Blog
/
Email
/
August 11, 2021

How One Email Compromised an Entire Logistics Company

A single phishing email led to a massive compromise at a logistics company in Europe. Discover the importance of email security with increasing SaaS usage.
No items found.
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.
No items found.
Default blog image
11
Aug 2021

Organizations are only as secure as their weakest link. In many cases, that weak link arises in the various cloud applications an organization relies on. Several high-profile groups including APT28 are known to exploit commonly-used passwords to bruteforce their way into businesses around the world. These ‘spray’ campaigns often target Microsoft Office 365 accounts and will only become more frequent as the use of SaaS increases.

This blog analyses how a single phishing email slipped under the radar of the gateway and other traditional tools in place, and eventually led to mass compromise at a logistics company in Europe.

Logistical nightmare

Logistics operators play a critical role across every industry sector. Managing the distribution of goods and services from the seller to the customer, they enable – or bottleneck – an efficient supply chain. Inevitably, logistics companies have become an attractive target for cyber-criminals, due to the high number of organizations they interact with, the pressure they’re under to deliver on time, and the sensitive data they often handle.

It is a simple equation for attackers: do they put in the hard work to infiltrate 20 well-defended organizations, or compromise just one, and from there gain access to all 20 or more? The majority of cyber-threats Darktrace has observed this year have gone for the latter – exploiting less protected third parties to gain a foothold across a range of businesses.

The vaccine supply in particular has fallen under attack, numerous times. Last autumn, threat actors infiltrated a German biomedical organization and launched a phishing campaign to harvest credentials and compromise several organizations involved in the COVID-19 cold chain.

Alongside ransomware, phishing attacks are one of the most pressing concerns facing the industry.

Breaking the chain

At a medium-sized logistics company, a user received one phishing email from a hijacked third party. The email came from a trusted source with a well established history of sending emails, so it easily passed the gateway.

Once the phishing email had reached the inbox, the user clicked on the malicious link and was led to a fake login page, where they were tricked into divulging their credentials.

Four days later, the attacker logged into the account from an unusual location, and proceeded to read files with sensitive information.

The next day, Darktrace detected a new email rule from another unusual location. Almost immediately, a large volume of outbound emails was sent from the account, all containing the suspicious link.

Figure 1: Timeline of the attack — the total dwell time was five days.

Supply and disrupt

Once you are inside an organization’s digital ecosystem, it is easy to move around and compromise more accounts. Most security tools and employees do not question an internal email sent by a trusted user, especially if the user is a senior figure with authority.

So, after this set of outbound emails, unusual activity from anomalous locations was duly seen on other company accounts. These users had been tricked into giving away their details from the emails supposedly sent by their colleague.

More sensitive customer files were read, followed by a second spike in outbound emails from these hijacked accounts.

This time, the emails were sent not internally, but to external contacts. The contacts likely were conducting business with the logistics company at the time, and so were used to receiving emails from the accounts.

In total, over 450 phishing emails were sent to a wide range of third parties. Many of these third parties in turn had their credentials compromised – repeating the cycle once again.

Figure 2: Cyber AI Analyst investigates the suspicious activity of a compromised user, providing a detailed summary with the unusual login location and actions carried out.

Hanging by a thread: The threat of third-party attacks

The source of the initial phishing email that kickstarted this attack was itself from a legitimate third party known to the customer, where presumably the same thing had occured.

This form of Vendor Email Compromise, which can be rinsed and repeated to form a vicious loop, is notoriously difficult for email security solutions to detect, and can lead to heavy reputational and financial damage. To complicate matters, acting against a suspicious email from a known sender can also cause severe business disruption if it turns out to be legitimate.

Because of this, security must move beyond the binary approach of ‘good’ and ‘bad’, towards a more holistic understanding of the contextual setting surrounding any email interaction.

Darktrace accurately detected the multiple anomalies when comparing it to other emails from senders of the same domain. It sent high-priority alerts to the security team, but could not prevent the email from reaching the inbox because it was only in detection mode.

Figure 3: Darktrace’s automatic summary of the initial phishing email gives an overview of the suspicious aspects of the email.

‍

The phishing links during the attack used a third-party tool called Piktochart, designed to create various type of files such as infographics, charts, and forms. While Piktochart has several legitimate applications, it can also be exploited. Gateways thus have a hard time distinguishing between legitimate and malicious Piktochart links. In this case, the gateway rewrote the initial link for analysis, but did not identify it as malicious.

In comparison, Darktrace for Email easily identified the email to be suspicious because it noticed it was out of character for that particular sender, and because the link itself was suspicious. In active mode, the AI would have locked the link and moved the email to the Junk folder, effectively preventing the very first step of the attack and avoiding any further compromise.

Figure 4: Piktochart was rarely seen on the deployment up until this point – the domain was 100% rare. Darktrace therefore easily detected the anomalous nature of this third-party tool usage.

The butterfly effect

Most cyber-attacks begin with just a single point of entry – that is all an attacker requires. One phishing email can be enough to bring a whole supply chain to its knees. With 94% of cyber-attacks beginning in the inbox, and suppliers and vendors in constant communication over multiple SaaS platforms – including Microsoft Teams and Google Cloud – email security tools must be capable of detecting when a trusted third party is acting abnormally.

Especially with the rise of remote working, SaaS usage has surged in businesses worldwide and many have been forced to turn to cloud and SaaS to enable a flexible workforce. While there are obvious benefits, these additions have expanded the attack surface and stretched the limits of traditional security and human security teams.

When it comes to logistics companies – who often act as the middle man in global operations – credential harvesting not only has serious consequences for the customer, but for anyone in the customer’s email contacts, and can lead to major breaches for numerous people and businesses.

Figure 5: Darktrace’s user interface reveals the two spikes in outbound emails that were sent out by compromised company accounts.

Thanks to Darktrace analyst Emma Foulger for her insights on the above threat find.

Learn more about the threats facing logistics providers

Darktrace model detections:

  • SaaS / Compliance / New Email Rule
  • SaaS / Unusual Login and New Email Rule
  • Antigena Email models included
  • Unusual / Unusual Login Location and New Unknown Link
  • Link / Account Hijack Link
  • Link / Outlook Hijack
  • Internal Compromise / Recipient Surge from Unusual Login Location (outbound emails)
  • Internal Compromise / Recipient Surge with Suspicious Content (outbound emails)

‍

No items found.
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.
No items found.

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