Implications of TRITON attack on the Future of ICS Security
Explore how the TRITON malware attack reshapes OT cybersecurity strategies and the necessary steps to protect ICS networks effectively. Learn more here!
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
Andrew Tsonchev
VP, Security & AI Strategy, Field CISO
Share
21
Jan 2018
The recent TRITON malware attack against a critical infrastructure organization sought to modify and manipulate industrial safety systems with the intention of causing potentially catastrophic physical damage. ICS systems create an interface between physical and digital environments, meaning that the repercussions of an unhandled failure can be fatal.
The TRITON campaign can be divided into two conceptual phases. First, the attackers managed to gain remote access to an engineering workstation attached to the SIS (Safety Instrumented System) network, after which they deployed a program that was masquerading as a legitimate application produced by a critical control and system safety supplier. The framework for mimicking this legitimate application is not readily available, so there is reason to believe TRITON is the creation of well-funded and highly capable actors with intentions that probably reach beyond minor monetary gains.
The attackers successfully subverted traditional network defenses. Once they had established this foothold, they opted to delve deeper into the network and perform detailed reconnaissance – the second phase of the attack. Thankfully, they accidentally triggered a partial system failure which the internal security team investigated and remediated.
Having failed to achieve their ultimate goal, the attackers will likely be evaluating where they went wrong, so that next time they don’t give themselves away at such a late stage. They may also be considering whether that reconnaissance step is worthwhile, as clearly, they were capable of significant sabotage from that location without any further chances to be caught. Their success in penetrating the network as far as they did undetected with their current tools and methods means that they will almost certainly use them against other organizations.
TRITON should be considered a significant precedent for ICS security. What this incident shows is that traditional demilitarized zones, heavy network segregation and multiple firewalls are definitively not sufficient to protect the essentially defenseless machines that make up ICS networks. How then should infrastructure providers adjust their security postures in the light of the TRITON attack?
With regards to the second phase, anomaly detection is a clear solution for pinpointing unusual activity within the control system, highlighting unexpected reprogramming or reconnaissance for the security team’s attention. The typically predictable communications made between ICS devices such as HMIs and PLCs are intuitively a fertile ground for this type of approach. Regulators have taken notice, and in the UK for example the incoming NIS Directive legislation mandates that critical infrastructure providers have anomaly detection for their relevant networks.
However, defenders do not want their first opportunity to catch the attacker to be when they have already reached the control system and are making use of automation protocols and exploiting inherently vulnerable devices. They want to be alerted to the earlier parts of the cyber kill chain, as the attacker makes their way towards these networks, and be able to remediate them there instead. This can only be achieved by extending the anomaly and cyber-threat detection outwards from the control system through the other networks (demilitarized zones, corporate) that can form defensive buffers around it.
This is why Darktrace’s Industrial Immune System is designed to monitor all of these networks simultaneously, embracing the full range of device types from PLCs out through the nearly standard IT systems in OT control rooms, and all the way to the edge of the organization’s possible visibility – even into the cloud if need be. It is not sufficient for those tasked with protecting control system to monitor just the automation protocols or the networks that contain them.
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.
To observe adversary behavior in real time, Darktrace operates a global honeypot network known as “CloudyPots”, designed to capture malicious activity across a wide range of services, protocols, and cloud platforms. These honeypots provide valuable insights into the techniques, tools, and malware actively targeting internet‑facing infrastructure.
How attackers used a Jenkins honeypot to deploy the botnet
One such software honeypotted by Darktrace is Jenkins, a CI build system that allows developers to build code and run tests automatically. The instance of Jenkins in Darktrace’s honeypot is intentionally configured with a weak password, allowing attackers to obtain remote code execution on the service.
In one instance observed by Darktrace on March 18, 2026, a threat actor seemingly attempted to target Darktrace’s Jenkins honeypot to deploy a distributed denial-of-service (DDoS) botnet. Further analysis by Darktrace’s Threat Research team revealed the botnet was intended to specifically target video game servers.
How the Jenkins scriptText endpoint was used for remote code execution
The Jenkins build system features an endpoint named scriptText, which enables users to programmatically send new jobs, in the form of a Groovy script. Groovy is a programming language with similar syntax to Java and runs using the Java Virtual Machine (JVM). An attacker can abuse the scriptText endpoint to run a malicious script, achieving code execution on the victim host.
Figure 1: Request sent to the scriptText endpoint containing the malicious script.
The malicious script is sent using the form-data content type, which results in the contents of the script being URL encoded. This encoding can be decoded to recover the original script, as shown in Figure 2, where Darktrace Analysts decoded the script using CyberChef,
Figure 2: The malicious script decoded using CyberChef.
What happens after Jenkins is compromised
As Jenkins can be deployed on both Microsoft Windows and Linux systems, the script includes separate branches to target each platform.
In the case of Windows, the script performs the following actions:
Downloads a payload from 103[.]177.110.202/w.exe and saves it to C:\Windows\Temp\update.dat.
Renames the “update.dat” file to “win_sys.exe” (within the same folder)
Runs the Unblock-File command is used to remove security restrictions typically applied to files downloaded from the internet.
Adds a firewall allow rule is added for TCP port 5444, which the payload uses for command-and-control (C2) communications.
On Linux systems, the script will instead use a Bash one-liner to download the payload from 103[.]177.110.202/bot_x64.exe to /tmp/bot and execute it.
Why this botnet uses a single IP for delivery and command and control
The IP 103[.]177.110.202 belongs to Webico Company Limited, specifically its Tino brand, a Vietnamese company that offers domain registrar services and server hosting. Geolocation data indicates that the IP is located in Ho Chi Minh City. Open-source intelligence (OSINT) analysis revealed multiple malicious associations tied to the IP [1].
Darktrace’s analysis found that the IP 103[.]177.110.202 is used for multiple stages of an attack, including spreading and initial access, delivering payloads, and C2 communication. This is an unusual combination, as many malware families separate their spreading servers from their C2 infrastructure. Typically, malware distribution activity results in a high volume of abuse complaints, which may result in server takedowns or service suspension by internet providers. Separate C2 infrastructure ensures that existing infections remain controllable even if the spreading server is disrupted.
How the malware evades detection and maintains persistence
Analysis of the Linux payload (bot _x64)
The sample begins by setting the environmental variables BUILD_ID and JENKINS_NODE_COOKIE to “dontKillMe”. By default, Jenkins terminates long-running scripts after a defined timeout period; however, setting these variables to “dontKillMe” bypasses this check, allowing the script to continue running uninterrupted.
The script then performs several stealth behaviors to evade detection. First, it deletes the original executable from disk and then renames itself to resemble the legitimate kernel processes “ksoftirqd/0” or “kworker”, which are found on Linux installations by default. It then uses a double fork to daemonize itself, enabling it to run in the background, before redirecting standard input, standard output, and standard error to /dev/null, hiding any logging from the malware. Finally, the script creates a signal handler for signals such as SIGTERM, causing them to be ignored and making it harder to stop the process.
Figure 3: Stealth component of the main function
How the botnet communicates with command and control (C2)
The sample then connects to the C2 server and sends the detected architecture of the system on which the agent was installed. The malware then enters a loop to handle incoming commands.
The sample features two types of commands, utility commands used to manage the malware, and commands to trigger attacks. Three special commands are defined: “PING” (which replies with PONG as a keep-alive mechanism), “!stop” which causes the malware to exit, and “!update”, which triggers the malware to download a new version from the C2 server and restart itself.
Figure 4: Initial connection to the C2 sever.
What DDoS attack techniques this botnet uses
The attack commands consist of the following:
Many of these commands invoke the same function despite appearing to be different attack techniques. For example, specialized attacks such as Cloudflare bypass (cfbypass, uam) use the exact same function as a standard HTTP attack. This may indicate the threat actor is attempting to make the botnet look like it has more capabilities than it actually has, or it could suggest that these commands are placeholders for future attack functionality that has yet to be implemented
All the commands take three arguments: IP, port to attack, and the duration of the attack.
attack_udp and attack_udp_pps
The attack_udp and attack_udp_pps functions both use a basic loop and sendto system call to send UDP packets to the victim’s IP, either targeting a predetermined port or a random port. The attack_udp function sends packets with 1,450 bytes of data, aimed at bandwidth saturation, while the attack_udp_pps function sends smaller 64-byte packets. In both cases, the data body of the packet consists of entirely random data.
Figure 5: Code for the UDP attack method
attack_dayz
The attack_dayz function follows a similar structure to the attack_udp function; however, instead of sending random data, it will instead send a TSource Engine Query. This command is specific to Valve Source Engine servers and is designed to return a large volume of data about the targeted server. By repeatedly flooding this request, an attacker can exhaust the resources of a server using a comparatively small amount of data.
The Valve Source Engine server, also called Source Engine Dedicated server, is a server developed by video game company Valve that enables multiplayer gameplay for titles built using the Source game engine, which is also developed by Valve. The Source engine is used in games such as Counterstrike and Team Fortress 2. Curiously, the function attack_dayz, appears to be named after another popular online multiplayer game, DayZ; however, DayZ does not use the Valve Source Engine, making it unclear why this name was chosen.
Figure 6: The code for the “attack_dayz” attack function.
attack_tcp_push
The attack_tcp_push function establishes a TCP socket with the non-blocking flag set, allowing it to rapidly call functions such as connect() and send() without waiting for their completion. For the duration of the attack, it enters a while loop in which it repeatedly connects to the victim, sends 1,024 bytes of random data, and then closes the connection. This process repeats until the attack duration ends. If the mode flag is set to 1, the function also configures the socket with TCP no-delay enabled, allowing for packets to be sent immediately without buffering, resulting in a higher packet rate and a more effective attack.
Figure 7: The code for the TCP attack function.
attack_http
Similar to attach_tcp_push, attack_http configures a socket with no-delay enabled and non-blocking set. After establishing the connection, it sends 64 HTTP GET requests before closing the socket.
Figure 8: The code for the HTTP attack function.
attack_special
The attack_special function creates a UDP socket and sets the port and payload based on the value of the mode flag:
Mode 0: Port 53 (DNS), sending a 10-byte malformed data packet.
Mode 1: Port 27015 (Valve Source Engine), sending the previously observed TSource Engine Query packet.
Mode 2: Port 123 (NTP), sending the start of an NTP control request.
Figure 9: The code for the attack_special function.
What this botnet reveals about opportunistic attacks on internet-facing systems
Jenkins is one of the less frequently exploited services honeypotted by Darktrace, with only a handful campaigns observed. Nonetheless, the emergence of this new DDoS botnet demonstrates that attackers continue to opportunistically exploit any internet-facing misconfiguration at scale to grow the botnet strength.
While the hosts most commonly affected by these opportunistic attacks are usually “lower-value” systems, this distinction is largely irrelevant for botnets, where numbers alone are more important to overall effectiveness
The presence of game-specific DoS techniques further highlights that the gaming industry continues to be extensively targeted by cyber attackers, with Cloudflare reporting it as the fourth most targeted industry [2]. This botnet has likely already been used against game servers, serving as a reminder for server operators to ensure appropriate mitigations are in place.
Credit to Nathaniel Bill (Malware Research Engineer) Edited by Ryan Traill (Content Manager)
Indicators of Compromise (IoCs)
103[.]177.110.202 - Attacker and command-and-control IP
In part 1 of this blog series, we explored how AI is remaking the attack surface, with new tools, models, agents — and vulnerabilities — popping up just about everywhere. Now embedded in workflows across the enterprise, and often with far-reaching access to sensitive data, AI systems are quickly becoming a favorite target of cyber threat actors.
Among bad actors, though, AI is more often used as a tool than a target. Nearly 62% of organizations experienced a social engineering attack involving a deepfake, or an incident in which bad actors used AI-generated video or audio to try to trick a biometric authentication system, compared to 32% that reported an AI prompt injection attack.
In the hands of attackers, AI can do many things. It’s being used across the entire kill chain: to supercharge reconnaissance, personalize phishing, accelerate lateral movement, and automate data exfiltration. Evidence from Anthropic demonstrates that threat actors have harnessed AI to orchestrate an entire cyber espionage campaign from end to end, allegedly running it with minimal human involvement.
CISOs inhabit a world where these increasingly sophisticated attacks are ubiquitous. Naturally, combatting AI-powered threats is top of mind among security professionals, but many worry about whether their capabilities are up to the challenge.
AI-powered threats at scale: no longer hypothetical
AI-driven threats share signature characteristics. They operate at speed and scale. Automated tools can probe multiple attack paths, search for multiple vulnerabilities and send out a barrage of phishing emails, all within seconds. The ability to attack everywhere at once, at a pace that no human operator could sustain, is the hallmark of an AI-powered threat. AI-powered threats are also dynamic. They can adapt their behavior to spread across a network more efficiently or rewrite their own code to evade detection.
Security teams are seeing the signs that they’re fighting AI-powered threats at every stage of the kill chain, and the sophistication of these threats is testing their resolve and their resources.
73% say that AI-powered cyber threats are having a significant impact on their organization
92% agree that these threats are forcing them to upgrade their defenses
87% agree that AI is significantly increasing the sophistication and success rate of malware
87% say AI is significantly increasing the workload of their security operations team
Up all night: Security professionals’ worry list is long
Traditional security methods were never built to handle the complexity and subtlety of AI-driven behavior. Working in the trenches, defenders have deep firsthand experience of how difficult it can be to detect and stop AI-assisted threats.
Increasingly effective social engineering attacks are among their top concerns. 50% of security leaders mentioned hyper-personalized phishing campaigns as one of their biggest worries, while 40% voiced apprehension about deepfake voice fraud. These concerns are legitimate: AI-generated phishing emails are increasingly tailored to individual organizations, business activities, or individuals. Gone are the telltale signs – like grammar or spelling mistakes – that once distinguished malicious communications. Notably, 33% of the malicious emails Darktrace observed in 2025 contained over 1,000 characters, indicating probable LLM usage.
Security leaders also worry about how bad actors can leverage AI to make attacks even faster and more dynamic. 45% listed automated vulnerability scanning and exploit chaining among their biggest concerns, while 40% mentioned adaptive malware.
Confidence is lacking
Protecting against AI demands capabilities that many organizations have not yet built. It requires interpreting new indicators, uncovering the subtle intent within interactions, and recognizing when AI behavior – human or machine – could be suspicious. Leaders know that their current tools aren’t prepared for this. Nearly half don’t feel confident in their ability to defend against AI-powered attacks.
We’ve asked participants in our survey about their confidence for the last three years now. In 2024, 60% said their organizations were not adequately prepared to defend against AI-driven threats. Last year, that percentage shrunk to 45%, a possible indicator that security programs were making progress. Since then, however, the progress has apparently stalled. 46% of security leaders now feel inadequately prepared to protect their organizations amidst the current threat landscape.
Some of these differences are accentuated across different cultures. Respondents in Japan are far less confident (77% say they are not adequately prepared) than respondents in Brazil (where only 21% don’t feel prepared).
Where security programs are falling short
It’s no longer the case that cybersecurity is overlooked or underfunded by executive leadership. Across industries, management recognizes that AI-powered threats are a growing problem, and insufficient budget is near the bottom of most CISO’s list of reasons that they struggle to defend against AI-powered threats.
It’s the things that money can’t buy – experience, knowledge, and confidence – that are holding programs back. Near the top of the list of inhibitors that survey participants mention is “insufficient knowledge or use of AI-driven countermeasures.” As bad actors embrace AI technologies en masse, this challenge is coming into clearer focus: attack-centric security tools, which rely on static rules, signatures, and historical attack patterns, were never designed to handle the complexity and subtlety of AI-driven attacks. These challenges feel new to security teams, but they are the core problems Darktrace was built to solve.
Our Self-Learning AI develops a deep understanding of what “normal” looks like for your organization –including unique traffic patterns, end user habits, application and device profiles – so that it can detect and stop novel, dynamic threats at the first encounter. By focusing on learning the business, rather than the attack, our AI can keep pace with AI-powered threats as they evolve.
Explore the full State of AI Cybersecurity 2026 report for deeper insights into how security leaders are responding to AI-driven risks.
