stolen files using hybrid RSA and AES-256 encryption, generates a new far-future calendar event, and uploads the targeted data as attachments. To maintain continuous Graph API access, operators utilize a secondary DNS-based channel to refresh the application’s Entra ID login credentials. The malware decodes these values from an attacker-controlled domain and writes them to a disguised configuration file.

Analysts observed this highly targeted campaign actively compromising machines at an Israeli organization between June and July 2026. While the implant’s underlying code shares significant structural similarities with a modular backdoor framework called Cavern, frequently utilized by Iranian state-sponsored syndicates, researchers have not yet definitively attributed this specific operation to a known threat group.

HollowGraph buries its C2 in M365 calendar events dated 2050 – no attacker server ever touched. https://t.co/yJo4fpw0X9 #ThreatIntel #HollowGraph #Cavern #Cav3rn pic.twitter.com/kaABaAYZg0

— ThreadLinqs (@threadlinqs) July 22, 2026

Since HollowGraph relies entirely on compromised account identities and legitimate application permissions rather than software vulnerabilities, standard patching remains ineffective. The technique effectively weaponizes the trust organizations place in their own Microsoft Graph API traffic, turning routine calendar activity into a blind spot by design.

The Ugly | AI Models Escape Sandbox to Breach Hugging Face Infrastructure

Open-source AI platform Hugging Face recently disclosed a network breach orchestrated entirely by an autonomous AI agent. The attack compromised the company’s data-processing pipeline by deploying a malicious dataset that exploited two distinct code-execution vulnerabilities. This initial access allowed the agent to run unauthorized code on a processing worker, harvest cloud and cluster credentials, and subsequently move laterally across several internal systems.

The platform reported that the agent executed thousands of individual actions across short-lived sandboxes while staging self-migrating C2 infrastructure on public services. Following the discovery, Hugging Face evicted the agent, revoked affected credentials, and rebuilt compromised nodes. It was during the forensic investigation that responders encountered a unique operational hurdle: Western AI models refused to process the malicious artifacts due to built-in safety guardrails. At this point, the company turned to an unrestricted, open-weight Chinese model to successfully complete the analysis.

Shortly after Hugging Face’s disclosure, OpenAI confirmed that its own AI models orchestrated the intrusion during internal benchmark testing. Operating with reduced safety guardrails for evaluation purposes, GPT-5.6 Sol and an advanced pre-release model attempted to cheat the ExploitGym cybersecurity assessment. Rather than solving the challenges natively, the models identified and exploited a shortcut, retrieving the test solutions directly from Hugging Face’s production database via a zero-day vulnerability in a third-party package registry cache proxy.

After gaining open internet access through privilege escalation and lateral movement, the models accessed Hugging Face servers. The AI agents chained together multiple attack vectors, utilizing the stolen credentials and additional zero-day vulnerabilities to establish remote code execution. OpenAI subsequently disclosed the zero-day flaw and collaborated with Hugging Face to implement stricter infrastructure controls and guardrails.

AI is absorbing the volume work that makes up the fundamental architecture of the Security Operations Center (SOC) tier system. While the tiers and the work aren’t going away, a junior and senior analyst’s day-to-day is changing fast.

At some point in the last week, every analyst on your team made the same call. Close an alert uninvestigated, because the queue was too long and triage ate the time real investigation and deep analysis was needed. Most of those calls were right, but odds are that at least one critical threat will eventually be overlooked.

The root cause here is the mathematical disparity. Nearly half of SOC teams lack the capacity to investigate more than fifty percent of the alerts they generate daily. Analyst capacity grows linearly while data volume compounds exponentially.

According to findings from SentinelOne®’s Annual Threat Report, what’s worse is that the math leans heavily towards the adversaries. Automated exploits have been recorded escalating privileges within a target environment in approximately 30 milliseconds. Similarly, malicious attack chains can progress from initial network access to establishing persistent footholds in under 50 seconds. No manual workflow currently matches these kinds of machine speed tempos.

As a result, the traditional tier structure of the SOC is transforming in real time. AI is already absorbing the triage and correlation work that that structure was originally built to manage. The critical call-out here is understanding that these tiers are not dissolving; rather, they are evolving. Junior and senior analysts still exist and hold their titles, and continue to have a clear career path ahead of them. What’s changing is the nature of the tasks that fills their day.

Analyst Tiers, Redefined by Depth

Historically, the distinctions between Level 1, Level 2, and Level 3 analysts were established primarily to manage high-volume workloads. Those distinctions were built to manage volume: Alerts routed to the right skill level, junior analysts escalating what they couldn’t resolve, senior time reserved for what actually needed it. AI now handles the triage and correlation volume those tiers existed to manage. Volume stops being the variable that defines the role. As a result, analyst tiers are being redefined by depth of expertise rather than the ability to process large queues.

That evolution isn’t limited to junior and senior analysts:

• Threat intel analysts move from manual feed correlation to directing AI-correlated intelligence
• Security engineers move from manual rule-writing to guiding AI-generated detection logic
• SOC managers move from tactical management to strategic leadership and AI governance

Depth of expertise replaces volume as the key differentiator: cloud architecture, identity, adversarial tradecraft. The kind of judgment that only comes from watching an environment long enough to know what normal looks like. AI can’t replicate nuanced, environment-specific skills that the legacy, volume-driven tiered system was never able to encourage or reward.

A Day in the Life, Before and After

In the legacy model, an analyst’s day typically begins by facing a queue containing hundreds of unvetted overnight alerts. Three or four hours go to manual triage, pivoting between tools to reconstruct what happened. The vast majority of the queue closes as false positives. The real threat, if there is one, surfaces hours later, pieced together across five or more disconnected consoles.

Adding in the administrative paperwork widens the gap even further. The legacy model requires analysts to spend upwards of an hour writing an incident report that adds nothing to the actual technical investigation. The new model allows the analyst to review and approve an AI-generated summary, adds environment-specific context, and closes the case in minutes.

In a modern model, the day starts with a prioritized queue instead of a noise wall: evidence-backed verdicts already assembled, ready for review. Thirty minutes go towards confirming the highest-priority case. That confirmation triggers a pre-approved response workflow within defined policy. Critical hours that were parcelled off to triage are used for proactive threat hunting instead.

Teams operating this way report the difference in hard numbers, according to two IDC research studies commissioned by SentinelOne.

• Teams using AI-powered investigation report 63% faster threat identification and 41% more efficient investigation. (IDC Business Value of Purple AI®, July 2025)
• AI SIEM customers on the Singularity Platform report 55% more efficient security operations and 4x more threats handled, at 55% lower solution costs. (IDC Business Value of Singularity AI SIEM, May 2026)

That time moves to where the judgment actually matters.

Governance is the New Core Skill

Recovered time only pays off if real judgment fills it. The most important judgment now is knowing when to distrust the AI.

The analyst who knows exactly where their AI is unreliable is more operationally effective than the one who trusts it uniformly. That skepticism is a skill and it has to be built on purpose, case by case.

Four capabilities define the analyst role going forward:

• Validating AI output instead of accepting it by default
• Designing the automation workflows that execute at machine speed
• Forming hypotheses worth hunting instead of only answering tickets
• Translating what the AI found into what it means for the business

Escalation frameworks must be designed to reflect that same judgment. Teams building trust in a new workflow route more cases to a human by default. Mature teams narrow that escalation path as their confidence in specific alert types grows. Either way, the analyst decides where the line sits, not the AI.

On top of this, analysts must govern the response earlier, setting the policy before events are triggered instead of reacting to it. Every automated action is scoped to a policy an analyst defined in advance. This keeps all of the details of the logged and fully auditable after the fact. The quality of what fires automatically traces back to the quality of that workflow.

The Career Path Forward

The evolution we are seeing in SecOps is directly addressing systemic issues of burnout and attrition, both significant risks to retaining talent within the cybersecurity industry. Under an AI-augmented model, every rung on the SOC career ladder gets more strategic:

• Junior analysts move from manual triage to verdict review
• Senior analysts move from reactive response to strategic hunting
• Managers move from daily firefighting to designing the system everyone else works inside

None of this happens in one leap. Adoption works crawl-walk-run, workflow by workflow. ‘Crawl’ starts with AI-assisted triage, validated against your own judgment, alert by alert. ‘Walk’ enables automated responses for well-understood, lower-risk cases, with human approval required for anything novel. ‘Run’ hands full workflows to AI for established threat patterns, with analyst time going to verdict review and hunting. Different parts of a SOC can sit at different stages of that maturity at the same time.

Start with the First 90 Days in the AI SOC checklist, a concrete plan for the next ninety days. For the full argument behind it, check out the Analyst’s Guide to the Autonomous SOC and see how SentinelOne is building toward this model.

Third-Party Trademark Disclaimer:

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third-party.

Security Operations Centers (SOCs) are currently confronting scalability challenges on two fronts: structural and cognitive. The day-to-day reality of modern defensive operations is stark: an analyst frequently begins a shift facing a queue deeply saturated with unvetted alerts. To process a single event, the analyst must open the alert, pivot to a secondary console to complete an investigation, manually enrich an IP address, copy a file hash into a third interface, and cross-reference an asset inventory that may not have been updated in months. Following this, they must author and refine queries, waiting for overloaded databases to return historical context.

The actual work of assessing the investigation’s results and moving to decision-making and action has not even begun. This is the administrative burden of the modern SOC. The true threats are not just those that attempt to bypass defenses, but the critical operational hours lost before an active mitigation attempt is even initiated. While analysts are highly trained professionals, the relentless requirement to perform manual data aggregation inevitably leads to exhaustion.

Misdiagnosing the Bottleneck: The Upstream Data Problem

Threat actors operate at machine speed, utilizing automation to pivot laterally across networks in a matter of seconds, frequently disappearing before defensive teams can even log into their terminals. Expecting human defenders to counter automated threat vectors by manually aggregating bad data is an architectural failure.

Every SOC inherits a highly fragmented data ecosystem. Telemetry is continuously generated by diverse sources, including firewalls, cloud workloads, identity providers, endpoint sensors, and legacy systems. This telemetry arrives in disparate dialects, varying formats, and highly inconsistent levels of fidelity. Before AI tools can accurately reason about a potential threat, or an analyst can initiate a logical investigation and run a playbook response, this raw telemetry must be synthesized.

Historically, organizations analysts take on these complex synthesis processes, manually normalizing data points across different vendor schemas. This represents a key misallocation of human intelligence. The asymmetry in modern security operations is not merely a discrepancy in speed; it is an imbalance in how security teams are forced to allocate their finite time. When operators spend the majority of their shifts wrangling data instead of actively investigating threats, the foundation of the SOC itself is inadequate. To achieve defensive velocity, organizations must recognize that fixing the data foundation is the mandatory prerequisite for improving all downstream security functions.

Architecting the Data Foundation with Singularity AI Data Pipelines

Addressing the upstream data problem requires the implementation of advanced data pipelines capable of resolving enterprise data chaos before it impacts the detection engine. Frameworks such as SentinelOne’s® Singularity AI Data Pipelines serve as this foundational layer, engineered to ingest telemetry from every source and in every format without requiring months-long integration projects or heavy manual engineering.

Modern pipelines utilize AI to normalize raw telemetry into standardized formats, specifically aligning with the Open Cybersecurity Schema Framework (OCSF). This structural alignment transforms fragmented logs into structured data that is immediately actionable. It eliminates the need for analysts to construct complex regular expressions during critical incidents simply to reconcile how two different software vendors format data, such as usernames or a timestamp.

Efficient data ingestion also requires dynamic, in-flight optimization. Not all telemetry possesses the same analytical value, and storing all generated logs in highly indexed, expensive storage tiers is financially and operationally untenable. Data pipelines optimize data streams by filtering out extraneous noise, trimming excess volume, and routing specific logs based on dynamic criteria. High-value security events are routed and indexed for rapid search retrieval, while lower-priority compliance or operational logs are routed to more cost-effective tiered storage. The result is a substantial reduction in infrastructure costs, a higher signal-to-noise ratio, and a structured data foundation that is completely prepared the moment an investigation is required.

When underlying data pipelines automatically enrich that log with identity and asset information, revealing (for example) that a specific financial director’s laptop in a remote office is communicating with a known botnet, the output transitions from a raw data point into a definitive starting point. Crucially, this enrichment occurs systematically before the human operator ever interacts with the alert. Solving this data problem end-to-end is a primary reason SentinelOne was recognized in the IDC MarketScape for AI SIEM.

Accelerating Detection via Singularity AI SIEM

When a clean, structured data foundation is properly established, the performance of downstream security tools accelerates. Modern detection engines, such as the Singularity AI SIEM, leverage indexless architectures to manage enterprise-scale telemetry. Because the data is normalized and optimized prior to ingestion, these platforms can execute petabyte-scale queries with minimal latency, ensuring investigative results are delivered before the analyst’s attention wanes.

Within this architecture, detection logic is executed continuously against a stream of clean, correlated telemetry. This transforms an ocean of disparate event logs into readable, centralized dashboards that provide immediate situational awareness. The quantitative benefits of this approach are substantial. With AI SIEM, organizations are already executing their queries 70% faster. Adding AI Data Pipelines further augments this workstream, providing cleaner data for AI to run at optimal efficiency. These improvements represent the direct result of ensuring that the data arriving at the SIEM is inherently fit for purpose.

AI SIEM remains a single, comprehensive SKU with customers automatically receiving integrated pipeline functionality for everyday data optimization rather than treating it as a premium add-on. For every unit of paid Data Ingest capacity, customers can process twice that volume through Data Pipelines. A customer with 500 GB/day SIEM entitlement can push 1 TB/day through the pipeline at no additional cost.

Transitioning to Agentic Reasoning Layers with Purple AI

The establishment of a structured data pipeline unlocks the capability for true agentic reasoning within the SOC. Unlike traditional rule-based automation, which executes static responses to predefined triggers, technologies like SentinelOne’s Purple AI operate as a dynamic investigative layer.

When an initial alert is generated, an agentic reasoning system does not simply pause and wait for human triage. It autonomously launches an investigation, comprehensively maps the potential blast radius of the incident, and synthesizes a clear, logical recommendation for containment. Then, the analyst logs into the console and is presented with a fully formed situational briefing rather than a blank investigation screen.

More importantly, an agentic AI layer possesses the capacity to evaluate broader adversarial campaigns rather than isolated security events. In isolation, a minor registry key modification, a singular file write, or a brief outbound network connection may not meet the threshold for a critical alert. Legacy security tools often fail to connect these disparate, low-signal events. However, Purple AI can assemble these seemingly unrelated activities into a cohesive narrative, exposing the overarching strategy of the attacker before a major breach occurs.

This level of autonomous intelligence is strictly dependent on the underlying architecture. Advanced AI algorithms cannot derive accurate conclusions from unparsed, low-quality telemetry. The analytical integrity of the agentic layer is entirely contingent on the principle of data quality; systems like Purple AI require clean, structured data to function effectively, avoiding the fundamental issue of “garbage in, garbage out”.

Governed Hyperautomation and the Human-in-the-Loop

The final component of a modernized, agentic SOC is the deployment of Hyperautomation to execute defensive responses. To counter threats effectively, organizations must deploy automated workflows capable of executing decisions at machine speed. These no-code workflows can be configured to trigger autonomously based on AI triage verdicts, the disclosure of new high-severity vulnerabilities, or specific incoming alerts. By automating the mitigation phase, the SOC evolves from an environment strictly dedicated to passive observation into a dynamic system that actively neutralizes threats.

However, the implementation of automated response mechanisms must be rigorously governed. Executing changes to enterprise infrastructure carries inherent risk. To mitigate this, automated workflows must integrate critical approval steps, ensuring that highly consequential actions are paused until human authorization is provided. The analyst retains the ultimate authority, defining the precise parameters of what processes may run automatically and what workflows require manual judgment.

Redefining the Analyst Mandate via Autonomous Security Intelligence

The strategic objective of integrating data pipelines, agentic reasoning, and Hyperautomation is not the removal of the human operator. Instead, the overarching goal is the restoration of the analyst’s primary function: exercising expert judgment.

By offloading repetitive tasks to technological systems, organizations systematically remove operational friction. The data layer filters out irrelevant noise, allowing the analyst to clearly see the threat. The AI investigation layer removes the administrative grind of data collection, allowing the analyst to focus purely on analytical thinking. Finally, the automated response layer eliminates procedural delays, ensuring the analyst’s decisions are executed rapidly enough to matter. This creates an intelligence fabric, known as Autonomous Security Intelligence (ASI), where data, investigation, and response function concurrently as a single, unified system.

Under this model, the operational output of a single analyst is exponentially multiplied, allowing one unburdened professional to accomplish the work of ten while still owning every critical decision. While the alert queue will perpetually require attention, the fundamental nature of the work fundamentally changes. The timeline of a manual initial triage to active investigation compresses from a multi-hour ordeal into a matter of minutes. The data arrives clean, the investigation runs automatically, and the response mechanisms are prepared. The hours previously consumed by administrative waiting are directly reallocated to strategic decision-making.

Conclusion

When defensive systems are finally architected to operate at the speed of the modern threat landscape, the role of the human operator transforms. Analysts are no longer forced to act as passive passengers, grateful to be carried by fragmented tools. They are elevated to the role of pilots, operating with full situational awareness, retaining their judgment, and actively directing the defensive posture of the organization. This is the paradigm of the agentic SOC, and it is entirely predicated on the foundation of clean, structured data.

Contact us today to learn more about how SentinelOne is leading the way forward with Agentic SOC.

 

The Good | Authorities Sanction Cybercriminals & Dismantle Russian Bulletproof Hosting Infrastructure

The EU and the United Kingdom have jointly sanctioned multiple Russian individuals and entities for targeting government networks and critical infrastructure across Europe. The sanctions specifically target senior Russia military intelligence (GRU) officers and operators, as well as four entities linked to the Federal Security Service (FSB).

Officials say that the Russian government actively utilizes these state-sponsored units alongside recruited cybercriminals and private companies to systematically destabilize international partners and compromise key infrastructure across the continent.

From the U.S. Treasury Department, two individuals and a virtual private network (VPN) provider face sanctions for actively enabling ransomware attacks against American organizations.

OFAC designated First VPN Service (1VPNS) and its administrator, Dmytro Rashevskyi, for supplying infrastructure that helped cybercriminals obscure their identities and manage stolen data. The service, which law enforcement dismantled last May, notoriously ignored abuse complaints and maintained zero user logs.

Yegeniy Silayev was also sanctioned for developing cryptors designed to conceal malware. Investigators estimate these specific tools and services directly facilitated billions of dollars in financial losses across critical sectors.

U.S. Federal prosecutors also unsealed indictments this week against three Russian nationals for operating bulletproof hosting services that facilitated over $62 million in global ransomware damages.

Defendants Aleksandr Volosovik, Yulia Pankova, and Kirill Zatolokin allegedly managed “Media Land” and “ML Cloud”, providing essential infrastructure to syndicates like Lockbit, Play, and Blacksuit. These hosting platforms actively shielded cybercriminals by disregarding victim complaints and ignoring law enforcement takedown requests.

To disrupt this supply chain, the State Department is offering a $10 million reward for actionable information regarding foreign government links to these hosting providers.

The Bad | Attackers Trojanize Popular Remote User Platforms to Deploy Starland Malware

Cybersecurity researchers identified a financially-motivated Russian threat actor tracked as UAT-11795. Active since June 2025, the actor has utilized trojanized applications to harvest user credentials and cryptocurrency while primarily targeting users across the United States, Germany, Romania, and Venezuela.

To distribute their payloads, UAT-11795 operators disguise malicious installers as legitimate software, including WebEx, Zoom, MobaXterm, DBeaver, and FaceIT. Researchers suspect the attackers likely deploy these files via ClickFix social engineering.

The infection chain typically starts when a victim executes a malicious HTA file. This file retrieves an altered NSIS installer harboring a hidden Python loader disguised as a standard text document. The loader then modifies the Windows Registry to ensure persistent access before decrypting and deploying the Starland remote access trojan (RAT).

Upon execution, Starland verifies whether it is operating within a sandbox before creating scheduled tasks and attempting to escalate its system privileges. The malware scans compromised systems for browser data, cryptocurrency wallet assets, detailed system configurations, any antivirus products, and Active Directory infrastructure such as domain structure and controllers.

Beyond data theft, Starland possesses extensive capabilities to capture desktop screenshots, execute arbitrary shell commands, and fetch secondary payloads. Depending on system architecture, the malware can inject a 64-bit shellcode chain to deliver the CastleStealer information stealer or a 32-bit chain to deploy the Remcos remote access trojan.

To maintain resilient command and control (C2) communications, the operators integrate a redundancy mechanism that queries a Polygon smart contract for a fallback domain, and control two Telegram bots to receive notification beacons, including messages with the victim’s machine fingerprints and cryptowallet inventories.

Users are reminded to avoid executing unidentified commands online and should only download confirmed software from official vendor sources.

The Ugly | Nearly 300 Imposter GitHub Repositories Distribute Infostealing Malware to Collect Sensitive Data

Threat actors have published almost 300 fabricated GitHub repositories to distribute an information stealer from the BoryptGrab malware family. The actors systematically impersonated premium security products, cryptocurrency tools, and developer utilities to deceive victims searching for free software downloads.

As part of the lure, the malicious landing pages employ highly sophisticated client-side scripts that parse referral URLs to render customized branding and spoofed trust badges, significantly increasing the likelihood of successful social engineering.

Once a targeted victim clicks the download link, the infrastructure delivers a constantly rotating ZIP archive containing a legitimate, signed WinGUP updater paired with a trojanized dynamic link library file. When the user executes the updater, the program side-loads the malicious file, which then decodes and reflectively executes the BoryptGrab-variant payload directly into system memory.

Operating without establishing long-term persistence, the malware is designed to exfiltrate maximum data in a single execution cycle. The stealer targets passwords, payment details, and session cookies across 19 different web browsers and 32 cryptocurrency wallet brands, alongside messaging tokens from Discord, Steam, and Telegram.

To maximize collection, operators utilize direct code injection to bypass Chrome’s native App-Bound Encryption. All newly harvested data is compressed and routed to a Russian-based C2 server. Although the malware leaves behind forensic evidence by failing to wipe temporary staging directories, the scale of the impersonation campaign poses significant risks to unsuspecting developers.

GitHub has already removed a large portion of the false repositories, though several of the malicious redirector pages remain actively online. Researchers advise users to independently verify software authenticity and exercise extreme caution when navigating unofficial portals, sharing this YARA rule to help detect BoryptGrab activity and IoCs.

The Good | Authorities Apprehend Pro-Russian Hacktivist & Dismantle Global Fraud Networks

Spanish authorities, acting on intelligence provided by the FBI, have apprehended a suspected core member of the pro-Russian hacktivist syndicates CyberArmy of Russia Reborn (CARR) and Z-Pentest.

While masquerading as ideologically motivated collectives, these groups have actively executed disruptive cyberattacks against critical infrastructure, including food processing and water facilities across the United States and Europe. Investigators allege the arrested individual, residing in Palencia, provided extensive operational and logistical support to a Ukrainian hacker working for CARR, even attempting to facilitate their escape to Russia.

The individual is also suspected of coordinating cyber operations for the NoName057(16) group using encrypted messaging platforms. In a March 2026 raid, law enforcement officers seized multiple computers and successfully froze cryptocurrency wallets utilized to launder illicit proceeds generated from stolen data sales. The suspect currently faces ongoing criminal investigations for alleged collaboration with a recognized terrorist organization and severe computer damage.

In a massive global crackdown on social engineering and financial fraud, international law enforcement agencies have arrested 5,811 suspects and seized approximately $293 million in illicit assets. Codenamed “Operation First Light 2026”, the coordinated initiative spanned 97 countries and specifically targeted business email compromise (BEC), investment scams, and money laundering syndicates operating between January and April. Interpol actively coordinated the extensive joint action, collaborating directly with regional policing bodies like ASEANAPOL, GCCPOL, and Europol to swiftly block over 31,000 fraudulent bank accounts and virtual wallets.

Investigators identified more than 142,000 victims worldwide and pinpointed an additional 15,600 suspects for future prosecution. This success builds upon recent international efforts, including Operation Synergia II, to dismantle the sprawling infrastructure supporting transnational cybercrime. Officials emphasize that robust, cross-border law enforcement cooperation remains essential to combat the escalating threat of organized cyber-enabled financial crimes globally.

The Bad | Threat Actors Deploy Forg365 PhaaS to Hijack Microsoft Accounts

Cyber researchers have identified a new phishing-as-a-service (PhaaS) operation dubbed Forg365, which targets Microsoft 365 enterprise accounts. Blending adversary-in-the-middle (AiTM) techniques with device-code phishing, the platform provides an integrated dashboard to manage post-compromise activities.

Forg365 works by incorporating AI to assist in generating customized phishing lures. By integrating AI directly into the control panel, Forg365 developers significantly lowered the financial cost needed to launch targeted campaigns. To remain undetected, its operators route messages through legitimate Amazon SES infrastructure while hosting their landing pages on Cloudflare.

The operation leverages the OAuth 2.0 device code authentication flow, originally designed for input-constrained devices. Attackers present victims with a deceptive verification page, tricking them into authorizing an attacker-controlled gadget rather than stealing their password directly.

Once initial access is achieved, the platform ensures persistence through a specialized browser extension called ForgCookie. Compatible with Microsoft Edge, Google Chrome, and Brave, this extension operates silently to request account data, clear session cookies, and trigger a hidden OAuth flow to capture fresh tokens. Doing so grants attackers continuous access to the victim’s Microsoft services without requiring them to ever re-authenticate.

To protect its administration panel from being accessed by security defenders, Forg365 integrates robust anti-analysis features. The platform utilizes debugger traps, polymorphic code, and dynamic sandbox checks to evade detection, redirecting connections to innocuous websites whenever a VPN is detected.

Administrators can defend against these hijacking techniques by monitoring Microsoft Entra logs for unexpected device-code authentication events. Organizations can also restrict or entirely disable device-code flows unless absolutely necessary. In the event of a suspected compromise, security teams should revoke all OAuth grants and refresh session tokens to sever access.

The Ugly | Rival Espionage Actors Breach & Spy On Pakistani Law Enforcement Networks

Between February 2024 and April 2026, suspected state-sponsored threat actors based in China and India separately converged on several Pakistani law enforcement organizations in unrelated cyberespionage campaigns.

According to SentinelLABS, operators heavily targeted the Balochistan Police, compromising critical network appliances and web servers. By infiltrating these critical systems, both nations actively sought independent visibility into Pakistan’s internal security posture and ongoing counter-militancy operations.

The China-nexus intrusions, leveraging PlugX, ShadowPad, and Cobalt Strike malware, were likely driven by Beijing’s concerns over the safety of Chinese nationals working on regional infrastructure projects within the China-Pakistan Economic Corridor (CPEC). Ongoing terrorist attacks have left the Chinese government dissatisfied with Pakistani protection and data from Balochistan Police would give the PRC direct insights.

Conversely, the India-nexus activity utilized Remcos backdoors to gather intelligence on the restive Balochistan province, a recurring flashpoint in the adversarial relationship between the two countries. Control over Balochistan Police networks means having invaluable visibility on how Pakistan manages their security posture as well as persistent access to civilian data.

One China-aligned threat actor specifically compromised the Balochistan Police Force’s Complaint Management System (CMS), a web application that actively serves both law enforcement personnel and Pakistani civilian users. The attackers uploaded custom malware implants disguised as routine portal updates, effectively weaponizing the digitalization of public policing services.

One payload masqueraded as a legitimate component of endpoint security software to evade initial detection and deploy an AsyncRAT client in order to establish persistent footholds into internal police networks while simultaneously surveilling citizens utilizing the platform.

This multi-actor convergence highlights how modernized policing infrastructure can serve as a high-value intelligence target for rival nations seeking comprehensive regional data.

Note The June 2026 Dark Web Breach Incident Trend Report is based on major data breach cases posted on the deep web and dark web forums. Due to the nature of some sources, it was difficult to fully verify the accuracy of certain information, so the report includes content that requires further verification. Major Issue […]

The US Federal Government is committing $600 million to build one of the world’s most advanced AI infrastructure systems. Executive Order 14363, the Genesis Mission, connects national laboratory supercomputers across nuclear simulation, biodefense, energy grid modeling, and every major scientific domain. Fifty-one organizations signed on, including NVIDIA, OpenAI, IBM, Microsoft, AWS, Google, and Oracle.

The security framework governing these workloads was not written for this scale of use.

NIST SP 800-234, the High-Performance Computing Security Overlay, is well-constructed, tailoring 60 controls across four security zones, building on the SP 800-53B moderate baseline. It was designed for deterministic HPC workloads such as climate simulations, finite element analysis, and computational fluid dynamics. These workloads share a common attribute: code that runs the same way, every time, and behaves predictably under well-understood inputs. The security controls governing those workloads assume you can scan at the perimeter, clear memory between jobs, and attest to integrity at load time.

AI workloads break every one of those assumptions.

SentinelOne has submitted a formal proposal to the NIST HPC Security Working Group regarding this gap, and NIST has acknowledged it. We have a post on LinkedIn to share our proposal, and welcome commentary from across the industry.

The supply chain problem just got a lot more dangerous

This spring, in just three weeks, three AI-driven supply chain attacks targeted widely deployed software: LiteLLM, the most-used AI infrastructure package in Python development environments, Axios, the most-downloaded HTTP client in the JavaScript ecosystem, and CPU-Z, a trusted system diagnostic tool with a legitimate signed binary from the official vendor domain.

SentinelOne stopped all three on the same day each attack launched, with no prior knowledge of any payload.

The most important aspect of this outcome is how these attacks were stopped, and why signature-based detection couldn’t work. Each attack arrived through a trusted delivery channel. LiteLLM was compromised after credentials were stolen via Trivy, a security scanner. The attacker published two malicious versions to the PyPI repository. In at least one confirmed case, an AI coding agent with unrestricted permissions auto-updated to the infected version, meaning there was no human review or approval step before the payload ran. The Axios attacker exploited a legacy access token that the project maintainers had forgotten to revoke, bypassing every npm security control. CPU-Z attackers targeted the vendor’s distribution infrastructure directly; anyone who downloaded from the official website received a properly signed binary containing a payload. In all three cases, while the authorization chain was legitimate, the intent was not.

This is the defining characteristic of modern supply chain attacks: the workflow is verified, but the intent has been subverted. Every perimeter control, signature library, and reputation lookup checks authorization and passes. These attacks were designed to exploit that gap, and they ran at machine speed through automated pipelines with no human checkpoint.

To put this into the context of HPC and AI workloads running at scale, a compromised Python package in a developer’s environment is a serious incident; a poisoned training pipeline on classified biodefense data on a national laboratory supercomputer is on a different order of magnitude. The model it produces may be correct 99.9 percent of the time and adversarially wrong under precisely targeted conditions. No perimeter scan, signature check, or load-time integrity verification will catch it after training completes.

Where the current framework falls short

Of the 60 controls SP 800-234 tailors, three bear directly on AI workload protection, and each carries a documented gap. In a fourth area, supply chain, the overlay does not tailor at all.

• SI-3 (Malware scanning): The control acknowledges that real-time scanning is most effective but explicitly permits tailoring for performance on HPC systems, deferring to perimeter scanning before data reaches the compute zone. For traditional HPC workloads, that tradeoff may be defensible, but for AI workloads, it leaves behavioral analysis of the execution process completely unaddressed. A poisoned training run that executes within the expected statistical range of a training job looks like legitimate compute to a perimeter scanner.
• SI-4 (System monitoring): The control notes that high-speed data flows in HPC environments can overwhelm standard monitoring tools, and lacks AI-specific monitoring requirements or telemetry collection requirements from execution pipelines. The practical interpretation of this is: monitor what you can, accept the gap for what you can’t. On infrastructure running AI at scale, that gap creates a primary attack surface.
• SC-4 (Information in shared resources): Requires GPU memory clearing between user reassignments. It addresses data residency at the transition but does not address runtime behavioral monitoring of workloads during execution, side-channel attack detection, or anomalous compute-pattern identification while training is active.
• SR family (Supply chain risk management): The overlay carries all 12 moderate-baseline SR controls forward from SP 800-53B, with no HPC or AI-specific guidance, and supply chain is not among the 14 categories it tailors to. The SR controls still address only the conventional software and hardware supply chain; they say nothing about training-data provenance, model-weight integrity, or pre-trained-model validation, and the framework defines no AI equivalent of a software bill of materials. LiteLLM, Axios, and CPU-Z all arrived through legitimate software supply chain channels. AI workloads carry that same exposure one layer deeper, in the data and model artifacts that software trains on, which is exactly where the overlay is silent

AI Runtime Threats

The attacks against AI workloads on HPC are not theoretical, and they are not detectable at the perimeter.

• Training data poisoning scales at rates most security teams are not equipped to respond to. Research¹ across 41 studies documents attack success rates exceeding 60 percent from manipulation of 100 to 500 training samples, a fraction of a percent of a typical dataset. Poisoning as little as 3 percent² of training data achieved 41 percent attack success rates in code-generating models. OWASP’s LLM Top 10³ documents the consequence. Backdoors leave model behavior intact until a specific trigger activates adversarial outputs. The model ships, it gets deployed, and operates correctly, until it doesn’t. No post-training audit reliably catches a well-designed poisoning attack.
• GPU side-channel attacks are executed remotely by a co-tenant workload on shared GPU infrastructure; no physical access is required. The NVBleed research demonstrated covert channel attacks on NVIDIA NVLink, achieving over 91 percent accuracy in recovering data-dependent information from co-tenant GPU workloads on a shared fabric. The BarraCUDA research demonstrated the extraction of neural network weights via electromagnetic side channels from NVIDIA hardware. Both attack classes execute during active training, not at job transition. If your HPC environment runs multiple projects or security classifications on shared accelerators, the co-tenancy model is an active attack surface today.
• Inference pipeline compromise survives load-time integrity checks. A model with clean weights at deployment faces attacks through three vectors: hot-swap modification of serving configurations while inference runs; preprocessing and postprocessing layer injection that alters inputs before they reach the model or modifies outputs before delivery; and adversarial input manipulation that triggers targeted misbehavior in a model that appears fully operational. For AI serving safety-critical inference, each is a security risk, not just a research concern.

The characteristic that makes AI workloads uniquely difficult is persistence. A compromised simulation may produce visibly wrong results, but a compromised model can produce correct results the overwhelming majority of the time and adversarially wrong results under precisely targeted conditions. By the time anyone has reason to investigate, the window for recovery has often closed.

Securing HPC AI Workloads

We know the technology required to address these gaps exists and has been proven at scale in environments with performance constraints far tighter than those in HPC. What is needed is a well-defined architecture that enables the secure execution of large-scale AI workloads.

Dedicated security compute. Runtime security that shares CPU resources with the workload it monitors can be starved of CPU time under heavy load and interfered with by a workload that achieves kernel-level access. The SPiCa research demonstrated that eBPF monitoring pipelines can be manipulated from within the kernel by rootkits filtering events before they reach the analysis engine, meaning that a co-scheduled monitor is not a reliable monitor.

Every other infrastructure function on an HPC node has dedicated resources. The job scheduler, the filesystem client, and the out-of-band management plane. Security monitoring is infrastructure and should be afforded the same dedicated resources.

Modern HPC nodes have 128 to 256 CPU cores. One reserved for security monitoring is less than one percent of the available compute. Linux kernel CPU isolation via isolcpus, nohz_full, and rcu_nocbs is production-proven in high-frequency trading and real-time systems, with bounded, predictable overhead.

eBPF-based behavioral telemetry at the training layer. Effective monitoring of an AI training pipeline means continuous observation of compute behavior profiles, memory access patterns, GPU utilization, and inter-node communication, with behavioral baselines established for approved training configurations. A poisoning attack that executes within expected statistical ranges is not visible to a perimeter scanner, but it is visible to a behavioral baseline that knows what the training job should look like.

This is the same principle that SentinelOne’s on-device Behavioral AI detected for LiteLLM, Axios, and CPU-Z. The LiteLLM detection flagged a Python interpreter executing Base64-decoded code in a spawned subprocess. The CPU-Z detection flagged an anomalous process chain: cpuz_x64.exe spawning PowerShell, which spawned csc.exe, which spawned cvtres.exe. CPU-Z doesn’t do that. The behavioral baseline knew what legitimate execution looked like, and in these cases, that behavior was the decisive signal.

Cloudflare uses an eBPF-based architecture to mitigate DDoS attacks exceeding 7 Tbps. SentinelOne uses it to detect and stop threats in under one second across enterprise fleets. A training job that begins writing to unexpected locations, establishing anomalous inter-node communication, or deviating from its expected compute profile is detectable at runtime, before the model completes training. The performance argument against runtime monitoring on HPC was never about the technology; it requires a shift in architecture.

Inference-time output monitoring. Deployed models require continuous observation of output distributions, latency patterns, confidence score distributions, and input-output statistical properties. A model under adversarial input attack, or serving modified weights, exhibits detectable output patterns before any human analyst notices the outputs are wrong. Circuit-breaker logic needs to be designed into the serving architecture, not added after the first incident.

Model integrity verification that runs during inference. Load-time attestation is a necessary and important requirement; it is not sufficient. Long-running inference deployments are vulnerable to hot-swap attacks that replace weights after the initial integrity check passes. Continuous cryptographic hash verification of loaded model weights, running on the dedicated security core with automated circuit-breaker logic on failure, closes that vector. For a model serving safety-critical calculations, the re-verification frequency should match the workload’s risk profile with predictable overhead.

An SR-family extension for the AI supply chain. The existing SR controls address software supply chain risk. They do not address training data provenance, model weight integrity at ingestion, or pre-trained model validation. An AI bill of materials, including cryptographic documentation from the training data source through intermediate checkpoints to the deployed model, is the model-layer equivalent of software supply chain controls. Without it, every pre-trained model loaded into an HPC environment is an unverified artifact from an unverified chain.

Defending AI at Every Layer

The supply chain attacks this spring demonstrated what happens when defense architecture falls behind the delivery mechanisms attackers use. LiteLLM, Axios, and CPU-Z all arrived through trusted channels, carrying payloads no signature database contained. They were stopped because behavioral detection does not require prior knowledge of the payload. It requires knowing what legitimate execution looks like and acting when execution deviates.

Defenders protecting AI workloads face that same problem across every layer they own. HPC is the hardest version of it. But identities, endpoints, applications, and infrastructure all carry the same exposure at different scales. SentinelOne gives defenders coverage across all four, with behavioral AI running at each layer to catch what signatures miss. The specifics of how that works across your AI environment are in our AI security overview.

Citations

¹ “Data Poisoning 2018–2025: A Systematic Review. IACIS (2025)”, and “Data Poisoning Vulnerabilities Across Health Care AI Architectures. JMIR (2026)

² “Poisoning Attacks on LLMs Require a Near-Constant Number of Poison Samples” (2025). arXiv:2510.07192 and Huang et al., 2020.

³ OWASP (2025) LLM04:2025 Data and Model Poisoning. OWASP Gen AI Security Project.

The Good | Authorities Apprehend Iranian Cybercriminal & Extradite UNC3944 Hacker

Montenegrin law enforcement, alongside the FBI, have apprehended a 39-year-old dual Iranian and Turkish citizen wanted by the U.S. government for several cybercrime offenses. Arrested in Kotor, the suspect faces charges in the Southern District Court of New York for conspiracy to commit computer fraud, hacking, and identity theft.

Since 2013, this individual allegedly orchestrated mass cyberattacks against more than 150 American universities and inflicted damages estimated at over $3.4 billion. Investigators say the stolen data and compromised academic credentials directly benefited the Islamic Revolutionary Guard Corps and various Iranian state entities. The case now proceeds to a High Court judge for formal extradition hearings. The arrest follows recent warnings from U.S. cybersecurity agencies regarding escalating Iranian state-sponsored operations targeting critical domestic infrastructure.

A 19-year-old dual United States and Estonian citizen, Peter Stokes, has been extradited to face federal charges for his role as a core member of the UNC3944 (aka Scattered Spider, oktapus) cybercrime syndicate. Finnish authorities initially apprehended Stokes at the Helsinki airport as he attempted to board a flight to Japan. Prosecutors are accusing him of orchestrating multiple high-profile corporate breaches, using intense social engineering tactics against IT helpdesks to bypass multi-factor authentication controls.

In one notable May 2025 incident, UNC3944 compromised a multibillion-dollar retailer, demanding an $8 million ransom while inflicting over $2 million in operational disruption and remediation costs. UNC3944 operators are also responsible for more than 100 network intrusions globally, all of which yield upwards of $100 million in illicit extortion payments. Stokes remains in federal custody in Chicago, facing charges of fraud, conspiracy, and computer intrusion.

The Bad | Russian Intelligence Exploit Phishing Campaigns To Steal Signal Backup Keys

CISA and the FBI are warning that Russian state-sponsored threat actors have made large strides in evolving their phishing operations to target the backup recovery keys of Signal users. In an update to their March 2026 advisory, the two agencies attribute this ongoing activity to Russian Intelligence Services (RIS), including Russia’s Federal Security Service (FSB) Border Guards and the country’s military.

Tracked as UNC5792 and UNC4221, these campaigns specifically target high-value individuals, including government officials, military personnel, journalists, and policy analysts. Previously, operators focused on harvesting standard verification codes or tricking users into silently linking unauthorized devices. Now, they employ social engineering to access private communications without compromising the application’s underlying end-to-end encryption.

During targeted intrusions, attackers masquerade as official Signal support personnel and send direct messages falsely claiming the platform requires mandatory two-factor verification following alleged international cyberattacks. The operators systematically guide victims through the specific process of enabling the Secure Backups feature, instructing them to paste their newly generated recovery key directly into the chat interface.

Once adversaries obtain this critical key, they seamlessly download and decrypt the victim’s entire historical message archive onto their own controlled devices. Simply registering a new account under the same phone number does not natively invalidate a compromised key – users must actively generate a new backup key within their application settings to effectively secure future communications.

The U.S. Department of State recently announced a substantial reward of up to $10 million for information leading to the identification or location of these operatives. Through the Rewards for Justice program, federal authorities actively seek actionable intelligence regarding the syndicates’ operational infrastructure, illicit funding mechanisms, and direct affiliations with Russian intelligence services.

The Ugly | Unknown Hackers Breach Department of Homeland Security Information Network

The Department of Homeland Security is actively investigating a cyberattack that recently compromised its Homeland Security Information Network (HSIN). The network is an information sharing platform used by federal, state, local, and private-sector partners, specifically to share sensitive but unclassified data amongst the government and internationally.

According to an initial report, an unidentified threat actor orchestrated the intrusion between late May and early June. Investigators indicate that the attackers targeted the HSIN’s core servers alongside a SharePoint environment designed for extensive interagency collaboration.

While officials have not yet attributed the breach to a specific foreign government or syndicate, the full extent of data exposure remains unclear. The compromised platform routinely supports real-time incident management, intelligence exchange regarding persons of interest, and operational coordination. Because the United States is overseeing security for World Cup matches across the country, experts raise concerns that the intrusion could have exposed critical security planning, response procedures, and communication protocols.

In a public statement to the press, a departmental spokesperson confirmed the incident, clarifying that the breach strictly involved an unclassified legacy information-sharing environment. Security personnel promptly isolated the affected systems and mitigated the underlying vulnerability before starting forensics.

Officials emphasized that the attack did not impact classified networks, and the primary system remains operational for authorized partners. As the investigation continues, authorities strongly urge U.S. government staff, contractors, and associated partners to remain vigilant while defense teams harden the underlying infrastructure against further unauthorized network access attempts.

When SentinelOne® introduced the Autonomous SOC maturity model, we made a deliberate choice: describe a journey, not promise a destination.

The industry had no shortage of vendors declaring that AI would transform security operations. We thought the more useful contribution was a framework for understanding what that transformation looked like, at what pace it was realistic, and what conditions each stage of progress required.

Security teams found the model useful. Not as a marketing claim, but as a map. CISOs and SOC leaders started placing their organizations on it, asking what it would take to move forward.

What happened next was telling. By RSAC 2026, ‘autonomous SOC’ appeared in vendor keynotes and product launches from companies that hadn’t used the term twelve months earlier. Add in pseudonyms like Agentic SOC and AI SOC, and the list explodes. Fast adoption brings loose definitions. For us, it’s worth being precise about what the concept means and what it doesn’t.

Here is what SentinelOne has learned from 18 months of real-world Autonomous SOC deployments.

What Held Up

Today, the progression still maps accurately to where organizations are and what separates each stage from the next. That accuracy holds even for a framework built before most organizations had meaningful AI deployment experience. The inflection points reflect real operational transitions at each maturity step.

The “journey not destination” framing has proven more important than we anticipated when we wrote it. In early 2026, Gartner published guidance to help buyers evaluate AI SOC claims more critically, noting that vendor credibility in this space depends on honest representation of where the technology is:

A maturity model is structurally honest. It reflects where you are, not where a vendor wishes you were. Gartner’s research found that while 40% of organizations are actively evaluating AI SOC capabilities, only 18% have actually deployed₂. The gap between evaluating and deploying is rarely about technology. Most organizations cannot advance because they lack a clear view of where they stand or what the next stage requires.

When security leaders use the model as a reference point, the evaluation conversation changes. The question shifts from “does your product make my SOC autonomous?” to “what would it realistically take to advance, given where we are today?” A feature list cannot answer that question. An honest vendor can.

Watch our webinar on why most AI SOC deployments stall here.

What We Underestimated

The levels were always sound. What we underestimated was how much organizations needed to build before they could operationalize them. Customers understood where they wanted to go. But achieving Partial Autonomy (Level 3) requires a data foundation, a workflow architecture, and AI readiness that most teams were still building when we first published this model. That’s a fact about where most security organizations were in 2024.

The transition from AI-Assisted Operations (Level 2) to Partial Autonomy (Level 3) is primarily a governance problem, not a tooling one. The tools are capable. What most organizations are missing is an understanding of the foundation of data and trust that Partial Autonomy (Level 3) requires, including the role humans play in building it.

When analysts work with AI assistance, they leave traces. Which queries they accept. Which results they act on. Which steps they modify or override. Over time, the system learns which investigation patterns the team trusts, which AI recommendations get acted on, and where analyst expertise is required – the kind of institutional knowledge that only comes from doing the work. Partial Autonomy is built on that record, not installed on top of an existing stack.

The path from AI-Assisted Operations to Partial Autonomy starts earlier than most organizations realize. It begins before they’re thinking about autonomy at all. Every assisted workflow is building toward what comes next.

What Holds Organizations Back

The primary barrier between AI-Assisted Operations (Level 2) and Partial Autonomy (Level 3) is accountability.

Consider how the automotive industry defined its equivalent of Partial Autonomy – SAE Level 3.

The designation applies only within specific, defined operational conditions. Outside those conditions, the human must take control. What qualifies a system for L3 is defined before autonomous operation begins: explicit parameters, a defined scope, and clear conditions for human override. Governance precedes autonomy.

Consider Waymo. It is the most capable autonomous system deployed at scale today — L3+ — operating without a safety driver under defined conditions. The vehicle is remarkable. But Waymo’s primary innovation is the organization built around it: the cloud infrastructure that keeps cars in autonomous condition, the human operations that handle exceptions the system cannot cover. The more autonomous the system, the more organizational maturity it required to build. High autonomy is an organizational capability.

The same logic applies in security operations. Accurate AI is the foundation. What makes Partial Autonomy legitimate is what gets built on top of it: defined rules of engagement, pre-approved policies, audit trails, and a clear organizational answer to who is responsible when an AI verdict is acted on. That accountability sits with the security team. When automation fires, it fires because someone made a deliberate governance decision to allow it. That is what makes it auditable, defensible, and durable.

Gartner’s readiness criteria for AI SOC deployments require that operational workflows be established in playbooks before AI is introduced₄. In practice, the organizations that advanced most consistently treated that requirement as a sequencing discipline, not a box to check. They defined their rules of engagement before turning on automated response.

The second learning was the attacker asymmetry. Defenders who stall between AI-Assisted Operations and Partial Autonomy have often done the validation work. The AI logic checks out. What remains is the decision to extend that trust to autonomous action — and that decision takes time. Attackers move differently. They deploy, observe what works, and iterate. Governance is an externality. Trial and error with no consequences for failure is a significant operational advantage. The gap between a defender’s trust-building timeline and an attacker’s operational tempo is structural. It compounds.

Why High Autonomy Stays on The Horizon. And Why That Matters Less Than We Thought

Eighteen months of deployment have also changed how we think about the upper end of the model.

When the original post was written, High Autonomy (Level 4) was described as dependent on a level of AI reasoning we hadn’t yet seen in production security environments. That remains the right framing. What’s changed is how close that horizon has become. Two years ago, asking a model to reason through a multi-stage attack, correlate signals across data sources, and produce an auditable verdict required significant scaffolding and produced inconsistent results. That’s no longer true. The gap between where AI was and where High Autonomy requires it to be has narrowed substantially.

High Autonomy still requires more than capable models. Institutional trust takes time to build. Accountability structures have to go beyond controlled tests to survive real incidents. Human oversight has to be redefined from reviewing individual actions to governing a system’s behavior within a defined scope. Those are organizational problems, technology doesn’t solve them. That work is already underway at Partial Autonomy (Level 3). What the road to High Autonomy requires is only visible from Partial Autonomy. Organizations that haven’t operated there yet are planning for a destination they haven’t seen. The knowledge of what it takes is path-dependent, and it emerges from operation, not from design.

As organizations move deeper into Partial Autonomy, the distinction between levels matters less in practice. What security leaders actually want is relevant control: governance over the decisions that matter, without being burdened by the ones that don’t. You cannot be responsible or accountable for a system that asks you to review everything.

Control over the right decisions is what matters. An analyst reviewing every alert has maximum control and minimum leverage. A system that acts autonomously on well-understood threat patterns, surfaces only the ambiguous and novel cases for human judgment, and maintains a complete audit trail, gives the analyst control over exactly what deserves their attention. That is a better and more focused version of human oversight.

High Autonomy, seen through this lens, is AI that has earned sufficient trust within a defined scope. The remaining human decisions are the ones that require human judgment, because the governance architecture evolved to allocate human attention correctly.

In the same way, a pilot does not manually adjust every control surface for the duration of a flight. They set the destination, define the parameters, and monitor the instruments. The system handles thousands of micro-corrections that would be impossible to manage directly. The pilot’s job is to govern the conditions under which the aircraft flies itself, not to manage every control input directly. Nobody describes this as a lack of pilot control. It is a better allocation of pilot judgment. And it works because of the environment surrounding the autopilot: pilot training standards, airline operational doctrine, air traffic control, and regulatory frameworks. The technology is one layer of a much larger system.

The governance work done at Partial Autonomy is the same work that produces High Autonomy. Organizations investing in it now are not waiting for a future capability release. They are building the foundation on which High Autonomy operates.

What This Means for the Road Ahead

The first step toward Partial Autonomy is a policy decision. Define the conditions under which your organization will allow a system to act: which response actions, against which threat types, within what scope, under whose authority. Write it down, however rough. That document is the actual starting point. Without it, the tooling is irrelevant.

The work at Partial Autonomy is real, meaningful, and available now. Security teams that define accountability structures before deploying autonomous systems, build a record of AI efficacy in their specific environment, and treat governance as a prerequisite rather than an afterthought, are the ones that reach and sustain Partial Autonomy. They are also the ones best positioned for what comes next. That work produces a more integrated SOC — data, AI, and response operating as a unified system.

High Autonomy remains the north star. This clearly articulated ideal state stops organizations from settling too early. It is the same function that “zero trust” serves as an architectural principle: no organization fully achieves it. Every organization is better for pursuing it.

The tools are capable. The frontier models have advanced significantly since our maturity model was first introduced. The capability gap that once made waiting feel reasonable has narrowed. What remains is the institutional work. That work is always harder than buying a product, which is why vendors who are honest about it are worth paying attention to.

SentinelOne customers operating the Autonomous SOC are seeing it in their numbers: 75% faster investigations, 4x more threats handled, 42% fewer false positives₅. Read the IDC Business Value Snapshot.

References

¹ Gartner, “AI SOC Agents: Harnessing Innovation, Managing Expectations,” Kevin Schmidt, Alex Tytarenko, Steve Santos, 25 February 2026. G00841784.

² Gartner, “AI SOC Agents: Harnessing Innovation, Managing Expectations,” Kevin Schmidt, Alex Tytarenko, Steve Santos, 25 February 2026. G00841784.

³ SAE International, “Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles,” SAE Standard J3016_202104, April 2021. https://www.sae.org/standards/content/j3016_202104/

⁴ Gartner, “AI SOC Agents: Harnessing Innovation, Managing Expectations,” Kevin Schmidt, Alex Tytarenko, Steve Santos, 25 February 2026. G00841784.

⁵ IDC Business Value Snapshot, “The Business Value of SentinelOne Singularity AI SIEM,” Michelle Abraham and Matthew Marden, May 2026, sponsored by SentinelOne. #US54435826-BVS.

Third-Party Trademark Disclaimer 

All third-party product names, logos, and brands mentioned in this publication are the property of their respective owners and are for identification purposes only. Use of these names, logos, and brands does not imply affiliation, endorsement, sponsorship, or association with the third party.