Blog
/
Network
/
October 20, 2025

Salty Much: Darktrace’s view on a recent Salt Typhoon intrusion

Salt Typhoon, a China-linked cyber espionage group, has been observed targeting global infrastructure using stealthy techniques such as DLL sideloading and zero-day exploits. Darktrace recently identified early-stage intrusion activity consistent with Salt Typhoon’s tactics, reinforcing the importance of anomaly-based detection over traditional signature-based methods when defending against persistent, state-sponsored threat.
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
Nathaniel Jones
VP, Security & AI Strategy, Field CISO
Written by
Sam Lister
Specialist Security Researcher
salt typhoonDefault blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog image
20
Oct 2025

What is Salt Typhoon?

Salt Typhoon represents one of the most persistent and sophisticated cyber threats targeting global critical infrastructure today. Believed to be linked to state-sponsored actors from the People’s Republic of China (PRC), this advanced persistent threat (APT) group has executed a series of high-impact campaigns against telecommunications providers, energy networks, and government systems—most notably across the United States.

Active since at least 2019, the group — also tracked as Earth Estries, GhostEmperor, and UNC2286 — has demonstrated advanced capabilities in exploiting edge devices, maintaining deep persistence, and exfiltrating sensitive data across more than 80 countries. While much of the public reporting has focused on U.S. targets, Salt Typhoon’s operations have extended into Europe, the Middle East, and Africa (EMEA) where it has targeted telecoms, government entities, and technology firms. Its use of custom malware and exploitation of high-impact vulnerabilities (e.g., Ivanti, Fortinet, Cisco) underscores the strategic nature of its campaigns, which blend intelligence collection with geopolitical influence [1].

Leveraging zero-day exploits, obfuscation techniques, and lateral movement strategies, Salt Typhoon has demonstrated an alarming ability to evade detection and maintain long-term access to sensitive environments. The group’s operations have exposed lawful intercept systems, compromised metadata for millions of users, and disrupted essential services, prompting coordinated responses from intelligence agencies and private-sector partners worldwide. As organizations reassess their threat models, Salt Typhoon serves as a stark reminder of the evolving nature of nation-state cyber operations and the urgent need for proactive defense strategies.

Darktrace’s coverage

In this case, Darktrace observed activity in a European telecommunications organization consistent with Salt Typhoon’s known tactics, techniques and procedures (TTPs), including dynamic-link library (DLL) sideloading and abuse of legitimate software for stealth and execution.

Initial access

The intrusion began with exploitation of CVE-2025-5777, a vulnerability affecting Citrix NetScaler Gateway appliances, in the first week of July 2025. From there, the actor pivoted to Citrix Virtual Delivery Agent (VDA) hosts in the client’s Machine Creation Services (MCS) subnet. Initial access activities in the intrusion originated from an endpoint potentially associated with the SoftEther VPN service, suggesting infrastructure obfuscation from the outset.

Tooling

Darktrace subsequently observed the threat actor delivering a backdoor assessed with high confidence to be SNAPPYBEE (also known as Deed RAT) [2][3] to multiple Citrix VDA hosts. The backdoor was delivered to these internal endpoints as a DLL alongside legitimate executable files for antivirus software such as Norton Antivirus, Bkav Antivirus, and IObit Malware Fighter. This pattern of activity indicates that the attacker relied on DLL side-loading via legitimate antivirus software to execute their payloads. Salt Typhoon and similar groups have a history of employing this technique [4][5], enabling them to execute payloads under the guise of trusted software and bypassing traditional security controls.

Command-and-Control (C2)

The backdoor delivered by the threat actor leveraged LightNode VPS endpoints for C2, communicating over both HTTP and an unidentified TCP-based protocol. This dual-channel setup is consistent with Salt Typhoon’s known use of non-standard and layered protocols to evade detection. The HTTP communications displayed by the backdoor included POST requests with an Internet Explorer User-Agent header and Target URI patterns such as “/17ABE7F017ABE7F0”. One of the C2 hosts contacted by compromised endpoints was aar.gandhibludtric[.]com (38.54.63[.]75), a domain recently linked to Salt Typhoon [6].

Detection timeline

Darktrace produced high confidence detections in response to the early stages of the intrusion, with both the initial tooling and C2 activities being strongly covered by both investigations by Darktrace Cyber AI AnalystTM investigations and Darktrace models. Despite the sophistication of the threat actor, the intrusion activity identified and remediated before escalating beyond these early stages of the attack, with Darktrace’s timely high-confidence detections likely playing a key role in neutralizing the threat.

Cyber AI Analyst observations

Darktrace’s Cyber AI Analyst autonomously investigated the model alerts generated by Darktrace during the early stages of the intrusion. Through its investigations, Cyber AI Analyst discovered the initial tooling and C2 events and pieced them together into unified incidents representing the attacker’s progression.

Cyber AI Analyst weaved together separate events from the intrusion into broader incidents summarizing the attacker’s progression.
Figure 1: Cyber AI Analyst weaved together separate events from the intrusion into broader incidents summarizing the attacker’s progression.

Conclusion

Based on overlaps in TTPs, staging patterns, infrastructure, and malware, Darktrace assesses with moderate confidence that the observed activity was consistent with Salt Typhoon/Earth Estries (ALA GhostEmperor/UNC2286). Salt Typhoon continues to challenge defenders with its stealth, persistence, and abuse of legitimate tools. As attackers increasingly blend into normal operations, detecting behavioral anomalies becomes essential for identifying subtle deviations and correlating disparate signals. The evolving nature of Salt Typhoon’s tradecraft, and its ability to repurpose trusted software and infrastructure, ensures it will remain difficult to detect using conventional methods alone. This intrusion highlights the importance of proactive defense, where anomaly-based detections, not just signature matching, play a critical role in surfacing early-stage activity.

Credit to Nathaniel Jones (VP, Security & AI Strategy, FCISO), Sam Lister (Specialist Security Researcher), Emma Foulger (Global Threat Research Operations Lead), Adam Potter (Senior Cyber Analyst)

Edited by Ryan Traill (Analyst Content Lead)

Appendices

Indicators of Compromise (IoCs)

IoC-Type-Description + Confidence

89.31.121[.]101 – IP Address – Possible C2 server

hxxp://89.31.121[.]101:443/WINMM.dll - URI – Likely SNAPPYBEE download

b5367820cd32640a2d5e4c3a3c1ceedbbb715be2 - SHA1 – Likely SNAPPYBEE download

hxxp://89.31.121[.]101:443/NortonLog.txt - URI - Likely DLL side-loading activity

hxxp://89.31.121[.]101:443/123.txt - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/123.tar - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/pdc.exe - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443//Dialog.dat - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/fltLib.dll - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/DisplayDialog.exe - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/DgApi.dll - URI - Likely DLL side-loading activity

hxxp://89.31.121[.]101:443/dbindex.dat - URI - Likely DLL side-loading activity

hxxp://89.31.121[.]101:443/1.txt - URI - Possible DLL side-loading activity

hxxp://89.31.121[.]101:443/imfsbDll.dll – Likely DLL side-loading activity

hxxp://89.31.121[.]101:443/imfsbSvc.exe - URI – Likely DLL side-loading activity

aar.gandhibludtric[.]com – Hostname – Likely C2 server

38.54.63[.]75 – IP – Likely C2 server

156.244.28[.]153 – IP – Possible C2 server

hxxp://156.244.28[.]153/17ABE7F017ABE7F0 - URI – Possible C2 activity

MITRE TTPs

Technique | Description

T1190 | Exploit Public-Facing Application - Citrix NetScaler Gateway compromise

T1105 | Ingress Tool Transfer – Delivery of backdoor to internal hosts

T1665 | Hide Infrastructure – Use of SoftEther VPN for C2

T1574.001 | Hijack Execution Flow: DLL – Execution of backdoor through DLL side-loading

T1095 | Non-Application Layer Protocol – Unidentified application-layer protocol for C2 traffic

T1071.001| Web Protocols – HTTP-based C2 traffic

T1571| Non-Standard Port – Port 443 for unencrypted HTTP traffic

Darktrace Model Alerts during intrusion

Anomalous File::Internal::Script from Rare Internal Location

Anomalous File::EXE from Rare External Location

Anomalous File::Multiple EXE from Rare External Locations

Anomalous Connection::Possible Callback URL

Antigena::Network::External Threat::Antigena Suspicious File Block

Antigena::Network::Significant Anomaly::Antigena Significant Server Anomaly Block

Antigena::Network::Significant Anomaly::Antigena Controlled and Model Alert

Antigena::Network::Significant Anomaly::Antigena Alerts Over Time Block

Antigena::Network::External Threat::Antigena File then New Outbound Block  

References

[1] https://www.cisa.gov/news-events/cybersecurity-advisories/aa25-239a

[2] https://www.trendmicro.com/en_gb/research/24/k/earth-estries.html

[3] https://www.trendmicro.com/content/dam/trendmicro/global/en/research/24/k/earth-estries/IOC_list-EarthEstries.txt

[4] https://www.trendmicro.com/en_gb/research/24/k/breaking-down-earth-estries-persistent-ttps-in-prolonged-cyber-o.html

[5] https://lab52.io/blog/deedrat-backdoor-enhanced-by-chinese-apts-with-advanced-capabilities/

[6] https://www.silentpush.com/blog/salt-typhoon-2025/

The content provided in this blog is published by Darktrace for general informational purposes only and reflects our understanding of cybersecurity topics, trends, incidents, and developments at the time of publication. While we strive to ensure accuracy and relevance, the information is provided “as is” without any representations or warranties, express or implied. Darktrace makes no guarantees regarding the completeness, accuracy, reliability, or timeliness of any information presented and expressly disclaims all warranties.

Nothing in this blog constitutes legal, technical, or professional advice, and readers should consult qualified professionals before acting on any information contained herein. Any references to third-party organizations, technologies, threat actors, or incidents are for informational purposes only and do not imply affiliation, endorsement, or recommendation.

Darktrace, its affiliates, employees, or agents shall not be held liable for any loss, damage, or harm arising from the use of or reliance on the information in this blog.

The cybersecurity landscape evolves rapidly, and blog content may become outdated or superseded. We reserve the right to update, modify, or remove any content.

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
Nathaniel Jones
VP, Security & AI Strategy, Field CISO
Written by
Sam Lister
Specialist Security Researcher

More in this series

No items found.

Blog

/

Network

/

May 22, 2026

Darktrace named a Leader in the 2026 Gartner® Magic Quadrant™ for Network Detection and Response (NDR) For the Second Consecutive Year

garnter ndr magic quadrantDefault blog imageDefault blog image

Continued recognition in NDR  

Darktrace has been recognized as a Leader in the 2026 Gartner® Magic Quadrant™ for Network Detection and Response (NDR), marking the second consecutive year in the Leaders quadrant.

We believe this consistency reflects sustained ability to execute, adapt, and deliver outcomes as the market evolves.

While we are immensely proud to be recognized by industry analysts as a Leader in NDR, that's just part of the story. Darktrace was also Named the Only 2025 Gartner® Peer Insights™ Customers’ Choice for Network Detection and Response based on direct customer feedback and real-world experience.

We believe the combination of these two signals is important. One reflects how the market is evaluated. The other reflects how technology performs in practice.

Why Darktrace continues to be recognized as a leader

We believe our position as a Leader for the second consecutive year reflects a combination of our sustained ability to execute in NDR, continued AI innovation, and proven delivery of security outcomes for customers and partners worldwide.

We also feel that our leadership in the NDR market is a testament to our unique and multi-layered AI approach, for which we were recognized as No.7 on Fast Company’s Most Innovative AI Companies of 2026 list, plus one of the hottest AI cybersecurity companies in CRN's AI 100.

Adapting to complex, real-world environments

Organizations are no longer protecting a single network perimeter. They are securing a mix of users, devices, applications, and data that move across hybrid environments.

Darktrace has focused on maintaining visibility and detection across these conditions, allowing security teams to understand activity as it scales.

Supporting organizations globally, not just technically

Security outcomes are shaped as much by deployment and support as they are by detection capability.

Darktrace continues to invest in regional presence across 29 countries around the world, helping organizations operationalize NDR in ways that align with local requirements, internal processes, and team structures.

Continuing to push AI beyond detection

AI in cybersecurity is often positioned as a way to improve detection accuracy. But the more important shift is how AI can influence decision-making and response.

Darktrace continues to develop models that learn from both live environments and historical incident data, combining real-time behavioral analysis with insights derived from prior attack patterns.

Using technologies such as the Incident Graph and DIGEST (Darktrace Incident Graph Evaluation for Security Threats), activity is not analyzed in isolation. Instead, relationships between users, devices, connections, and events are mapped over time, allowing the system to reconstruct how an incident is unfolding and how similar incidents have progressed in the past.

By evaluating these patterns, Darktrace can assess the likelihood that an incident will escalate, prioritizing the activity that poses the greatest risk and surfacing the most relevant context for investigation.

This shifts security operations from simply identifying anomalies to understanding their trajectory, helping teams anticipate potential impact and respond earlier with greater precision.

Why NDR is shifting from reactive detection to proactive, AI-driven security

Traditional approaches to NDR have been built around reactively identifying threats once they become clearly visible. That model is increasingly difficult to rely on.

Attackers are no longer operating in ways that stand out. They use valid credentials, trusted tools, and low-and-slow techniques that blend into everyday activity. By the time something looks obviously malicious, the impact is often already underway.

This is the core limitation of reactive detection. It depends on recognizing something that already looks like a threat.

As a result, many of the most consequential incidents today fall into a gap.

Insider activity, compromised credentials, and novel attacks rarely trigger traditional alerts because they do not follow known patterns. On the surface, they often appear legitimate, making them difficult to distinguish from normal behavior without deeper context.

This is why we believe this Gartner recognition reflects a broader shift in NDR toward autonomous, proactive and pre‑emptive security operations.

By understanding normal behavior within an environment, it is possible to identify subtle deviations rather than waiting for confirmation of threats as they are taking place.

Darktrace’s Self-Learning AI is designed for behavioral understanding. By continuously learning each organization’s normal patterns, it can detect deviations in real time, enabling a proactive and pre-emptive model of NDR where security teams can respond to early signs of risk as they emerge, reducing the window in which attacks can develop.

In multiple cases, this behavioral approach has led to early threat detection where Darktrace identified completely unknown threats, including pre-CVE zero-day activity. By detecting subtle behavioral changes before vulnerabilities were publicly disclosed or widely understood, organizations can mitigate threats before they do damage.

This shift is subtle but important. Modern NDR solutions must shift from a system that explains what happened to one that helps prevent threats from developing in the first place, and Darktrace is proud to be at the forefront of this shift - helping organizations build and maintain a state of proactive network resilience.

Continuing to innovate at the forefront of NDR

In our view, recognition as a Leader reflects where the market is today. Continuing to innovate defines what comes next.

As businesses evolve, new technologies like AI tools and agents introduce new security risks and challenges; security teams need more than simple detection. They need a complete understanding of risk as it develops, the ability to investigate it in context, and to contain threats at machine speed.  

Darktrace / NETWORK is built to deliver across that full spectrum. Its Self-Learning AI continuously adapts to each organization’s environment, identifying subtle behavioral changes that signal emerging threats. Integrated investigation and autonomous response reduce the time between detection and action, allowing teams to move with greater speed and confidence.

This combination enables organizations to detect and contain known, unknown, and insider threats as they develop, while also strengthening resilience over time.

As a two-time Leader in the Gartner® Magic Quadrant™ for NDR and the only 2025 Gartner® Peer Insights™ Customers’ Choice, we feel Darktrace continues to evolve its platform to meet the demands of modern environments, delivering a more complete and adaptive approach to network security.

[related-resource]

Disclaimer: The 2026 Gartner® Magic Quadrant™ for Network Detection and Response (NDR) ,The 2026 Gartner® Magic Quadrant™ for Network Detection and Response (NDR), Thomas Lintemuth, Charanpal Bhogal, Nahim Fazal, 18 May 2026.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner’s research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

GARTNER is a registered trademark and service mark of Gartner, Inc. and/or its affiliates in the U.S. and internationally and is used herein with permission. All rights reserved. Magic Quadrant is a registered trademark of Gartner, Inc. and/or its affiliates and is used herein with permission. All rights reserved.

Continue reading
About the author
Mikey Anderson
Product Marketing Manager, Network Detection & Response

Blog

/

/

May 21, 2026

Prompt Security in Enterprise AI: Strengths, Weaknesses, and Common Approaches

prompt securityDefault blog imageDefault blog image

How enterprise AI Agents are changing the risk landscape  

Generative AI Agents are changing the way work gets done inside enterprises, and subsequently how security risks may emerge. Organizations have quickly realized that providing these agents with wider access to tooling, internal information, and granting permissions for the agent to perform autonomous actions can greatly increase the efficiency of employee workflows.

Early deployments of Generative AI systems led many organizations to scope individual components as self-contained applications: a chat interface, a model, and a prompt, with guardrails placed at the boundary. Research from Gartner has shown that while the volume and scope of Agentic AI deployments in enterprise environments is rapidly accelerating, many of the mechanisms required to manage risk, trust, and cost are still maturing.

The issue now resides on whether an agent can be influenced, misdirected, or manipulated in ways that leads to unsafe behavior across a broader system.

Why prompt security matters in enterprise AI

Prompt security matters in enterprise AI because prompts are the primary way users and systems interact with Agentic AI models, making them one of the earliest and most visible indicators of how these systems are being used and where risk may emerge.

For security teams, prompt monitoring is a logical starting point for understanding enterprise AI usage, providing insight into what types of questions are being asked and tasks are being given to AI Agents, how these systems are being guided, and whether interactions align with expected behavior. Complete prompt security takes this one step further, filtering out or blocking sensitive or dangerous content to prevent risks like prompt injection and data leakage.

However, visibility only at the prompt layer can create a false sense of security. Prompts show what was asked, but not always why it was asked, or what downstream actions were triggered by the agent across connected systems, data sources, or applications.

What prompt security reveals  

The primary function of prompt security is to minimize risks associated with generative and agentic AI use, but monitoring and analysis of prompts can also grant insight into use cases for particular agents and model. With comprehensive prompt security, security teams should be able to answer the following questions for each prompt:

  • What task was the user attempting to complete?
  • What data was included in the request, and was any of the data high-risk or confidential?
  • Was the interaction high-risk, potentially malicious, or in violation of company policy?
  • Was the prompt anomalous (in comparison to previous prompts sent to the agent / model)?

Improving visibility at this layer is a necessary first step, allowing organizations to establish a baseline for how AI systems are being used and where potential risks may exist.  

Prompt security alone does not provide a complete view of risk. Further data is needed to understand how the prompt is interpreted, how context is applied, what autonomous actions the agent takes (if any), or what downstream systems are affected. Understanding the outcome of a query is just as important for complete prompt security as understanding the input prompt itself – for example, a perfectly normal, low-risk prompt may inadvertently result in an agent taking a high-risk action.

Comprehensive AI security systems like Darktrace / SECURE AI can monitor and analyze both the prompt submitted to a Generative AI system, as well as the responses and chain-of-thought of the system, providing greater insight into the behavior of the system. Darktrace / SECURE AI builds on the core Darktrace methodology, learning the expected behaviors of your organization and identifying deviations from the expected pattern of life.

How organizations address prompt security today

As prompt-level visibility has become a focus, a range of approaches have emerged to make this activity more observable and controllable. Various monitoring and logging tools aim to capture prompt inputs to be analyzed after the fact.  

Input validation and filtering systems attempt to intervene earlier, inspecting prompts before they reach the model. These controls look for known jailbreak patterns, language indicative of adversarial attacks, or ambiguous instructions which could push the system off course.

Importantly, for a prompt security solution to be accurate and effective, prompts must be continually observed and governed, rather than treated as a point-in-time snapshot.  

Where prompt security breaks down in real environments

In more complex environments, especially those involving multiple agents or extensive tool use, AI security becomes harder to define and control.

Agent-to-Agent communications can be harder to monitor and trace as these happen without direct user interaction. Communication between agents can create routes for potential context leakage between agents, unintentional privilege escalation, or even data leakage from a higher privileged agent to a lower privileged one.

Risk is shaped not just by what is asked, but by the conditions in which that prompt operates and the actions an agent takes. Controls at the orchestration layer are starting to reflect this reality. Techniques such as context isolation, scoped memory, and role-based boundaries aim to limit how far a prompt’s influence can extend.  

Furthermore, Shadow AI usage can be difficult to monitor. AI systems that are deployed outside of formal governance structures and Generative AI systems hosted on unknown endpoints can fly under the radar and can go unseen by monitoring tools, leaving a critical opening where adversarial prompts may go undetected. Darktrace / SECURE AI features comprehensive detection of Shadow AI usage, helping organizations identify potential risk areas.

How prompt security fits in a broader AI risk model

Prompt security is an important starting point, but it is not a complete security strategy. As AI systems become more integrated into enterprise environments, the risks extend to what resources the system can access, how it interprets context, and what actions it is allowed to take across connected tools and workflows.

This creates a gap between visibility and control. Prompt security alone allows security teams to observe prompt activity but falls short of creating a clear understanding of how that activity translates into real-world impact across the organization.

Closing that gap requires a broader approach, one that connects signals across human and AI agent identities, SaaS, cloud, and endpoint environments. It means understanding not just how an AI system is being used, but how that usage interacts with the rest of the digital estate.

Prompt security, in that sense, is less of a standalone solution and more of an entry point into a larger problem: securing AI across the enterprise as a whole.

Explore how Darktrace / SECURE AI brings prompt security to enterprises

Darktrace brings more than a decade of AI expertise, built on an enterprise‑wide platform designed to operate in and understand the behaviors of the complex, ambiguous environments where today’s AI now lives. With Darktrace / SECURE AI, enterprises can safely adopt, manage, monitor, and build AI within their business.  

Learn about Darktrace / SECURE AI here.

Sign up today to stay informed about innovations across securing AI.

[related-resource]

Continue reading
About the author
Jamie Bali
Technical Author (AI) Developer
Your data. Our AI.
Elevate your network security with Darktrace AI