March 17, 2026

When Reality Diverges from the Playbook: Darktrace Identifies Encryption in a World Leaks Ransomware Attack

World Leaks, a rebrand of Hunters International, are known for their extortion-only attack model, abandoning the tactic of file encryption. However, contrary to these claims, Darktrace detected a World Leaks compromise where a ransomware payload was deployed, and customer data was encrypted.

Written by

Tiana Kelly

Senior Cyber Analyst & Team Lead

Inside the SOC

Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.

Written by

Tiana Kelly

Senior Cyber Analyst & Team Lead

Mar 2026

As-a-Service Cybercrime Models

As-a-Service cybercrime models reduce the barrier to entry for cyber criminals as they no longer need expertise in every domain. Threat actors can increasingly outsource or supplement missing skills through the broader cybercrime-as-a-service ecosystem, and thus these models continue to grow in popularity within the cybercriminal underground. This has led to multiple templates in this sphere, such as Phishing-as-a-Service, Botnet-as-a-Service, DDoS-as-a-Service, and notably Ransomware-as-a-Service (RaaS) [1].

What is Extortion-as-a-Service?

Extortion-as-a-Service (EaaS) businesses function as a formalized way for cyber threat actors to offer extortion services to others for a fee or profit share and represents an evolution of extortion operations from the double-extortion ransomware model. Advancing from the RaaS model, extortion has become a distinct profit stream, separate from the encryption payload. This separation of functions, data theft, negotiation, and publicity, sets the stage for EaaS [1].

The EaaS model reflects a broader trend in cybercriminal activity, in which threat actors increasingly prioritize data theft and public exposure over traditional ransomware encryption. This shift reduces operational complexity while increasing pressure on victims through reputational damage. This approach has become increasingly popular among threat actors as, unlike encryption-based attacks, these operations are more difficult to detect and remediate [2]. It reflects a trend of ‘hack-and-leak’ operations that prioritize stealth, speed, and reputational damage over traditional encryption-based ransomware attacks [3].

World Leaks Overview

World Leaks emerged in early 2024 as a direct rebrand of the Hunters International ransomware group, which was notorious for encrypting victims’ data and demanding payment for decryption keys. In mid-2025, Hunters International shifted to an extortion-only model due to law enforcement scrutiny and reduced profitability, rebranding itself as World Leaks.

World Leaks functions as an affiliate-based EaaS operation which provides proprietary Storage Software exfiltration tooling to affiliates while maintaining a four-platform infrastructure consisting of a main data leak site hosted on the Dark Web where victim data is published, a victim negotiation portal with live chat, an affiliate management panel, and an insider journalist platform granting media outlets 24-hour advance access to stolen data before public release [4]. Since its emergence, World Leaks has published data stolen from dozens of organizations globally on its data leak site, serving both as a pressure tactic and a means for building reputation among cyber criminals.

World Leaks (known associations include Hive Ransomware, Secp0 Ransomware, and UNC6148) have been known to target the industrial (manufacturing) sector, along with healthcare organizations, technology firms and more generally, industries with valuable intellectual property [4]. Victims targeted have spanned multiple countries, with most located in the US, as well as Canada and several countries across Europe [5].

World Leaks’ Tactics, Techniques, and Procedures (TTPs) [3][4]

World Leaks’ typical attack pattern involves the exploitation of credentials with inadequate access controls, e.g. lacking multi-factor authentication (MFA), moving through reconnaissance, lateral movement and data exfiltration, notably without an encryption element.

Initial Access:

Initial access is typically gained through the exploitation of compromised virtual private network (VPN) credentials lacking MFA through valid accounts, as well as phishing campaigns. The targeting of internet-facing VPN infrastructure, RDP, and public-facing applications also represent common attack vectors in World Leaks incidents.

Lateral Movement:

SMB, RDP, and SSH are used for lateral movement via remote services. Notably, the group is also known to use PsExec and Rclone as part of their lateral movement activities.

Persistence:

Registry key modifications, scheduled tasks creation, account manipulation.

Exfiltration:

Data exfiltration is carried out through custom storage software tooling via TOR connections. Cloud storage services used for exfiltration particularly include MEGA. World Leaks also carry out direct data transfer through established command-and-control (C2) infrastructure.

Unlike Hunters International, which combined encryption with extortion, World Leaks claims to have abandoned the use of encryption. Some reports note that operations since January 2025 represent a pivot toward eliminating encryption entirely, instead relying on custom exfiltration tooling with SOCKSv5 proxy and TOR-based communications [4]. However, in early 2026, Darktrace detected an incident that directly contradicted this claim: World Leaks carried out an attack that involved both the exfiltration and encryption of customer data.

Darktrace’s Coverage of World Leaks Ransomware

Organizations today face a growing challenge: keeping pace with increasingly fast-moving threats. This incident highlights a common problem, when time-limited mitigations expire or human security teams cannot respond quickly enough, attackers are often able to regain the upper hand. A recent Darktrace detection of World Leaks ransomware provides a clear example of this challenge in practice.

In January 2026, Darktrace identified the presence of ransomware and data encryption linked to World Leaks within the network of an organization within the healthcare sector. Although Darktrace’s Autonomous Response capability was active in the customer’s environment and initially blocking suspicious connectivity, buying time for the customer to remediate, the attack continued once these mitigative actions expired. Darktrace continued to apply Autonomous Response actions as the attack progressed, working to inhibit the attackers at each stage of the intrusion.

Investigations carried out by Darktrace revealed that threat actors likely gained initial access via a Fortigate appliance in mid-October, indicating a three-month dwell time, before employing living-off-the-land (LOTL) techniques for lateral movement. C2 communications were established using Cloudflare Tunnel (formerly Argo Tunnel). As part of the Actions on Objectives attack phase, a significant volume of data was exfiltrated to the MEGA cloud storage platform, followed by the encryption of customer data.

‍

Initial access/ Lateral movement

Darktrace analysts identified the likely patient-zero device within the network as a Fortigate appliance. In October 2025, this device was seen conducting brute-force activity using the compromised ‘administrator’ credential to gain a foothold deeper within the customer’s environment. Masquerading as a privileged user, the threat actor then went on to launch activity on remote devices via PsExec, a common administrative tool that allows users to execute processes on remote systems without manually installing client software, providing significant power to attackers when abused. Around the time, Darktrace detected an unknown device on the network attempting to authenticate via NTLM. As this device had not previously been seen on the network, it likely belonged to the attacker.

Reconnaissance

As part of the reconnaissance phase of the attack, port and network scanning was carried out in an attempt to identify open UDP and TCP ports within the network.

Lateral movement & C2

Around one month after entering the customer’s network, the World Leaks threat actors began tunnelling activity using Cloudflare Tunnel. Darktrace detected connections to several hostnames including: region2.v2.argotunnel[.]com; h2.cftunnel[.]com; region1.v2.argotunnel[.]com. This tunnelling activity continued until January of 2026, when encryption occurred. Cloudflare tunnels are known to be abused by attackers as they enable the use of temporary infrastructure to scale operations, allowing rapid deployment and teardown. Furthermore, leveraging of Cloudflare’s infrastructure to create these rate-limited tunnels (used to relay traffic from an attacker-controlled server to a local machine) makes such malicious activity harder to detect by both defenders and traditional security measures, particularly those that rely on static blocklists [6].

Further lateral movement was carried out using common remote management tools such as Windows Remote Management (WinRM) RDP, allowing the World Leaks threat actors to access local devices within the victim organization’s network.

As this attack progressed, Darktrace detected multiple files being written over SMB. These files included Windows\Temp\chromeremotedesktophost.msi, which was written from the patient-zero device to another internal device as part of lateral movement efforts. Following this transfer, and prior to subsequent data exfiltration activity, a network server was observed connecting to the hostname remotedesktop-pa[.]googleapis[.]com, an API endpoint required for Chrome Remote Desktop, indicating that Chrome RDP was used by the threat actor in this stage of the attack.

Other files written over SMB included the script programdata\syc\OpenSSHUtils.psm1 (which can be used legitimately to configure OpenSSH) and the executable programdata\syc\ssh‑sk‑helper.exe (a legitimate OpenSSH component used to support security keys). These files were written from the suspected patient‑zero device to an internal domain controller using the ‘administrator’ credential.

Thereafter, SSH connections to external IP address 51.15.109[.]222 were observed, providing another channel between the malicious actors and victim machines. Darktrace recognized that the use of SSH by the devices seen connecting to this IP address was highly anomalous, indicating that this suspicious activity formed part of the attack.

Writes of the script programdata\syc\OpenSSHUtils.psm1 were also observed into January, highlighting the continuation of the attack that had begun three months earlier.

On December 19 and 20, Darktrace detected a DNS server within the customer’s network making anomalous outgoing connections to an external IP address not previously seen in the environment: 193.161.193[.]99. This IP address has been reported by open-source-intelligence (OSINT) as being associated with C2 infrastructure, having been linked to several remote access trojans (RATs) and botnets in the past.

This activity a shift towards the infrastructure-as-a-service (IaaS) model, underscoring the growing trend around As-a-Service Cybercrime models and the increasing the industrialization of botnets. The presence of extensive digital botnets, often leased to other criminal organizations, means the group gaining initial access is not necessarily the same group conducting ransomware deployment or data theft; botnets now act as shared underlying infrastructure enabling multiple forms of cybercriminal activity [7].

Furthermore, connections to this IP address (193.161.193[.]99) were made over port 1194, which is associated with OpenVPN, suggesting that World Leaks may have leveraged it to obfuscate C2 communication with attacker-controlled infrastructure.

*Figure 1: Darktrace’s detection of the IP address 193.161.193[.]99, noting that it was first seen within the customer’s network on December 19, 2025.*

Data exfiltration

In November, Darktrace detected the threat actors carrying out one of their Attack on Objective tactics: data exfiltration. Multiple local devices within the compromised network began transferring data to Backblaze and MEGA domains, both of which provide cloud storage services; 80+GB of data was transferred to MEGA in late December 2025. Endpoints associated with this activity included: backblazeb2[.]com and gfs302n520[.]userstorage[.]mega[.]co[.]nz, as well as related user agents such as AS40401 BACKBLAZE) and MegaClient/10.3.0/64.

Notably, Darktrace researchers identified two known World Leaks TTPs in this attack: the use of MEGA, a known tool abused by the group, and Rclone, a command-line tool used to manage files on cloud storage, which was observed in the user agent of the MEGA data-transfer connections: rclone/v1.69.0 [4].

Figure 2: Cyber AI Analyst Incident highlighting data upload activity to backblaze[.]com endpoints.\

Ransomware deployment & encryption

The encryption stage of this attack was confirmed by the presence of a ransom note found on the network in a file with a seemingly randomized nine-character string preceding README.txt, attributing the incident to World Leaks, along with an extension with the same nine characters appended to encrypted files. Darktrace also observed SMB writes of files named world.exe and task.bat, with the compromised ‘localadmin’ credential used during the SMB logins. It is likely that these files served as the vector for the ransomware payload.

*Figure 3: Packet Capture (PCAP) of the ransom note claiming that the attack was carried out by World Leaks.*

Conclusion

Though traditional ransomware relies on encryption, recent trends show that cyber threat actors no longer need to rely on noisy encryption tools and can eliminate much of the risk and technical complexity associated with encrypting systems. This is the model reportedly preferred by World Leaks after their rebrand from Hunters International.

In addition to reducing noise around these attacks, extortion‑only operations may be favored by threat actors over encryption‑focused ones for several reasons, including the fact that traditional security tools may struggle to detect data theft compared to encryption, that attackers leave less evidence behind when encryption is avoided, and that the long‑term impacts of stolen data on organizations can be greater than the loss of systems caused by encryption processes, which can be restored [8]. This is supported by analysis of data leak sites suggesting that almost 1,500 incidents in 2025 relied on data theft alone. Attackers can simply steal victim data and attempt to extort a ransom by threatening to publish it, without needing to deploy ransomware at all [9]. Furthermore, although World Leaks aims to function as an affiliate‑based EaaS operation, security teams should remain aware that their affiliates may have different criminal objectives.

Contrary to reports that World Leaks’ typical attack style has an extortion‑only objective, Darktrace detected an incident in which a World Leaks attack did end with the encryption of customer data. This highlights the need for adaptive defenses and reinforces the importance of network defenders staying proactive in the face of attacks, particularly as they may progress in ways that are unexpected compared to previous trends associated with a given threat actor.

Credit to Tiana Kelly (Senior Cyber Analyst and Analyst Manager) and Emily Megan Lim (Senior Cyber Analyst)

Edited by Ryan Traill (Content Manager)

Appendices

IoCs

world.exe – Executable File – Possible Ransomware Payload
task.bat – Script File – Possible Ransomware Payload
‘^[A-Z][a-z]{3}[A-Z][a-z][A-Z]{3}[.]README[.]txt' – Ransom Note
[.]^[A-Z][a-z]{3}[A-Z][a-z][A-Z]{3} – Ransomware file extension

· 51.15.109[.]222 – IP Address - Possible C2 Infrastructure

· 193.161.193[.]99 – IP Address – Probable C2 Infrastructure

Darktrace Model Detections (Enhanced Monitoring models denoted with an asterisk)

· Device / Attack and Recon Tools

· Device / Suspicious SMB Scanning Activity

· Device / Anomalous NTLM Brute Force

· Compliance / Connection to Tunnelling Service

· Device / Suspicious New User Agents

· Device / New or Unusual Remote Command Execution

· Compliance / SMB Drive Write

· Anomalous Connection / Uncommon 1 GiB Outbound

· Compromise / Ransomware / Ransom or Offensive Words Written to SMB

· Device / Multiple Lateral Movement Model Alerts*

· Device / SMB Lateral Movement

· Unusual Activity / Sustained Anomalous SMB Activity

· Device / Large Number of Model Alerts

· Compromise / Ransomware / SMB Reads then Writes with Additional Extensions

· Compromise / Ransomware / Suspicious SMB Activity*

· Anomalous File / Internal / Additional Extension Appended to SMB File

· Unusual Activity / SMB Access Failures

· Unusual Activity / Enhanced Unusual External Data Transfer*

· Device / Suspicious File Writes to Multiple Hidden SMB Shares

· Anomalous Server Activity / Rare External from Server

· Unusual Activity / Unusual Mega Data Transfer*

· Device / Possible SMB/NTLM Brute Force

· Anomalous Connection / Unusual Admin RDP Session

· Anomalous Connection / Active Remote Desktop Tunnel

· Anomalous Connection / Data Sent to Rare Domain

· Anomalous Connection / New or Uncommon Service Control

· Anomalous Connection / New or Uncommon Service Enumeration

· Anomalous Connection / Rare WinRM Outgoing

· Anomalous Connection / SMB Enumeration

· Anomalous Connection / Unusual Admin RDP Session

· Anomalous Connection / Unusual Incoming Long Remote Desktop Session

· Anomalous Connection / Upload via Remote Desktop

· Anomalous File / Internal / Executable Uploaded to DC

· Anomalous File / Internal / Unusual SMB Script Write

· Compliance / SSH to Rare External Destination

· Device / Anomalous Github Download

· Device / Anonymous NTLM Logins

· Device / Network Scan

· Device / New or Uncommon WMI Activity

· Device / New User Agent To Internal Server

· Device / Possible Brute-Force Activity

· Device / RDP Scan

· Device / SMB Session Brute Force (Admin)

· Device / SMB Session Brute Force (Non-Admin)

· Device / Suspicious Network Scan Activity

· Unusual Activity / Successful Admin Brute-Force Activity

· Unusual Activity / Unusual External Data to New Endpoint

· Unusual Activity / Unusual External Data Transfer

· Unusual Activity / Unusual File Storage Data Transfer

· User / New Admin Credentials on Server

Cyber AI Analyst Incidents

· Scanning of Multiple Devices

· Large Volume of SMB Login Failures to Multiple Devices

· Suspicious Chain of Administrative Connections

· SMB Write of Suspicious File

· Suspicious DCE-RPC Activity

· Unusual External Data Transfer

· Unusual External Data Transfer to Multiple Related Endpoints

· Unusual External Data Transfer to Endpoints

MITRE ATT&CK Mapping

· Initial Access – T1190 – Exploit Public-Facing Application

· Defense Evasion, Initial Access, Persistence, Privilege Escalation – T1078 – Valid Accounts

· Resource Development – T1588.001 – Obtain Capabilities: Malware

· Reconnaissance – T1590.005 – Gather Victim Network Information: IP Addresses

· Reconnaissance – T1592.004 – Gather Victim Host Information: Client Configurations

· Reconnaissance – T1595.001 – Active Scanning: Scanning IP Blocks

· Reconnaissance – T1595.002 – Active Scanning: Vulnerability Scanning

· Reconnaissance – T1595.003 – Active Scanning: Wordlist Scanning

· Discovery – T1018 – Remote System Discovery

· Discovery – T1046 – Network Service Discovery

· Discovery – T1083 – File and Directory Discovery

· Discovery – T1135 – Network Share Discovery

· Command and Control – T1219 – Remote Access Tools

· Command and Control – T1219.002 – Remote Access Tools: Remote Desktop Software

· Command and Control – T1571 – Non-Standard Port

· Command and Control – T1572 – Protocol Tunneling

· Command and Control – T1573.001 – Encrypted Channel: Symmetric Cryptography

· Credential Access – T1110 – Brute Force

· Credential Access – T1110.001 – Brute Force: Password Guessing

· Defense Evasion – T1006 – Direct Volume Access

· Defense Evasion – T1564.005 – Hide Artifacts: Hidden File System

· Defense Evasion – T1564.012 – Hide Artifacts: File/Path Exclusions

· Execution – T1047 – Windows Management Instrumentation

· Execution – T1569.002 – System Services: Service Execution

· Lateral Movement – T1021 – Remote Services

· Lateral Movement – T1021.001 – Remote Services: Remote Desktop Protocol

· Lateral Movement – T1021.002 – Remote Services: SMB/Windows Admin Shares

· Lateral Movement – T1021.006 – Remote Services: Windows Remote Management

· Lateral Movement – T1080 – Taint Shared Content

· Lateral Movement – T1210 – Exploitation of Remote Services

· Lateral Movement – T1570 – Lateral Tool Transfer

· Collection – T1039 – Data from Network Shared Drive

· Collection – T1074 – Data Staged

· Exfiltration – T1041 – Exfiltration Over C2 Channel

· Exfiltration – T1048 – Exfiltration Over Alternative Protocol

· Exfiltration – T1567.002 – Exfiltration Over Web Service: Exfiltration to Cloud Storage

References

[1] https://www.levelblue.com/blogs/levelblue-blog/extortion-as-a-service-the-latest-threat-actor-criminal-ecosystem/

[2] https://blackpointcyber.com/wp-content/uploads/2025/12/World-Leaks.pdf

[3] https://blackpointcyber.com/threat-profile/world-leaks-ransomware/

[4] https://www.halcyon.ai/threat-group/worldleaks

[5] https://www.moxfive.com/resources/moxfive-threat-actor-spotlight-world-leaks

[6] https://thehackernews.com/2024/08/cybercriminals-abusing-cloudflare.html

[7] https://www.trendmicro.com/vinfo/tw/security/news/threat-landscape/the-industrialization-of-botnets-automation-and-scale-as-a-new-threat-infrastructure

[8] https://www.morphisec.com/blog/ransomware-without-encryption-why-pure-exfiltration-attacks-are-surging-and-why-theyre-so-hard-to-catch/

[9] https://sed-cms.broadcom.com/sites/default/files/2026-01/RWN-2026-WP100_1.pdf

執筆者

Tiana Kelly

Senior Cyber Analyst & Team Lead

Inside the SOC

Written by

Tiana Kelly

Senior Cyber Analyst & Team Lead

Watch the NIS2 Webinar

Blog

AI

June 25, 2026

Shadow AI Detection: The First Step Toward Securing AI

Why shadow AI is emerging

Imagine you’re an employee under pressure, deadlines stacking up, repetitive tasks piling higher by the day. You find a free AI tool online that promises to automate the work in seconds; no approvals are needed. It feels like a simple win, paste in some data, write a quick prompt, and move faster.

But in that moment, something changed.

Sensitive customer information is entered into a tool your organization doesn’t monitor, doesn’t govern, and can’t see and suddenly, that data is no longer where it should be, and no one knows where it’s gone.

This is the reality of Shadow AI: employees using unsanctioned AI tools to move faster, while unintentionally creating risk that exists entirely outside visibility and control.

This is not just a one off case, research across businesses indicate that nearly half of employees report using unsanctioned AI tools, often prioritizing speed and productivity over security. Additionally, 51% of employees report connecting AI tools to work systems or apps without IT approval, creating significant operational risk where the average cost of security incidents in organizations with a high level of shadow AI usage can reach $670k.

While shadow AI is often top of mind for security professionals, it is just one component of how AI use can increase risk. Understanding and managing shadow AI use should be considered as part of a broader, comprehensive risk management strategy that aims to secure AI systems, including human and agent identities, interactions, human-AI partnerships, and behaviors operating across the digital enterprise from visibility and governance through detection, response, and recovery.

Effective risk management calls for a layered and interdisciplinary strategy. It requires addressing issues across governance and visibility; identity, access and agent control, data security and privacy, secure MLOps / LLMOps, runtime security, behavior-based detection, autonomous response and recovery.

This blog explores a specific governance and visibility use case linked to shadow AI and reveals the challenges it presents as well as the defensive strategies that security teams can adopt.

Why shadow AI is hard to detect

When it comes to AI, what organizations can easily see does not always reflect the full scope of AI activity occurring within the tools, applications, and workflows used across an enterprise. As a result, organizations using traditional rule-based methods to flag unusual activity may struggle to distinguish unsanctioned AI usage from legitimate operational behavior, particularly as SaaS applications, APIs, and orchestration layers increasingly have AI embedded into normal business workflows. Identifying threats using previously observed intelligence or depending on hard to maintain allow and block lists does not provide a dynamic enough strategy to manage risk. Also, many organizations are focusing on identifying Shadow AI in their governed infrastructure, like gateways, endpoints, or SASE, which is foundational. But, organizations require visibility and Shadow AI detection across all networked infrastructure from on-prem, hybrid, data centers, and cloud infrastructure that may not have endpoint agent visibility. This uncovers the utilization of MCP, data flows, and autonomous agents across these domains.

For example, employees interact with AI assistants across approved SaaS platforms every day. However, browser extensions and other types of plug-ins can route prompts that include enterprise data to embedded AI services in ways that are not visible to the security team. AI enabled workflows may invoke multiple APIs, orchestration layers, and cloud services behind the scenes, making it difficult for traditional security tooling to determine where data is processed, stored, or retransmitted. Because much of this activity occurs within trusted browser sessions and encrypted SaaS traffic, conventional network monitoring, DLP, and application allowlisting controls often lack the context needed to accurately identify or govern these interactions

Identifying AI tools in the environment is one part of the equation. Understanding the behavior surrounding their use is where the real challenge lies. An AI application is not inherently risky, but the way users or other assets interact with it may be. Sensitive data exposure, abnormal access patterns, and misuse of AI-assisted workflows often appear legitimate in isolation and only become visible through behavioral analysis across the broader environment.

What Shadow AI visibility does and doesn’t show

Comprehensive Shadow AI visibility allows organizations to answer several important questions:

What types of AI are we using? What AI platforms, agents, MCP clients/servers, and services are active across the enterprise?

Who is using AI services? Which users, business units, or systems are interacting with those AI services?

Is our data safe? Is sensitive or regulated data being exposed through prompts, workflows, or integrations?

Are AI systems behaving as expected? Are AI systems behaving anomalously or operating outside approved governance processes?

Are our AI systems under attack? Is an attacker attempting to manipulate prompts, influence agent behavior, or abuse AI-enabled workflows?

Answering these questions is foundational to broader AI governance efforts. However, it is limited to helping teams understand initial interactions and fails to offer insight into dependencies and outcomes that are critical to securing AI across an enterprise.

Deeper visibility that includes the ability to understand dependencies and outcomes are not always available in AI security point products. Answering the questions below requires understanding runtime behavior and operational outcomes:

What actions did the AI interaction trigger?

What systems, applications, or data did it access? Did the AI operate beyond its intended permissions or scope?

Could a low-risk interaction lead to high-risk outcomes?

What is the risk and context understanding of an anomalous activity to assist in prioritization of analysis and autonomous response action?

The distinction between these two sets of questions offers two different layers of AI security. The first set of questions focuses on discovery and interaction visibility. The second set focuses on providing visibility that includes the context and outcomes that are critical for managing follow-on risks associated with obfuscated downstream activities.

Together, these layers help organizations move beyond simply identifying AI usage toward understanding how AI behaves operationally across the enterprise.

How organizations are addressing shadow AI

Most organizations still approach shadow AI as an application control problem, relying on policies, browser restrictions, and allow/block lists. However, AI adoption is evolving faster than most governance processes can realistically keep pace with. New assistants, plugins, and embedded AI features appear continuously, creating pressure to enable business productivity while simultaneously containing risk.

Existing governance processes were designed for a more traditional SaaS adoption cycle, where new applications could be reviewed, approved, and monitored over longer time horizons. AI adoption operates differently. New capabilities can appear overnight inside existing platforms employees already use, making it difficult for security and governance teams to maintain an accurate understanding of enterprise AI exposure. This means that many organizations are experiencing significant operational overhead, particularly in large environments where AI usage is decentralized across teams, departments, and third-party services.

Where should organizations start when securing their AI systems?

Shadow AI identification is an on-going critical component for AI Risk/Governance Boards as well as security organizations. As organizations seek AI certifications like ISO 42001 AI Management Systems, visibility into all AI adoption from enterprise use to custom innovation and development is crucial. Shadow AI identification provides organizations with the visibility needed to decide whether an AI tool should be brought into governed environments to reduce data loss (DLP) risks or whether policies should be established and enforced to restrict their use.

As organizations rapidly innovate and adopt AI, they are taking on more and more risk. Organizations need to have a strategy in place to mitigate the assumed risk, especially with third-party adoption. Visibility, monitoring, governance enforcement, behavioral-based detection of non-deterministic systems, and autonomous investigation and containment becomes critical to mitigating the risk of AI systems.

How Darktrace secures AI and shadow AI

Attackers are using AI to move faster, scale tactics, and make threats more adaptive and convincing. Internally, organizations are grappling with new forms of risk created by generative AI, autonomous agents, shadow AI, and increasingly complex digital environments.

Darktrace helps organizations protect both people and AI in a world where AI is now central to how business gets done. Darktrace / SECURE AI helps organizations discover and control shadow AI by surfacing unsanctioned or unexpected AI activity where it appears – including MCP detections, distinguishing misuse of legitimate tools and unapproved services, and applying policy to contain data exposure while guiding users toward sanctioned options.

Stay up to date on AI security

Sign up for the Secure AI Readiness Program here: This gives you exclusive access to the latest news on the latest AI threats, updates on emerging approaches shaping AI security, and insights into the latest innovations, including Darktrace’s ongoing work in this area.

Ready to talk with a Darktrace expert on securing AI? Register here to receive practical guidance on the AI risks that matter most to your business, paired with clarity on where to focus first across governance, visibility, risk reduction, and long-term readiness.

About the author

Blog

Network

June 25, 2026

From Click to Command: Behavioral Detection of AppleScript-Led MacOS Intrusions

Introduction

Darktrace’s Threat Research team is publishing this analysis to help defenders understand an active pattern of macOS tradecraft observed in multiple customer environments. This post summarizes the behaviors observed, how they were assessed, and what defenders can do now.

Across multiple environments, Darktrace observed a consistent MacOS intrusion pattern beginning with ClickFix-style user-assisted “update” execution and transitioning into AppleScript-driven post-compromise activity and sustained outbound signaling.

While individual indicators were low-confidence, the repeated convergence of weak behavioral signals — including HTTP POST beaconing, rare or IP-only destinations, SSL anomalies, and abnormal client characteristics — provided a defensible indication of command-and-control establishment Darktrace detection and response in these cases was driven by behavior over artifacts. In the highest-confidence instances, automated containment disrupted outbound signaling before sustained tasking could occur.

Background

ClickFix-style activity typically relies on user-assisted execution and plausible “update” pretexting, followed by post-execution use of native tools to keep the footprint light. In MacOS environments, AppleScript and other built-in scripting mechanisms enable flexible post-compromise workflows while minimizing stable file-based indicators.

Following execution, affected devices exhibited a consistent behavioral pattern. AppleScript or equivalent native scripting activity was observed initiating follow-on workflows, after which outbound communications began to establish a structured rhythm.

These communications were characterized by repeated HTTP POST requests to low-prevalence or IP-only endpoints, often combined with unusual SSL properties and client identifiers that diverged from baseline device behavior. Individually, these signals were weak. When correlated across time and devices, they formed a pattern consistent with control establishment rather than benign software activity.

In higher-confidence cases, Autonomous Response actions were able to reduce or halt outbound signaling, interrupting the attacker’s ability to maintain control.

Detection Timeline

In representative cases, the sequence unfolded as follows:

Stage 1 – Initial Execution

Initial activity began with suspicious or masqueraded execution on a MacOS endpoint, consistent with ClickFix-style user deception.

Stage 2 – Post-Execution Scripting

This was followed closely by native scripting activity, most commonly AppleScript, indicating the transition into post-execution workflow.

Stage 3 – Outbound Communications

Outbound communications then emerged, initially sporadic but quickly forming a consistent cadence of HTTP POST requests to rare external endpoints.

Stage 4 – Anomaly Convergence

As activity persisted, additional anomalies became visible — unusual SSL characteristics, abnormal user agents, and connections to infrastructure with no prior network prevalence.

Stage 5 – Autonomous Response

In the most mature stages of the activity, automated containment actions disrupted outbound communications on affected devices, limiting the attacker’s ability to continue tasking while investigations progressed.

Darktrace coverage and detections

The following use-case highlights systems likely affected by malicious macOS intrusion activity linked by Microsoft to the Democratic People’s Republic of Korea (DPRK) [1], with indications of suspicious behavior observed between March 1 and May 3, 2026. The activity overlaps with patterns described in recent reporting on DPRK-nexus MacOS intrusions [1], though attribution confidence in this case remains moderate and based on behavioral alignment rather than solely infrastructure linkage.

Analyst confidence emerged through the correlation of multiple weak signals across time and devices. This included model coverage for rare external communications, sustained beaconing patterns, repeated HTTP POSTs, and anomalous client characteristics. Where enabled, Autonomous Response actions disrupted the most active outbound paths to reduce the attacker’s ability to maintain control while Darktrace’s investigation continued.

Notably, this highly anomalous behavior included:

Outbound connections to the rare external endpoint, zoom[.]uswebob[.]us associated with IP address, 148.72.73[.]98 [2][3] over port 443
Outbound connections to the rare external endpoint, check02id[.]com associated with IP address, 83.136.210[.]180 [4] over port 7365
Outbound connections to the rare external endpoints, 104.145.210[.]107 [5] over port 8443 and 83.136.208[.]48 [6] over port 443
Outbound connections to the rare external endpoint, 83.136.208[.]246 [7] over port 6783 with observed URI `/api/daemon` and a PowerShell user agent

Darktrace’s detection initially highlighted a desktop device (running MacOS) engaging in anomalous behavior as early as March 12, 2026. Starting on March 12, the source device triggered a ‘Possible Doppelganger Attack’ alert including connectivity to the hostname "zoom[.]uswebob[.]us · 148.72.73[.]98" over port 443 (TCP, HTTPS, H2). This model highlights a device connecting to a location that is rare but masquerades as legitimate software, such as Zoom in this case, a commonly used technique to blend into expected traffic [2] [3].

Figure 1: Initial connectivity observed to the rare external hostname, zoom[.]uswebob[.]us · 148.72.73[.]98, over port 443.

This was followed roughly seven later by a connection to 104.145.210[.]107 over port 8443, during which approximately 250 KiB of data of inbound data and 30 MiB of outbound data was observed, triggering the ‘Unusual Activity / Unusual External Data to New Endpoint’ in Darktrace.

Quickly after this connection, Darktrace’s Autonomous Response intervened, blocking the device’s access to the unusual external location and halting the data exfiltration attempt.

Figure 2: Darktrace’s detection of unusual data exfiltration, shortly followed by an Autonomous Response action to block it.

The device continued to consistently trigger model alerts relating to unusual external connectivity, including 'Posting HTTP to IP Without Hostname', 'Anomalous Connection / Rare External SSL Self-Signed' alerts, until well after 3 PM that day.

Figure 3: Additional external connectivity to new IP without a hostname, including connectivity to 83.136.208[.]246, alongside an anomalous ‘curl/8.7.1’ user agent and ‘/api/daemon’ URI.

Figure 4: Continued external SSL connectivity to IP 83.136.208[.]48, including connectivity to 83.136.208[.]246, alongside an anomalous ‘curl/8.7.1’ user agent and ‘/api/daemon’ URI.

Figure 5: Continued external HTTP connectivity to hostname, check02id[.]com · 83.136.210[.]180, alongside an anomalous ‘Go-http-client/1,1’ user agent.

From March 13 to March 28, the device continued exhibit unusual connectivity to various endpoints (e.g., 83.136.208[.]48, 83.136.208[.]246, check02id[.]com · 83.136.210[.]180), with the 'Multiple HTTP POSTs to Rare Hostname' model consistently triggering.

Windows OS Case

Pivoting over to an additional device, this time running Windows OS, anomalous behavior was also observed between March 30 and April 20. Notably, on March 30, the device was observed making a large number of suspicious external connection attempts to 83.136.208[.]246 over port 6783, all of which failed.

A further indicator was observed on April 1 with PowerShell connectivity to the same rare endpoint (83.136.208[.]246, port 6783), using the URI '/api/daemon' and the user agent 'Mozilla/5.0 (Windows NT; Windows NT 10.0; fr-FR) WindowsPowerShell/5.1.26100.7920'. Additional alerts included 'New User Agent to IP Without Hostname' and 'Anomalous Github Download', alongside activity involving the same endpoint.

Figure 6 : ‘Anomalous Powershell to Rare External Destination’ and ‘Github Download’ model alerts. This behavior involved connectivity with the endpoints ‘83.136.208[.]246’ and ‘github[.]com’.

The device continued triggering 'Posting HTTP to IP Without Hostname' & 'PowerShell to External Rare' alerts between April 4 and April 20 across multiple related endpoints (i.e., 83.136.208[.]48, 83.136.208[.]246, check02id[.]com · 83.136.210[.]180).

Darktrace’s Autonomous Response capability was able to block suspicious PowerShell attempts to unusual external locations, as shown below in an example from April 20.

Figure 7: Autonomous Response intervening to block an unusual PowerShell connection to an external destination.

Cyber AI Analyst investigations

In higher-confidence instances, Darktrace’s Cyber AI Analyst investigations helped connect otherwise separate model alerts into a single incident narrative, highlighting the attacker’s progression from post-execution scripting into sustained outbound signaling. This contextual stitching is particularly valuable in macOS scenarios where static artefacts are limited, and behavioral sequencing defines the intrusion.

Cyber AI Analyst investigations highlighted alerts on March 12, including unusual repeated connections and possible SSL command-and-control (C2) to multiple endpoints:

Figure 8: Cyber AI Analyst investigation linking events into a unified incident.

Autonomous Response

In addition to the containment actions detailed earlier, Autonomous Response implemented multiple additional measures to contain suspicious activity throughout the course of this attack. Whenever unusual external connectivity was detected, Darktrace blocked it, closing down potential C2 channels. Likewise, when data exfiltration attempts were identified, these connections were stopped to prevent the potential loss of sensitive data.

Furthermore, in cases where a device was deemed to have carried out a significant number of anomalous activities, Darktrace enforced a “pattern of life” on the device, preventing it from deviating from its expected behavior while allowing legitimate business operations to continue uninterrupted.

Figure 10: Autonomous Response actions implemented by Darktrace in response to suspicious connectivity in April, including the “Enforce Pattern of Life” action.

Conclusion

macOS intrusion tradecraft continues to shift toward native tooling and lightweight control channels designed to evade signature-led controls.

The repeated convergence of rare destinations, POST-based signaling, and anomalous client behavior — observed across time and across devices — provided sufficient evidence to act early and with confidence.

As macOS tradecraft continues to evolve, the defender advantage increasingly lies not in signatures, but in the ability to reason from behavior.

‍

Credit to Justin Torres (Senior Cyber Analyst), Nathaniel Jones (VP, Security & AI Strategy, FCISO)

Edited by Ryan Traill (Content Manager)

Appendices

Darktrace Model Alert Coverage:

/ NETWORK-based model alerts:

· Anomalous Connection::Multiple HTTP POSTs to Rare Hostname

· Anomalous Connection::Rare External SSL Self-Signed

· Anomalous Connection::Powershell to Rare External

· Anomalous Connection::New User Agent to IP Without Hostname

· Anomalous Connection::Posting HTTP to IP Without Hostname

· Compromise::Fast Beaconing to DGA

· Compromise::Large Number of Suspicious Failed Connections

· Device::Anomalous Github Download

· Device::New PowerShell User Agent

· Unusual Activity::Unusual External Data to New Endpoint

/ NETWORK-based Autonomous Response model alerts:

· Antigena / Network::Significant Anomaly::Antigena Significant Anomaly from Client Block

· Antigena / Network::Significant Anomaly::Antigena Controlled and Model Breach

· Antigena / Network::Significant Anomaly::Antigena Breaches Over Time Block

Indicators of Compromise (IoCs)

IP/Hostname:

· zoom[.]uswebob[.]us · 148.72.73[.]98

· 83.136.208[.]246

· check02id[.]com · 83.136.210[.]180

· 83.136.208[.]48

· 104.145.210[.]107

URIs:

· /api/daemon

Destination Port Usage:

· 6783

· 5202

· 443

· 7365

· 8443

ASN:

· AS400897 PETROSKY

· AS398256 AS-ULTAHOST

User agents:

· Mozilla/5.0 (Windows NT; Windows NT 10.0; fr-FR) WindowsPowerShell/5.1.26100.7920

· Go-http-client/1.1

· curl/8.7.1

MITRE ATT&CK Mapping

(Technique Name - Tactic - ID - Sub-Technique of)

· Browser Session Hijacking - COLLECTION - T1185

· Web Protocols - COMMAND AND CONTROL - T1071.001 - T1071

· Install Digital Certificate - RESOURCE DEVELOPMENT - T1608.003 - T1608

· PowerShell - EXECUTION - T1059.001 - T1059

· Domain Generation Algorithms - COMMAND AND CONTROL - T1568.002 - T1568

· Non-Standard Port - COMMAND AND CONTROL - T1571

· Malware - RESOURCE DEVELOPMENT - T1588.001 - T1588

· Web Service - COMMAND AND CONTROL - T1102

· Code Repositories - COLLECTION - T1213.003 - T1213

· Exploitation of Remote Services - LATERAL MOVEMENT - T1210

· Exfiltration Over C2 Channel - EXFILTRATION - T1041

· Exfiltration to Cloud Storage - EXFILTRATION - T1567.002 - T1567

References:

[1] https://www.microsoft.com/en-us/security/blog/2026/04/16/dissecting-sapphire-sleets-macos-intrusion-from-lure-to-compromise/

[2] https://radar.securityalliance.org/advisory-on-dprk-unc1069-fake-microsoft-teams-and-zoom-calls/

[3] https://www.virustotal.com/gui/domain/uswebob.us

[4] https://www.virustotal.com/gui/ip-address/83.136.210.180/community

[5] https://www.virustotal.com/gui/ip-address/104.145.210.107/community

[6] https://www.virustotal.com/gui/ip-address/83.136.208.48/community

[7] https://www.virustotal.com/gui/ip-address/83.136.208.246/community

[8] https://www.darktrace.com/blog/applescript-abuse-unpacking-a-macos-phishing-campaign

About the author

Justin Torres

Cyber Analyst

あなたのデータ × DarktraceのAI

唯一無二のDarktrace AIで、ネットワークセキュリティを次の次元へ

デモを予約

Check out this article by Darktrace: Darktrace Identifies Encryption in a World Leaks Ransomware Attack

When Reality Diverges from the Playbook: Darktrace Identifies Encryption in a World Leaks Ransomware Attack

As-a-Service Cybercrime Models

What is Extortion-as-a-Service?

World Leaks Overview

World Leaks’ Tactics, Techniques, and Procedures (TTPs) [3][4]

Initial Access:

Lateral Movement:

Persistence:

Exfiltration:

Darktrace’s Coverage of World Leaks Ransomware

Initial access/ Lateral movement

Reconnaissance

Lateral movement & C2

Data exfiltration

Ransomware deployment & encryption

Conclusion

Appendices

IoCs

Darktrace Model Detections (Enhanced Monitoring models denoted with an asterisk)

MITRE ATT&CK Mapping

References

From Click to Command: Behavioral Detection of AppleScript-Led MacOS Intrusions

Shadow AI Detection: The First Step Toward Securing AI

ブログをお楽しみ頂けましたか？

More in this series

From Click to Command: Behavioral Detection of AppleScript-Led MacOS Intrusions

Hola VPN Abuse: From Proxy Traffic to Malware and Cryptomining

How Attackers Abuse the Chinese Nezha Monitoring Tool

Blog

AI

June 25, 2026

Shadow AI Detection: The First Step Toward Securing AI

Why shadow AI is emerging

Why shadow AI is hard to detect

What Shadow AI visibility does and doesn’t show

How organizations are addressing shadow AI

Where should organizations start when securing their AI systems?

How Darktrace secures AI and shadow AI

Stay up to date on AI security

Blog

Network

June 25, 2026

From Click to Command: Behavioral Detection of AppleScript-Led MacOS Intrusions

Introduction

Background

Detection Timeline

Darktrace coverage and detections

Windows OS Case

Cyber AI Analyst investigations

Autonomous Response

Conclusion

Appendices

Darktrace Model Alert Coverage:

Indicators of Compromise (IoCs)

MITRE ATT&CK Mapping

References: