May 10, 2024

Exploitation of ConnectWise ScreenConnect Vulnerabilities

Uncover the tactics used to exploit ConnectWise vulnerabilities and strategies to protect your systems.

Inside the SOC

Darktrace cyber analysts are world-class experts in threat intelligence, threat hunting and incident response, and provide 24/7 SOC support to thousands of Darktrace customers around the globe. Inside the SOC is exclusively authored by these experts, providing analysis of cyber incidents and threat trends, based on real-world experience in the field.

Written by

Justin Torres

Cyber Analyst

May 2024

Introduction

Across an ever changing cyber landscape, it is common place for threat actors to actively identify and exploit newly discovered vulnerabilities within commonly utilized services and applications. While attackers are likely to prioritize developing exploits for the more severe and global Common Vulnerabilities and Exposures (CVEs), they typically have the most success exploiting known vulnerabilities within the first couple years of disclosure to the public.

Addressing these vulnerabilities in a timely manner reduces the effectiveness of known vulnerabilities, decreasing the pace of malicious actor operations and forcing pursuit of more costly and time-consuming methods, such as zero-day related exploits or attacking software supply chain operations. While actors also develop tools to exploit other vulnerabilities, developing exploits for critical and publicly known vulnerabilities gives actors impactful tools at a low cost they are able to use for quite some time.

Between January and March 2024, the Darktrace Threat Research team investigated one such example that involved indicators of compromise (IoCs) suggesting the exploitation of vulnerabilities in ConnectWise’s remote monitoring and management (RMM) software ScreenConnect.

What are the ConnectWise ScreenConnect vulnerabilities?

CVE-2024-1708 is an authentication bypass vulnerability in ScreenConnect 23.9.7 (and all earlier versions) that, if exploited, would enable an attacker to execute remote code or directly impact confidential information or critical systems. This exploit would pave the way for a second ScreenConnect vunerability, CVE-2024-1709, which allows attackers to directly access confidential information or critical systems [1].

ConnectWise released a patch and automatically updated cloud versions of ScreenConnect 23.9.9, while urging security temas to update on-premise versions immediately [3].

If exploited in conjunction, these vulnerabilities could allow a malicious actor to create new administrative accounts on publicly exposed instances by evading existing security measures. This, in turn, could enable attackers to assume an administrative role and disable security tools, create backdoors, and disrupt RMM processes. Access to an organization’s environment in this manner poses serious risk, potentially leading to significant consequences such as deploying ransomware, as seen in various incidents involving the exploitation of ScreenConnect [2]

Darktrace Coverage of ConnectWise Exploitation

Darktrace’s anomaly-based detection was able to identify evidence of exploitation related to CVE-2024-1708 and CVE-2024-1709 across two distinct timelines; these detections included connectivity with endpoints that were later confirmed to be malicious by multiple open-source intelligence (OSINT) vendors. The activity observed by Darktrace suggests that threat actors were actively exploiting these vulnerabilities across multiple customer environments.

In the cases observed across the Darktrace fleet, Darktrace DETECT™ and Darktrace RESPOND™ were able to work in tandem to pre-emptively identify and contain network compromises from the onset. While Darktrace RESPOND was enabled in most customer environments affected by the ScreenConnect vulnerabilities, in the majority of cases it was configured in Human Confirmation mode. Whilst in Human Confirmation mode, RESPOND will provide recommended actions to mitigate ongoing attacks, but these actions require manual approval from human security teams.

When enabled in autonomous response mode, Darktrace RESPOND will take action automatically, shutting down suspicious activity as soon as it is detected without the need for human intervention. This is the ideal end state for RESPOND as actions can be taken at machine speed, without any delays waiting for user approval.

Looking within the patterns of activity observed by Darktrace , the typical attack timeline included:

Darktrace observed devices on affected customer networks performing activity indicative of ConnectWise ScreenConnect usage, for example connections over 80 and 8041, connections to screenconnect[.]com, and the use of the user agent “LabTech Agent”. OSINT research suggests that this user agent is an older name for ConnectWise Automate [5] which also includes ScreenConnect as standard [6].

Figure 1: Darktrace DETECT model alert highlighting the use of a remote management tool, namely “screenconnect[.]com”.

This activity was typically followed by anomalous connections to the external IP address 108.61.210[.]72 using URIs of the form “/MyUserName_DEVICEHOSTNAME”, as well as additional connections to another external, IP 185.62.58[.]132. Both of these external locations have since been reported as potentially malicious [14], with 185.62.58[.]132 in particular linked to ScreenConnect post-exploitation activity [2].

Figure 2: Darktrace DETECT model alert highlighting the unusual connection to 185.62.58[.]132 via port 8041.

Figure 3: Darktrace DETECT model alert highlighting connections to 108.61.210[.]72 using a new user agent and the “/MyUserName_DEVICEHOSTNAME” URI.

Same Exploit, Different Tactics?

While the majority of instances of ConnectWise ScreenConnect exploitation observed by Darktrace followed the above pattern of activity, Darktrace was able to identify some deviations from this.

In one customer environment, Darktrace’s detection of post-exploitation activity began with the same indicators of ScreenConnect usage, including connections to screenconnect[.]com via port 8041, followed by connections to unusual domains flagged as malicious by OSINT, in this case 116.0.56[.]101 [16] [17]. However, on this deployment Darktrace also observed threat actors downloading a suspicious AnyDesk installer from the endpoint with the URI “hxxp[:]//116.0.56[.]101[:]9191/images/Distribution.exe”.

Figure 4: Darktrace DETECT model alert highlighting the download of an unusual executable file from 116.0.56[.]101.

Further investigation by Darktrace’s Threat Research team revealed that this endpoint was associated with threat actors exploiting CVE-2024-1708 and CVE-2024-1709 [1]. Darktrace was additionally able to identify that, despite the customer being based in the United Kingdom, the file downloaded came from Pakistan. Darktrace recognized that this represented a deviation from the device’s expected pattern of activity and promptly alerted for it, bringing it to the attention of the customer.

Figure 5: External Sites Summary within the Darktrace UI pinpointing the geographic locations of external endpoints, in this case highlighting a file download from Pakistan.

Darktrace’s Autonomous Response

In this instance, the customer had Darktrace enabled in autonomous response mode and the post-exploitation activity was swiftly contained, preventing the attack from escalating.

As soon as the suspicious AnyDesk download was detected, Darktrace RESPOND applied targeted measures to prevent additional malicious activity. This included blocking connections to 116.0.56[.]101 and “*.56.101”, along with blocking all outgoing traffic from the device. Furthermore, RESPOND enforced a “pattern of life” on the device, restricting its activity to its learned behavior, allowing connections that are considered normal, but blocking any unusual deviations.

Figure 6: Darktrace RESPOND enforcing a “pattern of life” on the offending device after detecting the suspicious AnyDesk download.

Figure 7: Darktrace RESPOND blocking connections to the suspicious endpoint 116.0.56[.]101 and “*.56.101” following the download of the suspicious AnyDesk installer.

The customer was later able to use RESPOND to manually quarantine the offending device, ensuring that all incoming and outgoing traffic to or from the device was prohibited, thus preventing ay further malicious communication or lateral movement attempts.

Figure 8: The actions applied by Darktrace RESPOND in response to the post-exploitation activity related to the ScreenConnect vulnerabilities, including the manually applied “Quarantine device” action.

Conclusion

In the observed cases of the ConnectWise ScreenConnect vulnerabilities being exploited across the Darktrace fleet, Darktrace was able to pre-emptively identify and contain network compromises from the onset, offering vital protection against disruptive cyber-attacks.

While much of the post-exploitation activity observed by Darktrace remained the same across different customer environments, important deviations were also identified suggesting that threat actors may be adapting their tactics, techniques and procedures (TTPs) from campaign to campaign.

While new vulnerabilities will inevitably surface and threat actors will continually look for novel ways to evolve their methods, Darktrace’s Self-Learning AI and behavioral analysis offers organizations full visibility over new or unknown threats. Rather than relying on existing threat intelligence or static lists of “known bads”, Darktrace is able to detect emerging activity based on anomaly and respond to it without latency, safeguarding customer environments whilst causing minimal disruption to business operations.

Credit: Emma Foulger, Principal Cyber Analyst for their contribution to this blog.

Appendices

Darktrace Model Coverage

DETECT Models

Compromise / Agent Beacon (Medium Period)

Compromise / Agent Beacon (Long Period)

Anomalous File / EXE from Rare External Location

Device / New PowerShell User Agent

Anomalous Connection / Powershell to Rare External

Anomalous Connection / New User Agent to IP Without Hostname

User / New Admin Credentials on Client

Device / New User Agent

Anomalous Connection / Multiple HTTP POSTs to Rare Hostname

Anomalous Server Activity / Anomalous External Activity from Critical Network Device

Compromise / Suspicious Request Data

Compliance / Remote Management Tool On Server

Anomalous File / Anomalous Octet Stream (No User Agent)

RESPOND Models

Antigena / Network::External Threat::Antigena Suspicious File Block

Antigena / Network::External Threat::Antigena File then New Outbound Block

Antigena / Network::Significant Anomaly::Antigena Enhanced Monitoring from Client Block

Antigena / Network::Significant Anomaly::Antigena Significant Anomaly from Client Block

Antigena / Network::Significant Anomaly::Antigena Controlled and Model Breach

Antigena / Network::Insider Threat::Antigena Unusual Privileged User Activities Block

Antigena / Network / External Threat / Antigena Suspicious File Pattern of Life Block

Antigena / Network / Insider Threat / Antigena Unusual Privileged User Activities Pattern of Life Block

List of IoCs

IoC - Type - Description + Confidence

185.62.58[.]132 – IP- IP linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

108.61.210[.]72- IP - IP linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

116.0.56[.]101 - IP - IP linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

/MyUserName_ DEVICEHOSTNAME – URI - URI linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

/images/Distribution.exe – URI - URI linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

24780657328783ef50ae0964b23288e68841a421 - SHA1 Filehash - Filehash linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

a21768190f3b9feae33aaef660cb7a83 - MD5 Filehash - Filehash linked with threat actors exploiting CVE-2024-1708 and CVE-2024-17091

MITRE ATT&CK Mapping

Technique – Tactic – ID - Sub-technique of

Web Protocols - COMMAND AND CONTROL - T1071.001 - T1071

Web Services - RESOURCE DEVELOPMENT - T1583.006 - T1583

Drive-by Compromise - INITIAL ACCESS - T1189 – NA

Ingress Tool Transfer - COMMAND AND CONTROL - T1105 - NA

Malware - RESOURCE DEVELOPMENT - T1588.001- T1588

Exploitation of Remote Services - LATERAL MOVEMENT - T1210 – NA

PowerShell – EXECUTION - T1059.001 - T1059

Pass the Hash - DEFENSE EVASION, LATERAL MOVEMENT - T1550.002 - T1550

Valid Accounts - DEFENSE EVASION, PERSISTENCE, PRIVILEGE ESCALATION, INITIAL ACCESS - T1078 – NA

Man in the Browser – COLLECTION - T1185 - NA

Exploit Public-Facing Application - INITIAL ACCESS - T1190 - NA

Exfiltration Over C2 Channel – EXFILTRATION - T1041 – NA

IP Addresses – RECONNAISSANCE - T1590.005 - T1590

Remote Access Software - COMMAND AND CONTROL - T1219 – NA

Lateral Tool Transfer - LATERAL MOVEMENT - T1570 – NA

Application Layer Protocol - COMMAND AND CONTROL - T1071 – NA

References:

[1] https://unit42.paloaltonetworks.com/connectwise-threat-brief-cve-2024-1708-cve-2024-1709/

[2] https://www.huntress.com/blog/slashandgrab-screen-connect-post-exploitation-in-the-wild-cve-2024-1709-cve-2024-1708

[3] https://www.huntress.com/blog/a-catastrophe-for-control-understanding-the-screenconnect-authentication-bypass

[4] https://www.speedguide.net/port.php?port=8041

[5] https://www.connectwise.com/company/announcements/labtech-now-connectwise-automate

[6] https://www.connectwise.com/solutions/software-for-internal-it/automate

[7] https://www.securityweek.com/slashandgrab-screenconnect-vulnerability-widely-exploited-for-malware-delivery/

[8] https://arcticwolf.com/resources/blog/cve-2024-1709-cve-2024-1708-follow-up-active-exploitation-and-pocs-observed-for-critical-screenconnect-vulnerabilities/https://success.trendmicro.com/dcx/s/solution/000296805?language=en_US&sfdcIFrameOrigin=null

[9] https://www.connectwise.com/company/trust/security-bulletins/connectwise-screenconnect-23.9.8

[10] https://socradar.io/critical-vulnerabilities-in-connectwise-screenconnect-postgresql-jdbc-and-vmware-eap-cve-2024-1597-cve-2024-22245/

[11] https://www.trendmicro.com/en_us/research/24/b/threat-actor-groups-including-black-basta-are-exploiting-recent-.html

[12] https://otx.alienvault.com/indicator/ip/185.62.58.132

[13] https://www.virustotal.com/gui/ip-address/185.62.58.132/community

[14] https://www.virustotal.com/gui/ip-address/108.61.210.72/community

[15] https://otx.alienvault.com/indicator/ip/108.61.210.72

[16] https://www.virustotal.com/gui/ip-address/116.0.56[.]101/community

[17] https://otx.alienvault.com/indicator/ip/116.0.56[.]101

Written by

Justin Torres

Cyber Analyst

Inside the SOC

Written by

Justin Torres

Cyber Analyst

•

August 3, 2026

Nathaniel Jones

SVP, Global Threat Intelligence

Watch the NIS2 Webinar

Blog

August 5, 2026

Testing a Prompt injection Attack Against an Enterprise AI Agent

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:

Claude generated a prompt injection payload.
Hidden instructions were embedded in an email.
The email passed traditional validation checks.
Darktrace analyzed the language and sender behavior.
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:

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.

About the author

Carlo Loregian

Solutions Engineer

Blog

August 4, 2026

Extending AI Security Visibility with Darktrace and Microsoft Agent 365

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.

About the author

The Darktrace Community

Your data. Our AI.

Elevate your network security with Darktrace AI

Get a demo

Check out this article by Darktrace: ScreenConnect Vulnerability: Detection & Response

Exploitation of ConnectWise ScreenConnect Vulnerabilities

Introduction

What are the ConnectWise ScreenConnect vulnerabilities?

Darktrace Coverage of ConnectWise Exploitation

Same Exploit, Different Tactics?

Darktrace’s Autonomous Response

Conclusion

Appendices

Darktrace Model Coverage

List of IoCs

References:

Why Trust is the New Attack Surface: Darktrace’s Mid-Year Threat Update 2026

Testing a Prompt injection Attack Against an Enterprise AI Agent

Enjoying the blog?

More in this series

Hiding in Plain Sight: Uncovering a Multi-Stage Ransomware Attack Through Behavioral Detection

From Click to Command: Behavioral Detection of AppleScript-Led MacOS Intrusions

Hola VPN Abuse: From Proxy Traffic to Malware and Cryptomining

Blog

August 5, 2026

Testing a Prompt injection Attack Against an Enterprise AI Agent

Key takeaways

How behavioral detection helps stop prompt injection attacks

How do prompt injection attacks work?

Prompt injection attack example

How we tested prompt injection attacks on an enterprise agent

How Darktrace detected a prompt injection attack

Why behavioral security makes a difference detecting prompt injection attacks

Why enterprise AI security requires more than guardrails

How Darktrace detects prompt injection attacks in emails

How to secure enterprise AI operations from prompt injection attacks

Blog

August 4, 2026

Extending AI Security Visibility with Darktrace and Microsoft Agent 365

Extending visibility into the Microsoft Agent 365 experience

Why unified visibility of AI agent risk signals matters

Extending Microsoft's AI security model, not replacing it

Building the future of secure AI