December 4, 2025

Atomic Stealer: Darktrace’s Investigation of a Growing macOS Threat

Atomic Stealer is rapidly emerging as a significant macOS threat, using infostealing and backdoor capabilities to target Apple users worldwide. Darktrace observed campaigns in 24 countries, detecting activity across multiple kill chain stages and providing recommended actions to contain attacks.

Written by

Isabel Evans

Cyber Analyst

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

Isabel Evans

Cyber Analyst

Atomic Stealer: Darktrace’s Investigation of a Growing macOS Threat

Dec 2025

The Rise of Infostealers Targeting Apple Users

In a threat landscape historically dominated by Windows-based threats, the growing prevalence of macOS information stealers targeting Apple users is becoming an increasing concern for organizations. Infostealers are a type of malware designed to steal sensitive data from target devices, often enabling attackers to extract credentials and financial data for resale or further exploitation. Recent research identified infostealers as the largest category of new macOS malware, with an alarming 101% increase in the last two quarters of 2024 [1].

What is Atomic Stealer?

Among the most notorious is Atomic macOS Stealer (or AMOS), first observed in 2023. Known for its sophisticated build, Atomic Stealer can exfiltrate a wide range of sensitive information including keychain passwords, cookies, browser data and cryptocurrency wallets.

Originally marketed on Telegram as a Malware-as-a-Service (MaaS), Atomic Stealer has become a popular malware due to its ability to target macOS. Like other MaaS offerings, it includes services like a web panel for managing victims, with reports indicating a monthly subscription cost between $1,000 and $3,000 [2]. Although Atomic Stealer’s original intent was as a standalone MaaS product, its unique capability to target macOS has led to new variants emerging at an unprecedented rate

Even more concerning, the most recent variant has now added a backdoor for persistent access [3]. This backdoor presents a significant threat, as Atomic Stealer campaigns are believed to have reached an around 120 countries. The addition of a backdoor elevates Atomic Stealer to the rare category of backdoor deployments potentially at a global scale, something only previously attributed to nation-state threat actors [4].

This level of sophistication is also evident in the wide range of distribution methods observed since its first appearance; including fake application installers, malvertising and terminal command execution via the ClickFix technique. The ClickFix technique is particularly noteworthy: once the malware is downloaded onto the device, users are presented with what appears to be a legitimate macOS installation prompt. In reality, however, the user unknowingly initiates the execution of the Atomic Stealer malware.

This blog will focus on activity observed across multiple Darktrace customer environments where Atomic Stealer was detected, along with several indicators of compromise (IoCs). These included devices that successfully connected to endpoints associated with Atomic Stealer, those that attempted but failed to establish connections, and instances suggesting potential data exfiltration activity.

Darktrace’s Coverage of Atomic Stealer

As this evolving threat began to spread across the internet in June 2025, Darktrace observed a surge in Atomic Stealer activity, impacting numerous customers in 24 different countries worldwide. Initially, most of the cases detected in 2025 affected Darktrace customers within the Europe, Middle East, and Africa (EMEA) region. However, later in the year, Darktrace began to observe a more even distribution of cases across EMEA, the Americas (AMS), and Asia Pacific (APAC). While multiple sectors were impacted by Atomic Stealer, Darktrace customers in the education sector were the most affected, particularly during September and October, coinciding with the return to school and universities after summer closures. This spike likely reflects increased device usage as students returned and reconnected potentially compromised devices to school and campus environments.

Starting from June, Darktrace detected multiple events of suspicious HTTP activity to external connections to IPs in the range 45.94.47.0/24. Investigation by Darktrace’s Threat Research team revealed several distinct patterns ; HTTP POST requests to the URI “/contact”, identical cURL User Agents and HTTP requests to “/api/tasks/[base64 string]” URIs.

Within one observed customer’s environment in July, Darktrace detected two devices making repeated initiated HTTP connections over port 80 to IPs within the same range. The first, Device A, was observed making GET requests to the IP 45.94.47[.]158 (AS60781 LeaseWeb Netherlands B.V.), targeting the URI “/api/tasks/[base64string]” using the “curl/8.7.2” user agent. This pattern suggested beaconing activity and triggered the ‘Beaconing Activity to External Rare' model alert in Darktrace / NETWORK, with Device A’s Model Event Log showing repeated connections. The IP associated with this endpoint has since been flagged by multiple open-source intelligence (OSINT) vendors as being associated with Atomic Stealer [5].

Figure 1: Darktrace’s detection of Device A showing repeated connections to the suspicious IP address over port 80, indicative of beaconing behavior.

Darktrace’s Cyber AI Analyst subsequently launched an investigation into the activity, uncovering that the GET requests resulted in a ‘503 Service Unavailable’ response, likely indicating that the server was temporarily unable to process the requests.

Figure 2: Cyber AI Analyst Incident showing the 503 Status Code, indicating that the server was temporarily unavailable.

This unusual activity prompted Darktrace’s Autonomous Response capability to recommend several blocking actions for the device in an attempt to stop the malicious activity. However, as the customer’s Autonomous Response configuration was set to Human Confirmation Mode, Darktrace was unable to automatically apply these actions. Had Autonomous Response been fully enabled, these connections would have been blocked, likely rendering the malware ineffective at reaching its malicious command-and-control (C2) infrastructure.

Figure 3: Autonomous Response’s suggested actions to block suspicious connectivity on Device A in the first customer environment.

In another customer environment in August, Darktrace detected similar IoCs, noting a device establishing a connection to the external endpoint 45.94.47[.]149 (ASN: AS57043 Hostkey B.V.). Shortly after the initial connections, the device was observed making repeated requests to the same destination IP, targeting the URI /api/tasks/[base64string] with the user agent curl/8.7.1, again suggesting beaconing activity. Further analysis of this endpoint after the fact revealed links to Atomic Stealer in OSINT reporting [6].

Figure 4: Cyber AI Analyst investigation finding a suspicious URI and user agent for the offending device within the second customer environment.

As with the customer in the first case, had Darktrace’s Autonomous Response been properly configured on the customer’s network, it would have been able to block connectivity with 45.94.47[.]149. Instead, Darktrace suggested recommended actions that the customer’s security team could manually apply to help contain the attack.

Figure 5: Autonomous Response’s suggested actions to block suspicious connectivity to IP 45.94.47[.]149 for the device within the second customer environment.

In the most recent case observed by Darktrace in October, multiple instances of Atomic Stealer activity were seen across one customer’s environment, with two devices communicating with Atomic Stealer C2 infrastructure. During this incident, one device was observed making an HTTP GET request to the IP 45.94.47[.]149 (ASN: AS60781 LeaseWeb Netherlands B.V.). These connections targeted the URI /api/tasks/[base64string, using the user agent curl/8.7.1.

Shortly afterward, the device began making repeated connections over port 80 to the same external IP, 45.94.47[.]149. This activity continued for several days until Darktrace detected the device making an HTTP POST request to a new IP, 45.94.47[.]211 (ASN: AS57043 Hostkey B.V.), this time targeting the URI /contact, again using the curl/8.7.1 user agent. Similar to the other IPs observed in beaconing activity, OSINT reporting later linked this one to information stealer C2 infrastructure [7].

Figure 6: Darktrace’s detection of suspicious beaconing connectivity with the suspicious IP 45.94.47.211.

Further investigation into this customer’s network revealed that similar activity had been occurring as far back as August, when Darktrace detected data exfiltration on a second device. Cyber AI Analyst identified this device making a single HTTP POST connection to the external IP 45.94.47[.]144, another IP with malicious links [8], using the user agent curl/8.7.1 and targeting the URI /contact.

Figure 7: Cyber AI Analyst investigation finding a successful POST request to 45.94.47[.]144 for the device within the third customer environment.

A deeper investigation into the technical details within the POST request revealed the presence of a file named “out.zip”, suggesting potential data exfiltration.

Figure 8: Advanced Search log in Darktrace / NETWORK showing “out.zip”, indicating potential data exfiltration for a device within the third customer environment.

Similarly, in another environment, Darktrace was able to collect a packet capture (PCAP) of suspected Atomic Stealer activity, which revealed potential indicators of data exfiltration. This included the presence of the “out.zip” file being exfiltrated via an HTTP POST request, along with data that appeared to contain details of an Electrum cryptocurrency wallet and possible passwords.

Read more about Darktrace’s full deep dive into a similar case where this tactic was leveraged by malware as part of an elaborate cryptocurrency scam.

Figure 9: PCAP of an HTTP POST request showing the file “out.zip” and details of Electrum Cryptocurrency wallet.

Although recent research attributes the “out.zip” file to a new variant named SHAMOS [9], it has also been linked more broadly to Atomic Stealer [10]. Indeed, this is not the first instance where Darktrace has seen the “out.zip” file in cases involving Atomic Stealer either. In a previous blog detailing a social engineering campaign that targeted cryptocurrency users with the Realst Stealer, the macOS version of Realst contained a binary that was found to be Atomic Stealer, and similar IoCs were identified, including artifacts of data exfiltration such as the “out.zip” file.

Conclusion‍

The rapid rise of Atomic Stealer and its ability to target macOS marks a significant shift in the threat landscape and should serve as a clear warning to Apple users who were traditionally perceived as more secure in a malware ecosystem historically dominated by Windows-based threats.

Atomic Stealer’s growing popularity is now challenging that perception, expanding its reach and accessibility to a broader range of victims. Even more concerning is the emergence of a variant embedded with a backdoor, which is likely to increase its appeal among a diverse range of threat actors. Darktrace’s ability to adapt and detect new tactics and IoCs in real time delivers the proactive defense organizations need to protect themselves against emerging threats before they can gain momentum.

Credit to Isabel Evans (Cyber Analyst), Dylan Hinz (Associate Principal Cyber Analyst)
Edited by Ryan Traill (Analyst Content Lead)

Appendices

References

1. https://www.scworld.com/news/infostealers-targeting-macos-jumped-by-101-in-second-half-of-2024

2. https://www.kandji.io/blog/amos-macos-stealer-analysis

3. https://www.broadcom.com/support/security-center/protection-bulletin/amos-stealer-adds-backdoor

4. https://moonlock.com/amos-backdoor-persistent-access

5. https://www.virustotal.com/gui/ip-address/45.94.47.158/detection

6. https://www.trendmicro.com/en_us/research/25/i/an-mdr-analysis-of-the-amos-stealer-campaign.html

7. https://www.virustotal.com/gui/ip-address/45.94.47.211/detection

8. https://www.virustotal.com/gui/ip-address/45.94.47.144/detection

9. https://securityaffairs.com/181441/malware/over-300-entities-hit-by-a-variant-of-atomic-macos-stealer-in-recent-campaign.html

10. https://binhex.ninja/malware-analysis-blogs/amos-stealer-atomic-stealer-malware.html

Darktrace Model Detections

Darktrace / NETWORK

Compromise / Beaconing Activity To External Rare
Compromise / HTTP Beaconing to New IP
Compromise / HTTP Beaconing to Rare Destination
Anomalous Connection / New User Agent to IP Without Hostname
Device / New User Agent
Compromise / Sustained TCP Beaconing Activity To Rare Endpoint
Compromise / Slow Beaconing Activity To External Rare
Anomalous Connection / Posting HTTP to IP Without Hostname
Compromise / Quick and Regular Windows HTTP Beaconing

Autonomous Response

Antigena / Network / Significant Anomaly::Antigena Alerts Over Time Block
Antigena / Network / Significant Anomaly::Antigena Significant Anomaly from Client Block
Antigena / Network / External Threat::Antigena Suspicious Activity Block

List of IoCs

45.94.47[.]149 – IP – Atomic C2 Endpoint
45.94.47[.]144 – IP – Atomic C2 Endpoint
45.94.47[.]158 – IP – Atomic C2 Endpoint
45.94.47[.]211 – IP – Atomic C2 Endpoint
out.zip - File Output – Possible ZIP file for Data Exfiltration

MITRE ATT&CK Mapping:

Tactic –Technique – Sub-Technique

Execution - T1204.002 - User Execution: Malicious File

Credential Access - T1555.001 - Credentials from Password Stores: Keychain

Credential Access - T1555.003 - Credentials from Web Browsers

Command & Control - T1071 - Application Layer Protocol

Exfiltration - T1041 - Exfiltration Over C2 Channel

Written by

Isabel Evans

Cyber Analyst

Inside the SOC

Written by

Isabel Evans

Cyber Analyst

Watch the NIS2 Webinar

Blog

Network

June 24, 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

Blog

June 24, 2026

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.

About the author

Nabil Zoldjalali

VP, Field CISO

Your data. Our AI.

Elevate your network security with Darktrace AI

Get a demo

Check out this article by Darktrace: Atomic Stealer: Darktrace’s Investigation of a Growing macOS Threat

Atomic Stealer: Darktrace’s Investigation of a Growing macOS Threat

The Rise of Infostealers Targeting Apple Users

What is Atomic Stealer?

Darktrace’s Coverage of Atomic Stealer

Appendices

References

Darktrace Model Detections

List of IoCs

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

A New Security Challenge: The Curious Case of Prompt Language Analysis

Enjoying the blog?

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

Network

June 24, 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:

Blog

June 24, 2026

A New Security Challenge: The Curious Case of Prompt Language Analysis

Why prompt analysis is emerging as a key AI security challenge

How prompt language differs from traditional security telemetry

Why existing security approaches only partially explain prompt risk

Prompts as behavioral signals, not just text to classify

Example: How context changes prompt risk entirely

What security teams need to analyze prompts effectively

How organizations should think about prompt analysis going forward

Why prompts become less useful when analyzed in isolation