Blog
/
Network
/
January 4, 2023

BlackMatter's Smash-and-Grab Ransom Attack Incident Analysis

Stay informed on cybersecurity trends! Read about a BlackMatters ransom attack incident and Darktrace's analysis on how RESPOND could have stopped the attack.
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
The Darktrace Analyst Team
Default blog image
04
Jan 2023

Only a few years ago, popular reporting announced that the days of smash-and-grab attacks were over and that a new breed of hackers were taking over with subtler, ‘low-and-slow’ tactics [1]. Although these have undoubtedly appeared, smash-and-grab have quickly become overlooked – perhaps with worrying consequences. Last year, Google saw repeated phishing campaigns using cookie theft malware and most recently, reports of hacktivists using similar techniques have been identified during the 2022 Ukraine Conflict [2 & 3]. Where did their inspiration come from? For larger APT groups such as BlackMatter, which first appeared in the summer of 2021, smash-and-grabs never went out of fashion.

This blog dissects a BlackMatter ransomware attack that hit an organization trialing Darktrace back in 2021. The case reveals what can happen when a security team does not react to high-priority alerts. 

When entire ransomware attacks can be carried out over the course of just 48 hours, there is a high risk to relying on security teams to react to detection notifications and prevent damage before the threat escalates. Although there has been hesitancy in its uptake [4], this blog also demonstrates the need for automated response solutions like Darktrace RESPOND.

The Name Game: Untangling BlackMatter, REvil, and DarkSide

Despite being a short-lived criminal organization on the surface [5], a number of parallels have now been drawn between the TTPs (Tactics, Techniques and Procedures) of the newer BlackMatter group and those of the retired REvil and DarkSide organizations [6]. 

Prior to their retirement, DarkSide and REvil were perhaps the biggest names in cyber-crime, responsible for two of last year’s most devastating ransomware attacks. Less than two weeks after the Colonial Pipeline attack, DarkSide announced it was shutting down its operation [7]. Meanwhile the FBI shutdown REvil in January 2022 after its devastating Fourth of July Kaseya attacks and a failed return in September [8]. It is now suspected that members from one or both went on to form BlackMatter.

This rebranding strategy parallels the smash-and-grab attacks these groups now increasingly employ: they make their money, and a lot of noise, and when they’re found out, they disappear before organizations or governments can pull together their threat intelligence and organize an effective response. When they return days, weeks or months later, they do so having implemented enough small changes to render themselves and their attacks unrecognizable. That is how DarkSide can become BlackMatter, and how its attacks can slip through security systems trained on previously encountered threats. 

Attack Details

In September 2021 Darktrace was monitoring a US marketing agency which became the victim of a double extortion ransomware attack that bore hallmarks of a BlackMatter operation. This began when a single domain-authenticated device joined the company’s network. This was likely a pre-infected company device being reconnected after some time offline. 

Only 15 minutes after joining, the device began SMB and ICMP scanning activities towards over 1000 different internal IPs. There was also a large spike of requests for Epmapper, which suggested an intent for RPC-based lateral movement. Although one credential was particularly prominent, multiple were used including labelled admin credentials. Given it’s unexpected nature, this recon quickly triggered a chain of DETECT/Network model breaches which ensured that Darktrace’s SOC were alerted via the Proactive Threat Notification service. Whilst SOC analysts began to triage the activity, the organization failed to act on any of the alerts they received, leaving the detected threat to take root within their digital environment. 

Shortly after, a series of C2 beaconing occurred towards an endpoint associated with Cobalt Strike [9]. This was accompanied by a range of anomalous WMI bind requests to svcctl, SecAddr and further RPC connections. These allowed the initial compromised device to quickly infect 11 other devices. With continued scanning over the next day, valuable data was soon identified. Across several transfers, 230GB of internal data was then exfiltrated from four file servers via SSH port 22. This data was then made unusable to the organization through encryption occurring via SMB Writes and Moves/Renames with the randomly generated extension ‘.qHefKSmfd’. Finally a ransom note titled ‘qHefKSmfd.README.txt’ was dropped.

This ransom note was appended with the BlackMatter ASCII logo:

Figure 1- The ASCII logo which accompanied BlackMatter’s ransom note

Although Darktrace DETECT and Cyber AI Analyst continued to provide live alerting, the actor successfully accomplished their mission.  

There are numerous reasons that an organization may fail to organize a response to a threat, (including resource shortages, out of hours attacks, and groups that simply move too fast). Without Darktrace’s RESPOND capabilities enabled, the threat actors could proceed this attack without obstacles. 

Figure 2- Cyber AI Analyst breaks down the stages of the attack [Note: this screenshot is from V5 of DETECT/Network] 

How would the attack have unfolded with RESPOND?

Armed with Darktrace’s evolving knowledge of ‘self’ for the customer’s unique digital environment, RESPOND would have activated within seconds of the first network scan, which was recognized as highly anomalous. The standard action taken here would usually involve enforcing the standard ‘pattern of life’ for the compromised device over a set time period in order to halt the anomaly while allowing the business to continue operating as normal.

RESPOND constantly re-evaluates threats as attacks unfold. Had the first stage still been successful, it would have continued to take targeted action at each corresponding stage of this attack. RESPOND models would have alerted to block the external connections to C2 servers over port 443, the outbound exfil attempts and crucially the SMB write activity over port 445 related to encryption.

As DETECT and RESPOND feed into one another, Darktrace would have continued to assess its actions as BlackMatter pivoted tactics. These actions buy back critical time for security teams that may not be in operation over the weekend, and stun the attacker into place without applying overly aggressive responses that create more problems than they solve.

Ultimately although this incident did not resolve autonomously, in response to the ransom event, Darktrace offered to enable RESPOND and set it in active mode for ransomware indicators across all client and server devices. This ensured an event like this would not occur again. 

Why does RESPOND work?

Response solutions must be accurate enough to fire only when there is a genuine threat, configurable enough to let the user stay in the driver’s seat, and intelligent enough to know the right action to take to contain only the malicious activity- without disrupting normal business operations. 

This is only possible if you can establish what ‘normal’ is for any one organization. And this is how Darktrace’s RESPOND product family ensures its actions are targeted and proportionate. By feeding off DETECT alerting which highlights subtle or large deviations across the network, cloud and SaaS, RESPOND can provide a measured response to the potential threat. This includes actions such as:

  • Enforcing the device’s ‘pattern of life’ for a given length of time 
  • Enforcing the ‘group pattern of life’ (stopping a device from doing anything its peers haven’t done in the past)
  • Blocking connections of a certain type to a certain destination
  • Logging out of a cloud account 
  • ‘Smart quarantining’ an endpoint device- maintaining access to VPNs and company’s AV solution

Conclusion 

In its report on BlackMatter [10], CISA recommended that organizations invest in network monitoring tools with the capacity to investigate anomalous activity. Picking up on unusual behavior rather than predetermined rules and signatures is an important step in fighting back against new threats. As this particular story shows, however, detection alone is not always enough. Turning on RESPOND, which takes immediate and precise action to contain threats, regardless of when and where they come in, is the best way to counter smash-and-grab attacks and protect organizations’ digital assets. There is little doubt that the threat actors behind BlackMatter will or have already returned with new names and strategies- but organizations with RESPOND will be ready for them.

Appendices

Darktrace Model Detections (in order of breach)

Those with the ‘PTN’ prefix were alerted directly to Darktrace’s 24/7 SOC team.

  • Device / ICMP Address Scan
  • Device / Suspicious SMB Scanning Activity
  • (PTN) Device / Suspicious Network Scan Activity
  • Anomalous Connection / SMB Enumeration
  • Device / Possible RPC Lateral Movement
  • Device / Active Directory Reconnaissance
  • Unusual Activity / Possible RPC Recon Activity
  • Device / Possible SMB/NTLM Reconnaissance
  • Compliance / Default Credential Usage
  • Device / New or Unusual Remote Command Execution
  • Anomalous Connection / New or Uncommon Service Control
  • Device / New or Uncommon SMB Named Pipe
  • Device / SMB Session Bruteforce
  • Device / New or Uncommon WMI Activity
  • (PTN) Device / Multiple Lateral Movement Model Breaches
  • Compromise / Sustained SSL or HTTP Increase
  • Compromise / SSL or HTTP Beacon
  • Compromise / Sustained TCP Beaconing Activity To Rare Endpoint
  • Device / Anomalous SMB Followed By Multiple Model Breaches
  • Device / Anomalous RDP Followed By Multiple Model Breaches
  • Anomalous Server Activity / Rare External from Server
  • Anomalous Connection / Anomalous SSL without SNI to New External
  • Anomalous Connection / Rare External SSL Self-Signed
  • Device / Long Agent Connection to New Endpoint
  • Compliance / SMB Drive Write
  • Anomalous Connection / Unusual Admin SMB Session
  • Anomalous Connection / High Volume of New or Uncommon Service Control
  • Anomalous Connection / Unusual Admin RDP Session
  • Device / Suspicious File Writes to Multiple Hidden SMB Shares
  • Anomalous Connection / Multiple Connections to New External TCP Port
  • Compliance / SSH to Rare External Destination
  • Anomalous Connection / Uncommon 1 GiB Outbound
  • Anomalous Connection / Data Sent to Rare Domain
  • Anomalous Connection / Download and Upload
  • (PTN) Unusual Activity / Enhanced Unusual External Data Transfer
  • Anomalous File / Internal / Additional Extension Appended to SMB File
  • (PTN) Compromise / Ransomware / Suspicious SMB Activity

List of IOCs 

Reference List 

[1] https://www.designnews.com/industrial-machinery/new-age-hackers-are-ditching-smash-and-grab-techniques 

[2] https://cybernews.com/cyber-war/how-do-smash-and-grab-cyberattacks-help-ukraine-in-waging-war/

[3] https://blog.google/threat-analysis-group/phishing-campaign-targets-youtube-creators-cookie-theft-malware/

[4] https://www.ukcybersecuritycouncil.org.uk/news-insights/articles/the-benefits-of-automation-to-cyber-security/

[5] https://techcrunch.com/2021/11/03/blackmatter-ransomware-shut-down/ 

[6] https://www.trellix.com/en-us/about/newsroom/stories/research/blackmatter-ransomware-analysis-the-dark-side-returns.html

[7] https://www.nytimes.com/2021/05/14/business/darkside-pipeline-hack.html

[8] https://techcrunch.com/2022/01/14/fsb-revil-ransomware/ 

[9] https://www.virustotal.com/gui/domain/georgiaonsale.com/community

[10] https://www.cisa.gov/uscert/ncas/alerts/aa21-291a

Credit to: Andras Balogh, SOC Analyst and Gabriel Few-Wiegratz, Threat Intelligence Content Production Lead

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
The Darktrace Analyst Team

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