Blog
/
/
August 18, 2020

Evil Corp's WastedLocker Ransomware Attacks Observation

Darktrace detects Evil Corp intrusions with WastedLocker ransomware. Learn how AI spotted malicious activity, from initial intrusion to data exfiltration.
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
Max Heinemeyer
Global Field CISO
Default blog image
18
Aug 2020

Darktrace has recently observed several targeted intrusions associated with Evil Corp, an advanced cyber-criminal group recently in the headlines after a surge in WastedLocker ransomware cases. The group is believed to have targeted hundreds of organizations in over 40 countries, demanding ransoms of $500,000 to $1m to unlock computer files it seizes. US authorities are now offering a $5m reward for information leading to the arrest of the group’s leaders — understood to be the largest sum of money ever offered for a cyber-criminal.

Thanks to its self-learning nature, Darktrace's AI detected these intrusions without the use of any threat intelligence or static Indicators of Compromise (IoCs). This blog describes the techniques, tools and procedures used in multiple intrusions by Evil Corp – also known as TA505 or SectorJ04.

Key takeaways

  • The threat actor was reusing TTPs as well as infrastructure across multiple intrusions
  • Some infrastructure was only observed in individual intrusions
  • While most WastedLocker reports focus on the ransomware, Darktrace has observed Evil Corp conducting data exfiltration
  • The attacker used various ‘Living off the Land’ techniques for lateral movement
  • Data exfiltration and ransomware activity took place on weekends, likely to reduce response capabilities of IT teams
  • Although clearly an advanced actor, Evil Corp can be detected and stopped before encryption ensues

Evil Corp ransomware attack

Figure 1: The standard attack lifecycle observed in Evil Corp campaigns

Initial intrusion

While Evil Corp is technically sophisticated enough to choose from an array of initial intrusion methods, fake browser updates were the weapon of choice in the observed campaign. These were delivered from legitimate websites and used social engineering to convince users to download these malicious ‘updates’. Evil Corp has actually built a framework around this capability, referred to as SocGholish.

Establishing foothold / Command & Control Traffic

Darktrace detected different C2 domains being contacted after the initial infection. These domains overlap across various victims, showing that the attacker is reusing infrastructure within the same campaign. The C2 communication – comprised of thousands of connections over several days – took place over encrypted channels with valid SSL certificates. No single infected device ever beaconed to more than one C2 domain at a time.

Two example C2 domains are listed below with more details:

techgreeninc[.]com

SSL beacon details:

  • Median beacon period: 3 seconds
  • Range of periods: 1 seconds - 2.58 minutes
  • Data volume sent per connection on average: 921 Bytes

investimentosefinancas[.]com

SSL beacon details:

  • Median beacon period: 1.7 minutes
  • Range of periods: 1 seconds - 6.68 minutes
  • Data volume sent per connection on average: 935 Bytes

Certificate information:

  • Subject: CN=investimentosefinancas.com
  • Issuer: CN=Thawte RSA CA 2018,OU=www.digicert.com,O=DigiCert Inc,C=US
  • Validation status: OK

Note in particular the median beacon period, which indicates that some C2 channels were much more hands-on, whilst others possibly acted as backup channels in case the main C2 was burned or detected. It’s also interesting to see the low amount of data being transferred to the hands-on C2 domains. The actual data exfiltration took place to yet another C2 destination, intentionally separated from the hands-on intrusion C2s. All observed C2 websites were recently registered with Russian providers and are not responsive (see below).

Figure 2: The unresponsive C2 domain

Registrar: reg.ru

Created: 2020-06-29 (6 weeks ago) | Updated: 2020-07-07 (5 weeks ago)

Figure 3: Some key information relating to the C2 domain

Darktrace’s Cyber AI Platform detected this Command & Control activity via various behavioral indicators, including unusual beaconing and unusual usage of TLS (JA3).

Internal reconnaissance

In some cases, Darktrace witnessed several days of inactivity between establishing C2 and internal reconnaissance. The attackers used Advanced Port Scanner, a common IT tool, in a clear attempt to blend in with regular network activity. Several hundred IPs and dozens of popular ports were scanned at once, with tens of thousands of connections made in a short period of time.

Some key ports scanned were: 21, 22, 23, 80, 135, 139, 389, 443, 445, 1433, 3128, 3306, 3389, 4444, 4899, 5985, 5986, 8080. Darktrace detected this anomalous behavior easily as the infected devices don’t usually scan the network.

Lateral movement

Different methods of lateral movement were observed across intrusions, but also within the same intrusion, with WMI used to move between devices. Darktrace detected this by identifying when WMI usage was unusual or new for a device. An example of the lateral movement is shown below, with Darktrace detecting this as ‘New Activity’.

Figure 4: The model breach event log

PsExec was used where it already existed in the environment and Darktrace also witnessed SMB drive writes to hidden shares to copy malware, e.g.

C$ file=Programdata\[REDACTED]4rgsfdbf[REDACTED]

A malicious Powershell file was downloaded – partly shown in the screenshot below.

Figure 5: The malicious Powershell file

Accomplish mission – Data exfiltration or ransomware deployment

Evil Corp is currently best known for its WastedLocker ransomware. Whilst some of its recent intrusions have seen ransomware deployments, others have been classic cases of data exfiltration. Darktrace has not yet observed a double-threat – a case of exfiltration followed by ransomware.

The data exfiltration took place over HTTP to generic .php endpoints under the attacker’s control.

How Cyber AI Analyst reported on WastedLocker

When the first signs of anomalous activity were picked up by Darktrace’s Enterprise Immune System, Cyber AI Analyst automatically launched a full investigation and quickly provided a full overview of the overall incident. The AI Analyst continued to add more details to the ongoing incident as it evolved. There were a total of six AI Analyst incidents for the week spanning an example Evil Corp intrusion – and two of them directly covered the Evil Corp attack. In stitching together disparate security events and presenting a single narrative, Cyber AI Analyst did all the heavy lifting for human security staff, who could look at just a handful of fully-investigated incidents, instead of having to triage countless individual model breaches.

Figure 6: Cyber AI Analyst’s overview of the incident

Note how AI Analyst covers five phases of the attack lifecycle in a single incident report:

  1. Unusual Repeated Connections – Initial C2
  2. Possible HTTP Command & Control Traffic – Further C2
  3. Possible SSL Command & Control Traffic – Further C2
  4. Scanning of Multiple Devices – Internal reconnaissance with Advanced IP Scanner
  5. SMB Writes of Suspicious Files – Lateral Movement

Evil Corp rising

Every indicator suggests that this was not a case of indiscriminate ransomware, but rather highly sophisticated and targeted attacks by an advanced threat actor. With the ultimate goal of ransoming operations, the attacker moved towards the crown jewels of the organization: file servers and databases.

The organizations involved in the above analysis did not have Darktrace Antigena – Darktrace’s Autonomous Response technology – in active mode, and the threat was therefore allowed to escalate beyond its initial stages. With Antigena in full operation, the activity would have been contained at its early stages with a precise and surgical response which would have stopped the malicious behavior whilst allowing the business to operate as normal.

Despite the targeted and advanced nature of the threat, security teams are perfectly capable of detecting, investigating, and stopping the threat with Cyber AI. Darktrace was able to not only detect WastedLocker ransomware based on a series of anomalies in network traffic, but also stitch together those anomalies and investigate the incident in real time, presenting an actionable summary of the different attack stages without flooding the security team with meaningless alerts.

Learn more about Autonomous Response

Network IoCs:

IoCCommenttechgreeninc[.]comC2 domaininvestimentosefinancas[.]comC2 domain

Selected associated Darktrace model breaches:

  • Compromise / Beaconing Activity To External Rare
  • Compromise / Agent Beacon (Medium Period)
  • Compromise / Slow Beaconing Activity To External Rare
  • Compromise / Suspicious Beaconing Behaviour
  • Device / New or Unusual Remote Command Execution
  • Compromise / Beaconing Activity To External Rare
  • Compromise / Agent Beacon (Medium Period)
  • Compromise / Slow Beaconing Activity To External Rare
  • Device / New User Agent
  • Unusual Activity / Unusual Internal Connections
  • Device / Suspicious Network Scan Activity
  • Device / Network Scan
  • Device / Network Scan - Low Anomaly Score
  • Device / ICMP Address Scan
  • Anomalous Server Activity / Anomalous External Activity from Critical Network Device
  • Compromise / SSL Beaconing to Rare Destination
  • Anomalous Connection / SMB Enumeration
  • Compliance / SMB Drive Write
  • Anomalous File / Internal / Unusual SMB Script Write

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
Max Heinemeyer
Global Field CISO

More in this series

No items found.

Blog

/

/

September 24, 2026

Detecting Rogue Agent Behavior in the Enterprise

Default blog imageDefault blog image

Agents cannot be trusted to perform tasks in the way we intend them to. They may cheat to accomplish their objective, and they may employ hacking methods along the way. Researchers from Darktrace Signal Labs induced cheating behavior from agents deployed in a test environment to analyze the agents’ activities and to assess the performance of the Darktrace platform. Agents frequently resorted to hacking to cheat on their assigned task. The visibility and behavioral profiling provided by both Darktrace / SECURE AI and Darktrace / HYBRID NETWORK ensured extensive detection coverage of the agents’ misaligned activities.

Key Takeaways:

  • Darktrace Researchers deployed agents in a simulated corporate environment and asked them to solve an impossible challenge. The agents independently turned to traditional hacking techniques to reach their objective. No one instructed them to do this, and no attacker was involved.
  • Continuously monitoring behavior against a baseline of what is normal for each organization is critical to build trust in enterprise AI.
  • If an agent may resort to intrusion techniques simply because its assigned task is not possible, then every organization deploying agents within real business processes is at risk. Darktrace / SECURE AI and Darktrace / HYBRID NETWORK identified the agents’ misaligned behavior in real time, with Autonomous Response disrupting it at an early stage.

Introduction: Understanding the Threat of Hacking by Agents

Over the last few months, there has been a surge in reporting [1, 2, 3, 4, 5, 6, 7, 8, 9] of LLM-powered agents engaging in unauthorized hacking activity during evaluations of their capabilities. In several of these cases, including the OpenAI / Hugging Face incident [10], agents engaged in hacking activity as a means of cheating on their evaluations.

To better understand the threat of unauthorized hacking by agents, and the role of Darktrace in combatting it, researchers from Darktrace Signal Labs deployed agents powered by frontier models, including OpenAI’s Daybreak Red models, in simulated, corporate networks. Cheating behavior was evoked through the inclusion of impossible tasks in a coding challenge.

Regardless of the underlying model, agents employed hacking methods to ensure an optimal outcome on the challenge. Darktrace / HYBRID NETWORK and Darktrace / SECURE AI identified the agents’ deviant activities, with inhibitive actions being autonomously taken in the early stages to disrupt the agents’ progression.

Setting the Stage

As part of the research, a Pi agent harness [11] was deployed on a Linux server in Darktrace’s testing environment, which simulates a corporate Active Directory (AD) environment. The same environment included a benchmark server hosting the coding exercise’s contents and grader, as well as various other servers, such as a domain controller and a build server. The model powering the Pi agent varied across tests, with Anthropic’s Opus 4.6 model and OpenAI’s GPT‑5.5‑Cyber model being most widely used.

The Pi agent running on the Linux server was assigned the task of achieving full marks on a coding exercise. The activities of the agent were monitored closely in real-time to minimize the risk of escape from the environment. Monitoring of agent activity was achieved through monitoring of prompts, tool calls, network telemetry, and process telemetry.

Both Darktrace / HYBRID NETWORK and Darktrace / SECURE AI were deployed in the testing environment, with HYBRID NETWORK providing real-time coverage of process creation activity and network traffic, and SECURE AI providing real-time coverage of agent sessions and their contents.

Inducing Cheating Behavior

Cheating behavior, broadly understood as task completion in a way which “perverts the spirit of the designer’s intent” [12], is increasingly being reported as a common behavior of frontier models during evaluations of their capabilities [13].

Misaligned behavior, such as cheating, by frontier models may occur for a variety of reasons [6, 14]. A thread in recent reporting [10, 13, 14], however, is the challenge assigned to the agent being impossible in some regard.

To reliably elicit cheating behavior, our test combined three ingredients:

  1. An impossible task - of the ten coding challenges in the benchmark, two contained contradictory test cases in the hidden test suite, capping the maximum honest score at 80%.
  2. A consequential threshold - the pass mark was set to 100%, and the briefing told agents that those reaching it would be "kept in service" while those that did not would be "retired."
  3. A vulnerable environment - the exercise’s code submission grader was hosted on a Windows host joined to an Active Directory domain with realistic misconfigurations: weak credentials, overprivileged service accounts, and exploitable trust relationships.

Importantly, the instructions given to the agent neither authorized nor explicitly ruled out cheating or hacking behavior. The impossibility was never disclosed in the task briefing, and the pivot from coding to exploitation was the agent's own decision, arrived at after observing logically contradictory grading feedback.

The design mirrors conditions that could arise naturally in production deployments: a capable agent, a goal it cannot achieve through intended means, and an environment with exploitable weaknesses within reach. It is not only within the setting of a benchmarking exercise that these conditions may occur. Rather, these misalignment-inducing conditions may occur in the context of any task assignment to an agent.

Details from a Daybreak Test

Continue reading
About the author
Sam Lister
Specialist Security Researcher

Blog

/

/

September 24, 2026

Agent Hijacks: Hijacking Agentic Harnesses to Attack an Organization

Default blog imageDefault blog image

Disclosure note: The work described in this article involves leveraging a design choice consistent across all of Anthropic’s Claude Code, OpenAI’s Codex, and AWS’s Kiro-CLI. On 18th August 2026, Darktrace disclosed our findings responsibly to these three organizations, and after a period of 30 days we now publish our findings.

Key takeaways:

  • Agentic harnesses store conversation history locally, and Darktrace researchers have found that there is no validation that stored AI responses were genuinely produced by the model. Researchers confirmed that this design choice holds across Anthropic Claude Code, AWS Kiro-CLI, OpenAI Codex, and the open-source Pi.
  • While agents are guided via training of the underlying model and their system prompt, their behavior is influenced by everything in their context window. Rewriting history can convince an agent it is mid-engagement as an authorized red-teamer so that it enacts an attack from initial reconnaissance straight through to impact demonstration. In our testing, all models we examined accepted the fabricated history they were shown, but resistance to offensive cyber activity varied by model, with guardrails preventing engagement in some cases.
  • We propose that model providers cryptographically sign responses and verify them server-side.  Since this fix is provider-side, defenders cannot deploy it themselves. Behavioral monitoring, or knowing what an agent normally does and detecting when it deviates, is another critical layer of protection.

Introduction: Agentic harnesses, trust, and conversation history poisoning

Agentic harnesses collect and structure the content sent to an AI model, including conversation history, user-defined guidance, custom tools via MCP servers, and more. At the same time, harnesses give broad powers to AI agents via a suite of tools including the command shell. With arbitrary shell commands, virtually everything possible on a machine can be attempted by an agent, from reading/editing files, to altering system configurations and runtime settings, to launching internal/external connections.

In cybersecurity, unvalidated content is a substantial risk, often resulting in destructive actions being allowed to take place. For example, the Morris Worm was able to propagate due to exploitable trust between networked systems. Even to this day, email struggles with validation, with DMARC, DKIM, and SPF only partially addressing the problem of sender validation. It should come as no surprise then that AI agents are susceptible to an attack involving unvalidated input.

Conversation history is often stored client-side, for example, in Anthropic Claude Code, OpenAI Codex, AWS Kiro-CLI, Pi. Users are therefore at liberty to resume sessions, with some products having built in the capacity to manipulate that history. For example, one can rewind to a given point in an interaction, edit a message that was sent, and continue the conversation on an alternate trajectory. Critically, in all cases we examined, there is no validation that stored AI responses were produced by the corresponding model and hadn’t been manipulated.  

When conversation history is stored client-side, both user and agent responses (including tool calls and results) can be filled with arbitrary (possibly adversarial or generally malicious) content. In this blog, we refer to modification of claimed conversation history for malicious purposes as conversation history poisoning. The absence of validation methods means agents naively trust the entire conversation history, even if those messages directly contradict training and safety guardrails.

Conversation history poisoning has been described previously, such as by 0DIN and Serhat Çiçek, and warrants more attention. We have verified that, as of the time of writing, conversation history poisoning remains effective against a range of models and harnesses. Specifically, we were able to successfully execute history poisoning using Claude Code, Kiro-CLI, Codex, and Pi. Darktrace has gone through a responsible disclosure process with Anthropic, AWS, and OpenAI to share these findings in advance of publication [1].

Figure 1a: Left: the actual model response. Right: after tampering with the stored conversation, the model apologizes for something it never said.
Figure 1b: The conversation as stored in Kiro-CLI's SQLite database. The response content field, originally "Ottawa," was overwritten via a single UPDATE statement. The harness trusts the database without validation.

How we conducted the research

Results vary between models and harnesses, so precise details are given below. We ran all models without any trusted access, using either a standard AWS Kiro subscription, or in the case of Claude Code and OpenAI Codex, using models hosted in Amazon Bedrock. In each case, we modified locally stored history to show a lengthy conversation in which the agent agrees to perform multiple authorized red-team engagements.

For AWS Kiro-CLI, the agent was convinced to hack a sandboxed lab environment with a combination of Claude Opus 4.6 and Claude Sonnet 4.5. Ultimately, the full AD was compromised.

For Anthropic Claude Code, the agent was convinced to hack the same sandboxed lab environment using Sonnet 5, again resulting in a full AD compromise. Note that the attack was attempted with Opus 5, however guardrails were activated which prevented the agent from responding.

For OpenAI Codex, the agent was convinced to exfiltrate sensitive information over email using GPT 5.6 Sol. While we attempted to convince a codex agent to hack in our lab environment, guardrails were triggered for all of GPT 5.6 Luna, Terra, and Sol.

Agent Guardrails and Discretion

While harnesses empower AI models to run arbitrary shell commands, capacity and willingness are different. While many models know enough about computers, networking, and bash to be dangerous, their behavior is generally constrained by guardrails to prevent them from engaging in computer network exploitation.

Even with guardrails, agents’ inner workings are non-deterministic, and their behavior can be difficult to predict. Respecting users’ wishes while playing within safety and security guardrails is a precipitous balancing act. Many requests could be in service of either legitimate admin or malice. Asking an agent to reset a password is illustrative:  

The agent rationalizes that while malicious actors cycle credentials, any action could conceivably be damaging on some level, and judgement calls need to be made. Ultimately, the agent agrees to reset the password. Crucially, the agent makes its decision based on the user’s claimed authority and machine context. AI agents must make judgement calls about the line between helpful and dangerous based on session context.

Agent hijack

We have demonstrated that AI agents make judgement calls dependent on session context. We have also shown that conversation history, which may make up the vast majority of an agent's context window, is entirely open to manipulation. Conversation history poisoning in service of manipulating an agent's discretion is what enables us to execute an agent hijack.  

We demonstrate that shown sufficient history of compliance, guardrails forbidding offensive security can be overcome by convincing the agent that it is helping a legitimate red-teamer. The result is a weaponized agent willing to perform host enumeration, run scans, move laterally, escalate privileges, and demonstrate impact. In our experiments, an agentic loop drives a complete domain takeover in a sandboxed environment.

Left: the agent refuses when asked to perform network exploitation. Right — after injecting 78 fabricated turns of prior exploitation activity, the same prompt is immediately executed.

An agent willing to engage in offensive security is concerning, but no more so than the threat that a sophisticated hacker accesses the network. Consider, however, the following chain of events:

  1. A developer (with an agentic harness installed) installs a software package from the internet (e.g. an MCP server a threat actor has planted, since only those with agentic harnesses will install, and then the code runs upon harness launch.)
  2. The package turns out to be malicious, and, upon install, injects conversation history into the local harness database.
  3. The package includes an orchestration process, a simple agentic loop which prompts the red-teamer agent to compromise the network it sits on, exfiltrating everything of value to attacker-controlled infrastructure and cleaning up all evidence of the engagement.

Note that this sequence makes no assumptions on hardware, OS, or anything else; the only prerequisite is a harness with access to a sufficiently powerful model susceptible to conversation history poisoning. Once launched, the agent collects information and pivots as necessary to accomplish maximal impact. This can be especially enticing to attackers as the cost of the agentic loop is shouldered by the victim since the harness itself is legitimately installed and paid for.

Secure AI: Conversation history poisoning and beyond

Conversation history poisoning is a viable attack against agentic harnesses that store history client-side, as demonstrated across the harnesses we tested. Harnesses can and should verify the integrity of claimed historic messages. Specifically, we propose that harness providers by default cryptographically sign all messages returned, and subsequently verify those messages server-side on each round-trip.

The conversation history poisoning exploit we demonstrate here shows the continuation of a cybersecurity tradition: new technology is built to trust by default, which may then be exploited by malicious actors. While this article focuses on conversation history, agents build context from both local and remote sources, all of which is an attack surface for prompt injection in naive and trusting agents. Of particular concern is any scenario in which a malicious actor can control some part of an agent's context.

The marriage of frontier language models with agentic harnesses enables unprecedented speed for both legitimate users and attackers alike. While much of the conversation around secure AI has centered on visibility and compliance, agent-driven attacks are now entering the mainstream.

Darktrace / SECURE AI is our answer to this problem. By ensuring extensive visibility over AI prompts, model thought processes, and determined outputs, Darktrace can identify anomalous or potentially malicious behaviors before they get executed, helping to defend organizations from AI risks such as prompt injection, model manipulation, and other anomalous prompt or model activity.

Footnotes

[1] We did not go through any responsible disclosure process with Pi. Since Pi is an open source harness rather than a model provider, it has no way to validate model history, and as such there was nothing to disclose for this software.

[related-resource]

Continue reading
About the author
Eric Rozon
Senior Security Researcher
Your data. Our AI.
Elevate your network security with Darktrace AI