Blog
/
/
December 7, 2021

Conti Ransomware Strategies in Modern Cybercrime

Uncover the strategies behind the Conti ransomware gang's double extortion methods and what it means for businesses facing cyber threats.
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
Justin Fier
SVP, Offensive Security
Default blog image
07
Dec 2021

In a previous blog, we outlined how the Ryuk ransomware strain developed by Russian hacking group ‘Wizard Spider’ has fallen into the hands of small-time cyber criminals.

Wizard Spider – who allegedly operate with support from the Russian government and remain under investigation by the FBI and Interpol – adopted Ryuk ransomware’s successor ‘Conti’ in 2020. Conti affects all Windows operating systems and has been involved in more than 400 incidents. Wizard Spider were soon rebranded in cyber press as the ‘Conti Ransomware Gang’, though the group does not necessarily see itself as a ‘gang’. It prefers to present itself as a business.

The ransomware bubble

Ransomware has become a multibillion-dollar industry – and the Conti Ransomware Gang reportedly made up 15% of it in 2020. With this scale of income, groups like Conti find themselves adopting some crude imitations of legitimate business practice. This corporate mimicry dictates that their victims be called ‘customers’, their extortion attempts ‘negotiations’ and their criminal peers ‘affiliates’. They even publish ‘press releases’ via a dedicated Dark Web site.

The gang’s Ransomware-as-a-Service ‘business model’ consists of employing affiliates, training them in Conti ransomware’s deployment and management, and then taking 30% of the profits themselves. With exact profits known only to the malware writers and not the affiliates, however, the percentage Conti takes is often much higher than the 30% they claim.

There may not be checks and regulations in place to address fraud in the cyber underworld, but one business complication which Conti have not been able to escape is that of the disgruntled employee.

Unhappy with the malpractice of their superiors, an underpaid affiliate leaked the Conti Ransomware Gang’s training materials and the IP addresses for their Cobalt Strike C2 servers in August 2021, declaring, “they recruit suckers and divide the money among themselves”.

Meanwhile, the US Government has also been taking action to try to disrupt the profit margins of groups like the Conti Ransomware Gang, going as far as to impose sanctions on cryptocurrency exchanges seen as facilitating ransomware transactions. However, leaks and legislation have proved far from fatal for Conti.

The reality is that these actions have not lost the Conti Ransomware Gang any of its so-called “customers”, and where there are customers there is profit. Any individual or organization entrusting their cyber security to conventional, rules-based measures is in their target market.

Darktrace’s AI recently detected a Conti attack conducted along the lines of one of the methods outlined in the August leak. The target organization – a US transportation company – was trialing Darktrace but, without Darktrace’s Autonomous Response set in active mode, the attack was allowed to go ahead. In examining how it progressed, however, it should become clear not only how threatening double extortion ransomware attacks like this one can be, but also how effectively they can be stopped by Darktrace at each stage of the attack.

Figure 1: Timeline of the attack

Conti Ransomware Gang diversifies the ransomware playbook

A single uninstalled Microsoft patch had left the target organization with dangerous ProxyShell vulnerabilities. Conti exploited these vulnerabilities, quickly gaining the rights to remotely execute Exchange PowerShell commands on the company’s server and steadily broadened its presence within the digital environment. This is a relatively new approach for the Conti Ransomware Gang, who previously relied upon phishing attacks and firewall exploits. By diversifying its approach, it stays ahead of patches and intelligence.

Two weeks after the initial breach, C2 connections were made to an unusual endpoint located in Finland using an SSL client which appeared innocuous but was 100% rare for the organization. Had Autonomous Response been set in active mode, Darktrace would have shut the connections down at this very early stage.

The IP address of this suspicious endpoint has since been identified as a Conti IoC (Indicator of Compromise), allowing it to be incorporated into rules-based security solutions. This would have done little good for the company in question, however, which was breached weeks before this intelligence was made available.

As Conti continued to conduct internal reconnaissance and move laterally through the company’s digital environment, Darktrace detected further unusual activity. The suspicious Finnish endpoint then employed new ‘Living off the Land’ techniques, installing the usually legitimate tools AnyDesk and Cobalt Strike onto various parts of the environment.

A series of SSL connections were made to AnyDesk endpoints and external hosts, one of which lasted 95 hours, indicating an active remote session conducted by one of Conti’s affiliates. At this stage, Darktrace had 10 distinct reasons to suspect an imminent attack.

Conti News: Closing the deal with double extortion ransomware

Double extortion has become the Conti Ransomware Gang’s new favourite sales tactic. If you refuse to pay its ransom, Conti will not only take your most important files from you, but also exfiltrate and publish them using its dedicated ‘Conti News’ website, or sell them directly to your competitors.

Having expanded their reach across the transport company’s network, the Conti affiliate began rapidly exfiltrating large quantities of company data to Conti’s preferred cloud storage site, MEGA. Over four days, more than 3TB of data was uploaded, and then encrypted.

To avoid detection by a human security team, encryption was launched at close to midnight – Conti’s ‘business’ does not respect business hours. When the company’s security team returned to work the next day, they were met with a ransom note.

This attack was able to progress because Darktrace was only being trialed at this stage and was therefore allowed to detect threats but not to take action against them. With Autonomous Response employed in active mode, this ransomware attack would have ended in the very early stages, when Darktrace detected its first suspicious connections.

Nonetheless, the Cyber AI Analyst was able to investigate and connect the dots of the attack automatically, making the organization’s remediation efforts drastically quicker and easier than they would have been without even this partial Darktrace deployment.

Figure 2: Cyber AI Analyst generated this incident report following the initiation of data exfiltration

How the Conti Ransomware Gang evades cyber intelligence

Security systems that rely on human intelligence to detect threats fit Conti’s ideal customer profile perfectly. By adapting and diversifying their approach, moving from Ryuk to Conti, and from spear phishing and firewall exploits to this new ProxyShell approach, Conti stay ahead of regulations and hold on to their vulnerable customer base.

Even if the Conti Ransomware Gang is brought down by leaks or legislation, other groups will rise to fill the gap in the market, eager for their own cut of the illicit gains. If these groups are to be truly stopped, they must be made unprofitable.

The US government has tried to do this by imposing fines upon ransom payers, but companies still often consider the losses involved in not recovering their data too great. As I have argued previously, ‘to pay or not to pay,’ is not the question we should be asking.

If you’re deciding whether to pay or not to pay, you’re already too far down the line. Darktrace stops groups like Conti at the first encounter. As this case has shown, Darktrace’s Self-Learning AI is able to identify threats weeks before human analysts and threat intelligence can do the same, and neutralize them at every stage of an attack with Autonomous Response.

Thanks to Darktrace analyst Sam Lister for his insights on the above threat find.

Darktrace model detections:

  • Device / Long Agent Connection to New Endpoint
  • Device / ICMP Address Scan
  • Anomalous Connection / SMB Enumeration
  • Anomalous Server Activity / Outgoing from Server
  • Compromise / Beacon to Young Endpoint
  • Anomalous Server Activity / Rare External from Server
  • Compromise / Fast Beaconing to DGA
  • Compromise / SSL or HTTP Beacon
  • Compromise / Sustained SSL or HTTP Increase
  • Compromise / Beacon for 4 Days
  • Anomalous Connection / Multiple HTTP POSTs to Rare Hostname
  • Unusual Activity / Enhanced Unusual External Data Transfer
  • Anomalous Connection / Data Sent to Rare Domain
  • Anomalous Connection / Uncommon 1 GiB Outbound
  • Compliance / SMB Drive Write
  • Anomalous File / Internal / Additional Extension Appended to SMB File
  • Anomalous Connection / Suspicious Read Write Ratio
  • Anomalous Connection / Suspicious Read Write Ratio and Unusual SMB
  • Anomalous Connection / Sustained MIME Type Conversion
  • Unusual Activity / Anomalous SMB Move & Write
  • Unusual Activity / Unusual Internal Data Volume as Client or Server
  • Device / Suspicious File Writes to Multiple Hidden SMB Shares
  • Compromise / Ransomware / Suspicious SMB Activity
  • Anomalous File / Internal / Unusual SMB Script Write
  • Anomalous File / Internal / Masqueraded Executable SMB Write
  • Device / SMB Lateral Movement
  • Device / Multiple Lateral Movement Model Breaches

‍

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
Justin Fier
SVP, Offensive Security

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