Blog
/
/
January 13, 2025

Agent vs. Agentless Cloud Security: Why Deployment Methods Matter

Cloud security solutions can be deployed with agentless or agent-based approaches or use a combination of methods. Organizations must weigh which method applies best to the assets and data the tool will protect.
Written by
Kellie Regan
Director, Product Marketing - Cloud Security
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
Kellie Regan
Director, Product Marketing - Cloud Security
Default blog image
13
Jan 2025

The rapid adoption of cloud technologies has brought significant security challenges for organizations of all sizes. According to recent studies, over 70% of enterprises now operate in hybrid or multi-cloud environments, with 93% employing a multi-cloud strategy[1]. This complexity requires robust security tools, but opinions vary on the best deployment method—agent-based, agentless, or a combination of both.

Agent-based and agentless cloud security approaches offer distinct benefits and limitations, and organizations often make deployment choices based on their unique needs depending on the function of the specific assets covered, the types of data stored, and cloud architecture, such as hybrid or multi-cloud deployments.

For example, agentless solutions are increasingly favored for their ease of deployment and ability to provide broad visibility across dynamic cloud environments. These are especially useful for DevOps teams, with 64% of organizations citing faster deployment as a key reason for adopting agentless tools[2].

On the other hand, agent-based solutions remain the preferred choice for environments requiring deep monitoring and granular control, such as securing sensitive high-value workloads in industries like finance and healthcare. In fact, over 50% of enterprises with critical infrastructure report relying on agent-based solutions for their advanced protection capabilities[3].

As the debate continues, many organizations are turning to combined approaches, leveraging the strengths of both agent-based and agentless tools to address the full spectrum of their security needs for comprehensive coverage. Understanding the capabilities and limitations of these methods is critical to building an effective cloud security strategy that adapts to evolving threats and complex infrastructures.

Agent-based cloud security

Agent-based security solutions involve deploying software agents on each device or system that needs protection. Agent-based solutions are great choices when you need in-depth monitoring and protection capabilities. They are ideal for organizations that require deep security controls and real-time active response, particularly in hybrid and on-premises environments.

Key advantages include:

1. Real-time monitoring and protection: Agents detect and block threats like malware, ransomware, and anomalous behaviors in real time, providing ongoing protection and enforcing compliance by continuously monitoring workload activities.  Agents enable full control over workloads for active response such as blocking IP addresses, killing processes, disabling accounts, and isolating infected systems from the network, stopping lateral movement.

2. Deep visibility for hybrid environments: Agent-based approaches allow for full visibility across on-premises, hybrid, and multi-cloud environments by deploying agents on physical and virtual machines. Agents offer detailed insights into system behavior, including processes, files, memory, network connections, and more, detecting subtle anomalies that might indicate security threats. Host-based monitoring tracks vulnerabilities at the system and application level, including unpatched software, rogue processes, and unauthorized network activity.

3. Comprehensive coverage: Agents are very effective in hybrid environments (cloud and on-premises), as they can be installed on both physical and virtual machines.  Agents can function independently on each host device onto which they are installed, which is especially helpful for endpoints that may operate outside of constant network connectivity.

Challenges:

1. Resource-intensive: Agents can consume CPU, memory, and network resources, which may affect performance, especially in environments with large numbers of workloads or ephemeral resources.

2. Challenging in dynamic environments: Managing hundreds or thousands of agents in highly dynamic or ephemeral environments (e.g., containers, serverless functions) can be complex and labor-intensive.

3. Slower deployment: Requires agent installation on each workload or instance, which can be time-consuming, particularly in large or complex environments.  

Agentless cloud security

Agentless security does not require software agents to be installed on each device. Instead, it uses cloud infrastructure and APIs to perform security checks. Agentless solutions are highly scalable with minimal impact on performance, and ideal for cloud-native and highly dynamic environments like serverless and containerized. These solutions are great choices for your cloud-native and multi-cloud environments where rapid deployment, scalability, and minimal impact on performance are critical, but response actions can be handled through external tools or manual processes.

Key advantages include:

1. Scalability and ease of deployment: Because agentless security doesn’t require installation on each individual device, it is much easier to deploy and can quickly scale across a vast number of cloud assets. This approach is ideal for environments where resources are frequently created and destroyed (e.g., serverless, containerized workloads), as there is no need for agent installation or maintenance.

2. Reduced system overhead: Without the need to run local agents, agentless security minimizes the impact on system performance. This is crucial in high-performance environments.

3. Broad visibility: Agentless security connects via API to cloud service providers, offering near-instant visibility and threat detection. It provides a comprehensive view of your cloud environment, making it easier to manage and secure large and complex infrastructures.

Challenges

1. Infrastructure-level monitoring: Agentless solutions rely on cloud service provider logs and API calls, meaning that detection might not be as immediate as agent-based solutions. They collect configuration data and logs, focusing on infrastructure misconfigurations, identity risks, exposed resources, and network traffic, but lack visibility and access to detailed, system-level information such as running processes and host-level vulnerabilities.

2. Cloud-focused: Primarily for cloud environments, although some tools may integrate with on-premises systems through API-based data gathering. For organizations with hybrid cloud environments, this approach fragments visibility and security, leading to blind spots and increasing security risk.

3. Passive remediation: Typically provides alerts and recommendations, but lacks deep control over workloads, requiring manual intervention or orchestration tools (e.g., SOAR platforms) to execute responses. Some agentless tools trigger automated responses via cloud provider APIs (e.g., revoking permissions, adjusting security groups), but with limited scope.

Combined agent-based and agentless approaches

A combined approach leverages the strengths of both agent-based and agentless security for complete coverage. This hybrid strategy helps security teams achieve comprehensive coverage by:

  • Using agent-based solutions for deep, real-time protection and detailed monitoring of critical systems or sensitive workloads.
  • Employing agentless solutions for fast deployment, broader visibility, and easier scalability across all cloud assets, which is particularly useful in dynamic cloud environments where workloads frequently change.

The combined approach has distinct practical applications. For example, imagine a financial services company that deals with sensitive transactions. Its security team might use agent-based security for critical databases to ensure stringent protections are in place. Meanwhile, agentless solutions could be ideal for less critical, transient workloads in the cloud, where rapid scalability and minimal performance impact are priorities. With different data types and infrastructures, the combined approach is best.

Best of both worlds: The benefits of a combined approach

The combined approach not only maximizes security efficacy but also aligns with diverse operational needs. This means that all parts of the cloud environment are secured according to their risk profile and functional requirements. Agent-based deployment provides in-depth monitoring and active protection against threats, suitable for environments requiring tight security controls, such as financial services or healthcare data processing systems. Agentless deployment complements agents by offering broader visibility and easier scalability across diverse and dynamic cloud environments, ideal for rapidly changing cloud resources.

There are three major benefits from combining agent-based and agentless approaches.

1. Building a holistic security posture: By integrating both agent-based and agentless technologies, organizations can ensure that all parts of their cloud environments are covered—from persistent, high-risk endpoints to transient cloud resources. This comprehensive coverage is crucial for detecting and responding to threats promptly and effectively.

2. Reducing overhead while boosting scalability: Agentless systems require no software installation on each device, reducing overhead and eliminating the need to update and maintain agents on a large number of endpoints. This makes it easier to scale security as the organization grows or as the cloud environment changes.

3. Applying targeted protection where needed: Agent-based solutions can be deployed on selected assets that handle sensitive information or are critical to business operations, thus providing focused protection without incurring the costs and complexity of universal deployment.

Use cases for a combined approach

A combined approach gives security teams the flexibility to deploy agent-based and agentless solutions based on the specific security requirements of different assets and environments. As a result, organizations can optimize their security expenditures and operational efforts, allowing for greater adaptability in cloud security use cases.

Let’s take a look at how this could practically play out. In the combined approach, agent-based security can perform the following:

1. Deep monitoring and real-time protection:

  • Workload threat detection: Agent-based solutions monitor individual workloads for suspicious activity, such as unauthorized file changes or unusual resource usage, providing high granularity for detecting threats within critical cloud applications.
  • Behavioral analysis of applications: By deploying agents on virtual machines or containers, organizations can monitor behavior patterns and flag anomalies indicative of insider threats, lateral movement, or Advanced Persistent Threats (APTs).
  • Protecting high-sensitivity environments: Agents provide continuous monitoring and advanced threat protection for environments processing sensitive data, such as payment processing systems or healthcare records, leveraging capabilities like memory protection and file integrity monitoring.

2. Cloud asset protection:

  • Securing critical infrastructure: Agent-based deployments are ideal for assets like databases or storage systems that require real-time defense against exploits and ransomware.
  • Advanced packet inspection: For high-value assets, agents offer deep packet inspection and in-depth logging to detect stealthy attacks such as data exfiltration.
  • Customizable threat response: Agents allow for tailored security rules and automated responses at the workload level, such as shutting down compromised instances or quarantining infected files.

At the same time, agentless cloud security provides complementary benefits such as:

1. Broad visibility and compliance:‍

  • Asset discovery and management: Agentless systems can quickly scan the entire cloud environment to identify and inventory all assets, a crucial capability for maintaining compliance with regulations like GDPR or HIPAA, which require up-to-date records of data locations and usage.
  • Regulatory compliance auditing and configuration management: Quickly identify gaps in compliance frameworks like PCI DSS or SOC 2 by scanning configurations, permissions, and audit trails without installing agents. Using APIs to check configurations across cloud services ensures that all instances comply with organizational and regulatory standards, an essential aspect for maintaining security hygiene and compliance.
  • Shadow IT Detection: Detect and map unauthorized cloud services or assets that are spun up without security oversight, ensuring full inventory coverage.

2. Rapid environmental assessment:

  • Vulnerability assessment of new deployments: In environments where new code is frequently deployed, agentless security can quickly assess new instances, containers, or workloads in CI/CD pipelines for vulnerabilities and misconfigurations, enabling secure deployments at DevOps speed.
  • Misconfiguration alerts: Detect and alert on common cloud configuration issues, such as exposed storage buckets or overly permissive IAM roles, across cloud providers like AWS, Azure, and GCP.
  • Policy enforcement: Validate that new resources adhere to established security baselines and organizational policies, preventing security drift during rapid cloud scaling.

Combining agent-based and agentless approaches in cloud security not only maximizes the protective capabilities, but also offers flexibility, efficiency, and comprehensive coverage tailored to the diverse and evolving needs of modern cloud environments. This integrated strategy ensures that organizations can protect their assets more effectively while also adapting quickly to new threats and regulatory requirements.

Darktrace offers complementary and flexible deployment options for holistic cloud security

Powered by multilayered AI, Darktrace / CLOUD is a Cloud Detection and Response (CDR) solution that is agentless by default, with optional lightweight, host-based server agents for enhanced real-time actioning and deep inspection. As such, it can deploy in cloud environments in minutes and provide unified visibility and security across hybrid, multi-cloud environments.

With any deployment method, Darktrace supports multi-tenant, hybrid, and serverless cloud environments. Its Self-Learning AI learns the normal behavior across architectures, assets, and users to identify unusual activity that may indicate a threat. With this approach, Darktrace / CLOUD quickly disarms threats, whether they are known, unknown, or completely novel. It then accelerates the investigation process and responds to threats at machine speed.

Learn more about how Darktrace / CLOUD secures multi and hybrid cloud environments in the Solution Brief.

References:

1. Flexera 2023 State of the Cloud Report

2. ESG Research 2023 Report on Cloud-Native Security

3. Gartner, Market Guide for Cloud Workload Protection Platforms, 2023

Written by
Kellie Regan
Director, Product Marketing - Cloud Security
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
Kellie Regan
Director, Product Marketing - Cloud Security

A Chain Reaction: Blockchain-Hosted Infostealer Campaign Targets Windows and macOS

Network
•
September 24, 2026
Nahisha Nobregas
SOC Analyst
Watch the NIS2 Webinar
Trending blogs
1
Prompt Security in Enterprise AI: Strengths, Weaknesses, and Common Approaches
May 20, 2026
2
How Empowering End Users can Improve Your Email Security and Decrease the Burden on the SOC
May 8, 2024
3
Elevating Network Security: Confronting Trust, Ransomware, & Novel Attacks
Jun 21, 2024
4
The State of AI in Cybersecurity: Unveiling Global Insights from 1,800 Security Practitioners
Apr 9, 2024
5
The State of AI in Cybersecurity: The Impact of AI on Cybersecurity Solutions
May 13, 2024

More in this series

No items found.
Continue reading
No items found.

Blog

/

Network

/

September 25, 2026

A Chain Reaction: Blockchain-Hosted Infostealer Campaign Targets Windows and macOS

Default blog imageDefault blog image

Key Insights

  • Darktrace detected a blockchain-hosted infostealer campaign targeting Windows and macOS devices across multiple customer environments.
  • The campaign combined ClickFix social engineering with trusted services and decentralized blockchain infrastructure to support malware delivery and C2 activity.
  • Compromised devices were observed connecting to rare and unusual external endpoints, including DGA C2 domains, blockchain-related endpoints, and cryptocurrency mining infrastructure.
  • The activity was associated with information-stealing malware strains including Atomic macOS Stealer (AMOS), Lumma, Rhadamanthys, Vidar, and Phexia.
  • Darktrace identified anomalous device behavior, beaconing patterns, rare external connections, cryptomining activity, and suspicious TLS/SSL communications without relying solely on prior knowledge or static indicators of compromise.
  • The campaign highlights how attackers are increasingly using legitimate and decentralized infrastructure to make detection, disruption, and attribution more challenging for defenders.

The Infostealer Ecosystem

The information stealer malware ecosystem continues to grow in value for threat actors across the digital threat landscape. Infostealers are increasingly delivered through Malware-as-a-Service (MaaS) operating models, distributed through affiliate networks, and designed to withstand infrastructure takedowns. This resilience was demonstrated by the recent takedown of Lumma Stealer malicious domains by Microsoft’s Digital Crimes Unit (DCU) [1].

Infostealers are used to gather and exfiltrate sensitive information, including non-human identity (NHI) data, from compromised systems across cloud, Software-as-a-Service (SaaS), Virtual Private Network (VPN), and development environments. They can also support ransomware operations by expanding the credentials and access paths available to threat actors, contributing to the high volume of identity-based attacks observed across the broader threat landscape [2][3].

Darktrace’s Observations of ClickFix and Infostealers

Throughout 2026, Darktrace has observed multiple campaigns using ClickFix social engineering to trick users into carrying out malicious actions and downloading initial payloads, including information stealers. More recently, Darktrace’s Threat Research team identified a specific ClickFix campaign involving a blockchain-hosted infostealer targeting Windows and macOS devices.

Darktrace identified affected customer environments across Europe, the United States, Asia, and the Middle East where blockchain-hosted infostealer malware appears to have been delivered to compromised systems following likely ClickFix-driven initial access. Darktrace investigated the activity and found that decentralized blockchain infrastructure, alongside widely trusted legitimate services, was used to support malware delivery and information theft across Windows and macOS systems.

Following initial access, compromised systems established C2 communication, with C2 configuration and payloads hosted on public blockchain infrastructure. The ultimate objective appears to be credential and cryptocurrency theft through the deployment of information stealers such as Atomic macOS Stealer (AMOS), Lumma, Rhadamanthys, and Vidar [5][6][7].

Darktrace’s Investigation

Affected devices across the Darktrace customer base were observed making outbound connections to rare external endpoints in patterns consistent with beaconing and C2 activity. Darktrace primarily detected devices making repeated connections to algorithmically generated domains (DGA) such as hf98x4d[.]site [8]. In many cases, these domains were linked through open-source intelligence (OSINT) to information-stealing malware families including AMOS and Phexia [5][6][7][8][9].

In multiple cases, devices were also observed connecting to blockchain-related endpoints, such as polygon[.]drpc[.]org, as well as legitimate public services, including GitHub. The use of decentralized blockchain infrastructure and trusted services such as GitHub to facilitate malware distribution and C2 activity can make disruption and attribution significantly more difficult for defenders.

Darktrace alsodetected a significant proportion of impacted devices making outboundconnections to cryptocurrency mining infrastructure associated with thelegitimate open-source XMRig mining software and the HashVault mining pool,including pool.hashvault[.]pro and donate[.]ssl[.]xmrig[.]com, which wereabused by the attackers, indicating, includingpool.hashvault[.]pro and donate[.]ssl[.]xmrig[.]com, indicating active cryptominingon compromised systems.

In one case, mining activity was observed before and during connections to the DGA endpoint hf98x4d[.]site. Due to its highly anomalous nature, Darktrace's Real-Time AI Analyst autonomously investigated the activity as it occurred, correlating the two events into a single cryptocurrency mining incident and providing comprehensive visibility into the broader attack.

‍

Figure 1: Real-Time AI Analyst investigation of suspicious SSL and C2 communications with hf98x4d[.]site over port 443.

‍

Figure 2: Real-Time AI Analyst investigation into cryptocurrency mining activity involving pool[.]hashvault[.]pro over SSL on port 443.

‍

Around the same time, Darktrace identified the same device initiating connections to the GitHub endpoint release-assets[.]githubusercontent[.]com while continuing to make repeated connections to hf98x4d[.]site.

‍

Figure 3: Darktrace's detection of an affected device connecting to a GitHub endpoint between repeated connections to the anomalous external endpoint hf98x4d[.]site.

On the network of another customer, Darktrace observed an affected device making highly unusual outbound connections consistent with beaconing activity. The device initiated multiple connections over port 443 to the external hostname polygon[.]drpc[.]org. According to OSINT, this hostname is a Remote Procedure Call (RPC) endpoint provided by dRPC, a legitimate service enabling decentralized applications (dApps), cryptocurrency wallets, and developer tools to interact with the Polygon blockchain [10].

The same device was later observed making repeated TLS/SSL connections to the previously mentioned DGA C2 domain. In addition, it made outbound connections to the external IP 195.242.214[.]34 over destination port 51820, an endpoint associated with the ProtonVPN service. Collectively, these connections to blockchain-related infrastructure, the DGA C2 domain, and ProtonVPN-associated infrastructure suggested the device had been affected by the campaign.

Conclusion

This campaign demonstrates how attackers can combine ClickFix social engineering with trusted services and decentralized blockchain infrastructure to create a resilient, cross-platform malware delivery chain. By using services such as GitHub alongside blockchain RPC endpoints and rapidly replaceable DGA domains, the activity can blend into legitimate traffic while making infrastructure disruption and attribution more difficult.

For defenders, it’s a reminder that trusted infrastructure does not automatically mean trusted activity. Security teams should look for the behaviors surrounding these connections, including unusual outbound communication, repeated beaconing, unexpected access to blockchain services, suspicious TLS/SSL activity and cryptomining. In this campaign, Darktrace identified and correlated these deviations without depending solely on previously known indicators, providing visibility as affected devices moved between legitimate services, decentralized infrastructure and malicious C2 endpoints

Credit to Nahisha Nobregas (Associate Principal Cyber Analyst), Manoel Kadja (Senior Cyber Analyst)

Edited by Ryan Traill (Content Manager)

Appendices

Darktrace Model Detections

▪ Compromise / Beaconing Activity To External Rare

▪ Compromise / Beacon to Young Endpoint

▪ Compromise / Fast Beaconing to DGA

▪ Compromise / High Volume of Connections with Beacon Score

▪ Compromise / DGA Beacon

▪ Compromise / Slow Beaconing Activity To External Rare

▪ Compromise / Agent Beacon (Long Period)

▪ Compromise / Agent Beacon (Medium Period)

▪ Compromise / Sustained SSL or HTTP Increase

▪ Compromise / Large Number of Suspicious Failed Connections

▪ Compromise / SSL Beaconing to Rare Destination

▪ Compromise / Beacon for 4 Days

▪ Compromise / High Priority Crypto Currency Mining

▪ Compromise / Monero Mining

▪ Device / Long Agent Connection to New Endpoint

▪ Device / New Connections On Suspicious Port

▪ Anomalous Connection / High Volume of Connections to Rare Domain

‍

‍

List of Indicators of Compromise (IoCs)

 
Indicator Description
hf98x4d[.]site C2 Endpoint (Hostname)
sj98xe4[.]xyz C2 Endpoint (Hostname)
citcix6[.]xyz C2 Endpoint (Hostname)
bduwih8[.]pro C2 Endpoint (Hostname)

‍

‍

MITRE ATT&CK Mapping

 
Tactic (ID) Technique
Persistence (T1176) Browser Extensions (T1176.001)
Persistence (T1176) Software Extensions
Command and Control (T1071) Web Protocols (T1071.001)
Command and Control (T1568) Domain Generation Algorithms (T1568.002)
Command and Control (T1071) Application Layer Protocol
Command and Control (T1102) One-Way Communication (T1102.003)
Command and Control (T1571) Non-Standard Port
Command and Control (T1104) Multi-Stage Channels
Command and Control (T1573) Encrypted Channel
Command and Control (T1008) Fallback Channels
Initial Access ICS (T0862) Supply Chain Compromise
Command and Control ICS (T0885) Commonly Used Port
Collection (T1185) Browser Session Hijacking
Impact (T1496) Compute Hijacking (T1496.001)
Impact (T1496) Resource Hijacking
Command and Control (T1071) Publish/Subscribe Protocols (T1071.001)
Lateral Movement (T1210) Exploitation of Remote Services

‍

References:

1.        https://www.microsoft.com/en-us/security/blog/2025/05/21/lumma-stealer-breaking-down-the-delivery-techniques-and-capabilities-of-a-prolific-infostealer/

2.        https://spycloud.com/resource/report/spycloud-annual-identity-exposure-report-2026/

3.        https://www.darktrace.com/blog/why-trust-is-the-new-attack-surface-darktraces-mid-year-threat-update-2026

4.        https://www.darktrace.com/blog/unpacking-clickfix-darktraces-detection-of-a-prolific-social-engineering-tactic

5.        https://abekweng.medium.com/inside-a-blockchain-hosted-malware-campaign-targeting-windows-and-macos-f5bcdeffed66

6.        https://cloud.google.com/blog/topics/threat-intelligence/unc5142-etherhiding-distribute-malware

7.        https://haveibeensquatted.com/blog/from-typosquatting-to-macos-backdoor-clickfix-blockchain-c2

8.        https://www.virustotal.com/gui/domain/hf98x4d.site/community

9.        https://x.com/FABO97662188/status/2074125545026244795

10.  https://www.virustotal.com/gui/url/b0e5c51a411065864119c305fddf218b7c120731f655932cc1c3307ad5b43f94/gti-summary

Continue reading
About the author
Nahisha Nobregas
SOC Analyst

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
Your data. Our AI.
Elevate your network security with Darktrace AI