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
The Darktrace Community
Share
20
Mar 2024
The state of email security and phishing attacks
Employees send and receive hundreds of emails a day to keep businesses moving. Unfortunately, it just takes one employee to interact with an undetected phishing email to potentially put an entire organization at risk from cyber disruption. Attackers know this, which is why they continue to develop and improve email phishing attacks.
Increased attack sophistication makes it harder than ever for traditional cyber security solutions like SEGs, firewalls, and spam filters to detect and mitigate increasingly novel and sophisticated email threats.
When there are tell-tale signs of a threat, these solutions can identify an incoming message as suspicious. Pointers such as emails from unknown senders, messages which contain an unusual amount of poor spelling and grammar or encourage the receiver to respond to an unexpected but supposedly urgent request.
That is, if the phishing attacks weren’t blocked by security measures before reaching the victim’s inbox. But, this is happening more and more often as phishing campaigns are becoming more advanced. Attackers are showing signs of consistently bypassing traditional protections and getting through to exploit victims.
Darktrace email threat reporting
In its End of Year Threat Report, Darktrace analyzed over 10 million phishing emails targeting customer environments between September 1 and December 31, 2023. Our findings signal that attackers are starting to take advantage of advancements in artificial intelligence (AI), including using Generative AI tools such as Large Language Models (LLMs) to create more convincing and sophisticated phishing messages – and at scale.
LLMs and Phishing
With the right AI prompts, attackers can use these LLMs to help write convincing email messages designed to target specific countries, companies or even individuals – all without the suspicious hallmarks which are traditionally associated with standard phishing attacks. The attackers don’t even need to speak the language of the individuals or groups they’re targeting. LLMs lower language barriers for attackers; using their native tongue, they can simply ask the Generative AI to write a message in the language of their choosing.
These techniques are designed to build trust and manipulate recipients into giving up sensitive information like user credentials, intellectual property or bank information or coerce them into downloading malicious payloads which can be used to launch further attacks on business infrastructure. With the appropriate research, attackers can tailor the messages to increase the chances of being successful, like making them look like a legitimate company email or request.
Social engineering phishing attacks
A year ago. Darktrace shared research which found a 135% increase in ‘novel social engineering attacks’ in the first two months of 2023, corresponding with the widespread adoption of ChatGPT. These novel phishing attacks showed a strong linguistic deviation compared to other phishing emails, which suggested to us that Generative AI was already providing an avenue for threat actors to craft sophisticated and targeted attacks at speed and scale
We’ve seen this trend continue. Our End of Year Threat Report found 38% of these emails were identified as utilizing novel social engineering techniques.
Attackers are also deploying another technique to make phishing emails look more convincing – they’re making the emails themselves longer and more sophisticated.
A potential victim might be suspicious of an ‘urgent’ email which prompts them to take action without an explanation - but if there’s additional context in the text, it adds an aura of legitimacy which is difficult to act against.
And threat actors know this; 28% of phishing emails analyzed by Darktrace over the period were identified as having “significant” amount of text – containing over 1,000 characters, which equates to over 200 words.
It’s a sign that attackers are innovating and bolstering their efforts to craft sophisticated phishing campaigns, potentially leveraging Generative AI tools to automate social engineering activity by creating longer, more convincing phishing emails.
QR code phishing
But this is far from the only innovative method which attackers are using to bypass traditional security defences. Among the 10 million plus emails analyzed during the reporting period, Darktrace/Email detected over 639,000 malicious QR codes within the messages.
Malicious QR codes placed within emails have become an increasingly common form of phishing attack, especially as QR codes have become a more common method for sharing links to information or buying links for products in recent years.
Attackers are deploying QR codes because they provide a way of directing unsuspecting victims to malicious websites or download links without needing to use a traditional phishing URL.
The advantage of implanting QR codes for attackers is that while phishing URLs are something which traditional security solutions are actively looking to identify and mitigate, malicious QR codes are more difficult for them to detect.
Applying AI to email security
Traditional security solutions which rely heavily on previously identified malicious emails and known bad senders are struggling to identify and defend against these novel and increasingly sophisticated email threats.
But by using AI that learns the unique digital environment and patterns of each business, Darktrace/Email can recognize the subtle deviations in expected email activity to determine whether any given email could represent a threat to the business. It is then able to make highly accurate decisions to mitigate and neutralize any email attack it faces helping to keep your organization safe from cyber disruption.
It’s therefore imperative that in the battle against ever-evolving, ever more sophisticated cyber threats, defenders are also embracing AI to keep businesses safe. By effectively applying AI to cyber security challenges, defenders can take a proactive approach to cyber security, staying one step ahead of malicious attackers, with real-time detection and automated response to known and unknown threats looking to disrupt the business via the inbox.
Darktrace/Email was recently awarded a 2024 AI Excellence Award for Machine Learning by Business Intelligence Group.
Join Darktrace on 9 April for a virtual event to explore the latest innovations needed to get ahead of the rapidly evolving threat landscape. Register today to hear more about our latest innovations coming to Darktrace’s offerings.
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.
Journey of a Threat: How Multi-Layered AI Works in Darktrace / EMAIL
Follow a malicious email as it moves through Darktrace / EMAIL’s multi-layered AI system, from raw data to final decision. Each layer works together to detect threats, understand intent, and take autonomous action.
How email-delivered prompt injection attacks can target enterprise AI – and why it matters
Prompt injection is a newly emerging threat, with only a handful of confirmed victims so far – targeting how AI systems use data rather than exploiting traditional software vulnerabilities. As agentic AI becomes embedded across enterprise environments, attackers may attempt to manipulate these systems through hidden instructions in everyday email content.
Darktrace Unites Human Behavior and Threat Detection Across Email, Slack, Teams, and Zoom
Introducing the adaptive era of email security: a unified platform that connects personalized coaching, collaboration tools, and user behavior into a self-improving defense system.
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.
Figure 9: Autonomous Response actions implemented by Darktrace in response to suspicious connectivity in mid-March.
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
A New Security Challenge: The Curious Case of Prompt Language Analysis
Why prompt analysis is emerging as a key AI security challenge
If securing AI has been one of the defining cybersecurity conversations of the past year, prompt analysis is quickly becoming one of its most interesting frontiers.
Security leaders are under pressure to understand how AI is being used across the business. In some organizations, that means governing employee use of chatbots. In others, it means overseeing copilots embedded into SaaS platforms, monitoring coding assistants, or assessing the growing footprint of autonomous agents. However different these use cases may appear on the surface, they share a common factor: humans and machines are usually interacting with enterprise systems through language.
How prompt language differs from traditional security telemetry
For years, defenders have become used to working with familiar forms of telemetry: email traffic, network connections, API calls, endpoint processes, authentication events. Prompt language is different. It is not simply another log source. It is an expression of intent, instruction, curiosity, urgency, and sometimes manipulation. It reflects the end-goal of a user or agent, but not always with enough surrounding context to interpret the risk correctly.
Why existing security approaches only partially explain prompt risk
A growing number of vendors are approaching the task of securing AI from the angle they know best. Perimeter vendors are extending web or browser controls into AI usage. Identity vendors are emphasizing agent permissions and access governance. Data security and DLP providers are focusing on content inspection and exfiltration risk. All of these perspectives matter, but individually can’t fully explain the problem.
The challenge with securing AI is not just that a new application category has emerged. It is that language has become a new operating layer in the enterprise.
Employees now use prompts to summarize documents, generate code, analyze spreadsheets, query internal knowledge, and trigger multi-step actions through agents. In each case, prompt language acts as the interface between human intent and machine execution. That makes prompts incredibly valuable from a security perspective as they can hint at misuse, policy violations, data exposure, or attempts to circumvent controls. However, they can also be deeply ambiguous when viewed in isolation. That ambiguity is the heart of the issue.
Prompts as behavioral signals, not just text to classify
A prompt by itself tells you what was asked. It does not necessarily tell you whether the request is expected, risky, accidental, or entirely legitimate in context. Two nearly identical prompts can carry very different meanings depending on the role and function of who issued them, what systems they can access, and what actions followed. In other words, prompts are not just text to classify. They are behavioral signals to interpret.
Example: How context changes prompt risk entirely
Consider a common enterprise scenario. An employee is pulled into a new project with an aggressive deadline. Almost overnight, their use of AI tools spikes. They begin prompting more frequently, working across unfamiliar documents, querying new data sources, and interacting with more systems than usual to accelerate delivery. Viewed narrowly, this may look suspicious. Prompt volume increases, file access patterns change, API and SaaS activity rise. From some vantage points, it may resemble insider risk or unmanaged AI usage.
But now add context. Imagine that, earlier that day, the employee received instructions from a senior leader asking them to support a time-sensitive initiative. Their communication history shows that this leader is a legitimate reporting-line superior. Their recent collaboration patterns align with the new project team. Their subsequent activity, while unusual for that individual’s baseline, is consistent with the business task they were assigned.
What initially looked like a risk event may actually be a normal response to business pressure. Without the surrounding context of communication, organizational relationships, and broader behavioral patterns, prompt activity alone could generate more noise than insight.
The reverse is also true. A prompt may appear benign on the surface while the context around it suggests elevated risk. A request that seems routine could originate from a compromised user, a newly connected external agent, a shadow AI workflow, or a user acting outside their normal role. The language itself may not contain anything obviously malicious, but the surrounding conditions may tell a very different story.
What security teams need to analyze prompts effectively
The future of prompt analysis is not just about understanding language. It is about understanding language in context.
To do that well, security teams need more than prompt inspection. They need to understand:
Who is issuing the prompt, whether human or agent
How that identity normally behaves across the enterprise
What systems, data, and workflows are connected to the interaction
Which relationships and communications explain the surrounding activity
Whether the downstream actions align with expected business behavior
When those layers are absent, prompt analysis can become another isolated control surface: useful in theory, but limited in practice. Security teams may detect unusual wording but miss the operational function behind it, overreact to benign changes in behavior, or miss subtle misuse because the prompt itself did not appear dangerous.
How organizations should think about prompt analysis going forward
Security teams have seen this pattern before. In the cloud, posture without runtime context left important gaps. In identity, access control without behavioral understanding missed misuse that looked legitimate on paper. In data security, content inspection without business context often created friction without resolving risk. AI is exposing the same lesson again: controls are strongest when they are coordinated, not isolated. As organizations work to secure AI and identify gaps across their security operations, prompt analysis will become an increasingly important source of insight, but only as part of a broader strategy.
Prompt analysis will undoubtedly become more common, as prompts are one of the clearest windows into how people and agents are using AI systems. However, what matters most is not simply collecting prompts or filtering dangerous phrases, but being able to place that language inside a wider behavioral and operational picture.
Organizations that already have a broader understanding of how work gets done across the enterprise will be better positioned to make sense of prompt language as this category matures. They will be better able to distinguish urgency from abuse, experimentation from exfiltration, and productive AI adoption from hidden risk.
Figure 1: Darktrace / SECURE AI reconstructs the full sequence of events, showing every user and agent interaction in context, with risky prompts highlighted and categorized, including PII, sensitive data, and other policy violations.
At Darktrace, this is the key lesson emerging from the market: prompt language does matter, but it does not stand alone. It is most valuable when treated as a new behavioral input that can enrich understanding across the enterprise, not as a self-contained source of truth.
Why prompts become less useful when analyzed in isolation
The curious case of prompt language analysis, then, is this: the more important prompts become, the less useful they are in a vacuum.
The real opportunity is not just to see what was asked. It is to understand why it was asked, what it meant in that moment, and what happened next.
For a deeper look at how organizations are approaching this challenge from the strengths of prompt analysis to its limitations in isolation see Prompt Security in Enterprise AI: Strengths, Weaknesses, and Common Approaches, which expands on the role prompt-level controls play within a broader, context-driven security strategy.
.avif)
![Initial connectivity observed to the rare external hostname, zoom[.]uswebob[.]us · 148.72.73[.]98, over port 443.](https://cdn.prod.website-files.com/626ff4d25aca2edf4325ff97/6a3c241a851b82605cdf62fb_Screenshot%202026-06-24%20at%2011.38.14%E2%80%AFAM.png)
