September 23, 2025

ShadowV2: An emerging DDoS for hire botnet

Darktrace exposed a cybercrime-as-a-service campaign using Python and Go-based malware, Docker containerization, and a full operator UI. With DDoS-as-a-service features, modular APIs, and advanced evasion, this platform highlights the need for defenders to monitor cloud workloads, container orchestration, and API activity to counter evolving threats.

Written by

Nate Bill

Threat Researcher

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

Nate Bill

Threat Researcher

ShadowV2: An emerging DDoS for hire botnet

Sep 2025

Introduction: ShadowV2 DDoS

Darktrace's latest investigation uncovered a novel campaign that blends traditional malware with modern devops technology.

At the center of this campaign is a Python-based command-and-control (C2) framework hosted on GitHub CodeSpaces. This campaign also utilizes a Python based spreader with a multi-stage Docker deployment as the initial access vector.

The campaign further makes use of a Go-based Remote Access Trojan (RAT) that implements a RESTful registration and polling mechanism, enabling command execution and communication with its operators.

ShadowV2 attack techniques

What sets this campaign apart is the sophistication of its attack toolkit.

‍The threat actors employ advanced methods such as HTTP/2 rapid reset, a Cloudflare under attack mode (UAM) bypass, and large-scale HTTP floods, demonstrating a capability to combine distributed denial-of-service (DDoS) techniques with targeted exploitation.

With the inclusion of an OpenAPI specification, implemented with FastAPI and Pydantic and a fully developed login panel and operator interface, the infrastructure seems to resemble a “DDoS-as-a-service” platform rather than a traditional botnet, showing the extent to which modern malware increasingly mirrors legitimate cloud-native applications in both design and usability.

Analysis of a ShadowV2 attack

Initial access

The initial compromise originates from a Python script hosted on GitHub CodeSpaces. This can be inferred from the observed headers:

User-Agent: docker-sdk-python/7.1.0

X-Meta-Source-Client: github/codespaces

The user agent shows that the attacker is using the Python Docker SDK, a library for Python programs that allows them to interact with Docker to create containers. The X-Meta-Source-Client appears to have been injected by GitHub into the request to allow for attribution, although there is no documentation online about this header.

The IP the connections originate from is 23.97.62[.]139, which is a Microsoft IP based in Singapore. This aligns with expectations as GitHub is owned by Microsoft.

This campaign targets exposed Docker daemons, specifically those running on AWS EC2. Darktrace runs a number of honeypots across multiple cloud providers and has only observed attacks against honeypots running on AWS EC2. By default, Docker is not accessible to the Internet, however, can be configured to allow external access. This can be useful for managing complex deployments where remote access to the Docker API is needed.

Typically, most campaigns targeting Docker will either take an existing image from Docker Hub and deploy their tools within it, or upload their own pre-prepared image to deploy. This campaign works slightly differently; it first spawns a generic “setup” container and installs a number of tools within it. This container is then imaged and deployed as a live container with the malware arguments passed in via environmental variables.

Figure 1: Attacker creates a blank container from an Ubuntu image.

Figure 2: Attacker sets up their tools for the attack.

Figure 3: Attacker deploys a new container using the image from the setup container.

It is unclear why the attackers chose this approach - one possibility is that the actor is attempting to avoid inadvertently leaving forensic artifacts by performing the build on the victim machine, rather than building it themselves and uploading it.

Malware analysis

The Docker container acts as a wrapper around a single binary, dropped in /app/deployment. This is an ELF binary written in Go, a popular choice for modern malware. Helpfully, the binary is unstripped, making analysis significantly easier.

The current version of the malware has not been reported by OSINT providers such as VirusTotal. Using the domain name from the MASTER_ADDR variable and other IoCs, we were able to locate two older versions of the malware that were submitted to VirusTotal on the June 25 and July 30 respectively [1] [2]. Neither of these had any detections and were only submitted once each using the web portal from the US and Canada respectively. Darktrace first observed the attack against its honeypot on June 24, so it could be a victim of this campaign submitting the malware to VirusTotal. Due to the proximity of the start of the attacks, it could also be the attacker testing for detections, however it is not possible to know for certain.

The malware begins by phoning home, using the MASTER_ADDR and VPS_NAME identifiers passed in from the Docker run environmental variables. In addition, the malware derives a unique VPS_ID, which is the VPS_NAME concatenated with the current unix timestamp. The VPS_ID is used for all communications with the C2 server as the identifier for the specific implant. If the malware is restarted, or the victim is re-infected, the C2 server will inform the implant of its original VPS_ID to ensure continuity.

Figure 4: Snippet that performs the registration by sending a POST request to the C2 API with a JSON structure.

From there, the malware then spawns two main loops that will remain active for the lifetime of the implant. Every second, it sends a heartbeat to the C2 by sending the VPS_ID to hxxps://shadow.aurozacloud[.]xyz/api/vps/heartbeat via POST request. Every 5 seconds, it retrieves hxxps://shadow.aurozacloud[.]xyz/api/vps/poll/<VPS ID> via a GET request to poll for new commands.

The poll mechanism shadow v2 — Figure 5: The poll mechanism.

At this stage, Darktrace security researchers wrote a custom client that ran on the server infected by the attacker that mimicked their implant. The goal was to intercept commands from the C2. Based on this, it was observed initiating an attack against chache08[.]werkecdn[.]me using a 120 thread HTTP2 rapid reset attack. This site appears to be hosted on an Amsterdam VPS provided by FDCServers, a server hosting company. It was not possible to identify what normally runs on this site, as it returns a 403 Forbidden error when visited.

Darktrace’s code analysis found that the returned commands contain the following fields:

Method (e.g. GET, POST)
A unique ID for the attack
A URL endpoint used to report attack statistics
The target URL & port
The duration of the attack
The number of threads to use
An optional proxy to send HTTP requests through

The malware then spins up several threads, each running a configurable number of HTTP clients using Valyala’s fasthttp library, an open source Go library for making high-performance HTTP requests. After this is complete, it uses these clients to perform an HTTP flood attack against the target.

Figure 6: A snippet showing the fasthttp client creation loop, as well as a function to report the worker count back to the C2.

In addition, it also features several flags to enable different bypass mechanisms to augment the malware:

WordPress bypass (does not appear to be implemented - the flag is not used anywhere)
Random query strings appended to the URL
Spoofed forwarding headers with random IP addresses
Cloudflare under-attack-mode (UAM) bypass
HTTP2 rapid reset

The most interesting of these is the Cloudflare UAM bypass mechanism. When this is enabled, the malware will attempt to use a bundled ChromeDP binary to solve the Cloudflare JavaScript challenge that is presented to new visitors. If this succeeds, the clearance cookie obtained is then included in subsequent requests. This is unlikely to work in most cases as headless Chrome browsers are often flagged, and a regular CAPTCHA is instead served.

Figure 7: The UAM bypass success snippet.

Additionally, the malware has a flag to enable an HTTP2 rapid reset attack mode instead of a regular HTTP flood. In HTTP2, a client can create thousands of requests within a single connection using multiplexing, allowing sites to load faster. The number of request streams per connection is capped however, so in a rapid reset attack many requests are made and then immediately cancelled to allow more requests to be created. This allows a single client to execute vastly more requests per second and use more server resources than it otherwise would, allowing for more effective denial-of-service (DoS) attacks.

Figure 8: The HTTP2 rapid reset snippet from the main attack function.

API/C2 analysis

As mentioned throughout the malware analysis section, the malware communicates with a C2 server using HTTP. The server is behind Cloudflare, which obscures its hosting location and prevents analysis. However, based on analysis of the spreader, it's likely running on GitHub CodeSpaces.

When sending a malformed request to the API, an error generated by the Pydantic library is returned:

{"detail":[{"type":"missing","loc":["body","vps_id"],"msg":"Field required","input":{"vps_name":"xxxxx"},"url":"https://errors.pydantic.dev/2.11/v/missing"}]}

This shows they are using Python for the API, which is the same language that the spreader is written in.

One of the larger frameworks that ships with Pydantic is FastAPI, which also ships with Swagger. The malware author left this publicly exposed, and Darktrace’s researchers were able to obtain a copy of their API documentation. The author appears to have noticed this however, as subsequent attempts to access it now returns a HTTP 404 Not Found error.

Figure 9: Swagger UI view based on the obtained OpenAPI spec.

This is useful to have as it shows all the API endpoints, including the exact fields they take and return, along with comments on each endpoint written by the attacker themselves.

It is very likely a DDoS for hire platform (or at the very least, designed for multi-tenant use) based on the extensive user API, which features authentication, distinctions between privilege level (admin vs user), and limitations on what types of attack a user can execute. The screenshot below shows the admin-only user create endpoint, with the default limits.

The admin-only user create endpoint shadow v2 — Figure 10: The admin-only user create endpoint.

The endpoint used to launch attacks can also be seen, which lines up with the options previously seen in the malware itself. Interestingly, this endpoint requires a list of zombie systems to launch the attack from. This is unusual as most DDoS for hire services will decide this internally or just launch the attack from every infected host (zombie). No endpoints that returned a list of zombies were found, however, it’s possible one exists as the return types are not documented for all the API endpoints.

The attack start endpoint shadow v2 — Figure 11: The attack start endpoint.

There is also an endpoint to manage a blacklist of hosts that cannot be attacked. This could be to stop users from launching attacks against sites operated by the malware author, however it’s also possible the author could be attempting to sell protection to victims, which has been seen previously with other DDoS for hire services.

Blacklist endpoints shadow v2 DDoS — Figure 12: Blacklist endpoints.

Attempting to visit shadow[.]aurozacloud[.]xyz results in a seizure notice. It is most likely fake the same backend is still in use and all of the API endpoints continue to work. Appending /login to the end of the path instead brings up the login screen for the DDoS platform. It describes itself as an “advanced attack platform”, which highlights that it is almost certainly a DDoS for hire service. The UI is high quality, written in Tailwind, and even features animations.

Conclusion

By leveraging containerization, an extensive API, and with a full user interface, this campaign shows the continued development of cybercrime-as-a-service. The ability to deliver modular functionality through a Go-based RAT and expose a structured API for operator interaction highlights how sophisticated some threat actors are.

For defenders, the implications are significant. Effective defense requires deep visibility into containerized environments, continuous monitoring of cloud workloads, and behavioral analytics capable of identifying anomalous API usage and container orchestration patterns. The presence of a DDoS-as-a-service panel with full user functionality further emphasizes the need for defenders to think of these campaigns not as isolated tools but as evolving platforms.

Appendices

References

1. https://www.virustotal.com/gui/file/1b552d19a3083572bc433714dfbc2b75eb6930a644696dedd600f9bd755042f6 ‍

2. https://www.virustotal.com/gui/file/1f70c78c018175a3e4fa2b3822f1a3bd48a3b923d1fbdeaa5446960ca8133e9c

IoCs

Malware hashes (SHA256)

● 2462467c89b4a62619d0b2957b21876dc4871db41b5d5fe230aa7ad107504c99

● 1b552d19a3083572bc433714dfbc2b75eb6930a644696dedd600f9bd755042f6

● 1f70c78c018175a3e4fa2b3822f1a3bd48a3b923d1fbdeaa5446960ca8133e9c

C2 domain

● shadow.aurozacloud[.]xyz

Spreader IPs

● 23.97.62[.]139

● 23.97.62[.]136

Yara rule

rule ShadowV2 {

meta:

author = "[email protected]"

description = "Detects ShadowV2 botnet implant"

strings:

$string1 = "shadow-go"

$string2 = "shadow.aurozacloud.xyz"

$string3 = "[SHADOW-NODE]"

$symbol1 = "main.registerWithMaster"

$symbol2 = "main.handleStartAttack"

$symbol3 = "attacker.bypassUAM"

$symbol4 = "attacker.performHTTP2RapidReset"

$code1 = { 48 8B 05 ?? ?? ?? ?? 48 8B 1D ?? ?? ?? ?? E8 ?? ?? ?? ?? 48 8D 0D ?? ?? ?? ?? 48 89 8C 24 38 01 00 00 48 89 84 24 40 01 00 00 48 8B 4C 24 40 48 BA 00 09 6E 88 F1 FF FF FF 48 8D 04 0A E8 ?? ?? ?? ?? 48 8D 0D ?? ?? ?? ?? 48 89 8C 24 48 01 00 00 48 89 84 24 50 01 00 00 48 8D 05 ?? ?? ?? ?? BB 05 00 00 00 48 8D 8C 24 38 01 00 00 BF 02 00 00 00 48 89 FE E8 ?? ?? ?? ?? }

$code2 = { 48 89 35 ?? ?? ?? ?? 0F B6 94 24 80 02 00 00 88 15 ?? ?? ?? ?? 0F B6 94 24 81 02 00 00 88 15 ?? ?? ?? ?? 0F B6 94 24 82 02 00 00 88 15 ?? ?? ?? ?? 0F B6 94 24 83 02 00 00 88 15 ?? ?? ?? ?? 48 8B 05 ?? ?? ?? ?? }

$code3 = { 48 8D 15 ?? ?? ?? ?? 48 89 94 24 68 04 00 00 48 C7 84 24 78 04 00 00 15 00 00 00 48 8D 15 ?? ?? ?? ?? 48 89 94 24 70 04 00 00 48 8D 15 ?? ?? ?? ?? 48 89 94 24 80 04 00 00 48 8D 35 ?? ?? ?? ?? 48 89 B4 24 88 04 00 00 90 }

condition:

uint16(0) == 0x457f and (2 of ($string*) or 2 of ($symbol*) or any of ($code*))

}

‍

^{The content provided in this blog is published by Darktrace for general informational purposes only and reflects our understanding of cybersecurity topics, trends, incidents, and developments at the time of publication. While we strive to ensure accuracy and relevance, the information is provided “as is” without any representations or warranties, express or implied. Darktrace makes no guarantees regarding the completeness, accuracy, reliability, or timeliness of any information presented and expressly disclaims all warranties.}

^{Nothing in this blog constitutes legal, technical, or professional advice, and readers should consult qualified professionals before acting on any information contained herein. Any references to third-party organizations, technologies, threat actors, or incidents are for informational purposes only and do not imply affiliation, endorsement, or recommendation.}

^{Darktrace, its affiliates, employees, or agents shall not be held liable for any loss, damage, or harm arising from the use of or reliance on the information in this blog.}

^{The cybersecurity landscape evolves rapidly, and blog content may become outdated or superseded. We reserve the right to update, modify, or remove any content without notice.}

Written by

Nate Bill

Threat Researcher

Inside the SOC

Written by

Nate Bill

Threat Researcher

Associate Principal Analyst & Regional Consultancy Lead

•

March 26, 2026

The Darktrace Community

Watch the NIS2 Webinar

Blog

Network

March 26, 2026

Phantom Footprints: Tracking GhostSocks Malware

Why are attackers using residential proxies?

In today's threat landscape, blending in to normal activity is the key to success for attackers and the growing reliance on residential proxies shows a significant shift in how threat actors are attempting to bypass IP detection tools.

The increasing dependency on residential proxies has exposed how prevalent proxy services are and how reliant a diverse range of threat actors are on them. From cybercriminal groups to state‑sponsored actors, the need to bypass IP detection tools is fundamental to the success of these groups. One malware that has quietly become notorious for its ability to avoid anomaly detection is GhostSocks, a malware that turns compromised devices into residential proxies.

What is GhostSocks?

Originally marketed on the Russian underground forum xss[.]is as a Malware‑as‑a‑Service (MaaS), GhostSocks enables threat actors to turn compromised devices into residential proxies, leveraging the victim's internet bandwidth to route malicious traffic through it.

How does Ghostsocks malware work?

The malware offers the threat actor a “clean” IP address, making it look like it is coming from a household user. This enables the bypassing of geographic restrictions and IP detection tools, a perfect tool for avoiding anomaly detection. It wasn’t until 2024, when a partnership was announced with the infamous information stealer Lumma Stealer, that GhostSocks surged into widespread adoption and alluded to who may be the author of the proxy malware.

Written in GoLang, GhostSocks utilizes the SOCKS5 proxy protocol, creating a SOCKS5 connection on infected devices. It uses a relay‑based C2 implementation, where an intermediary server sits in between the real command-and-control (C2) server and the infected device.

How does Ghostsocks malware evade detection?

To further increase evasion, the Ghostsocks malware wraps its SOCKS5 tunnels in TLS encryption, allowing its malicious traffic to blend into normal network traffic.

Early variants of GhostSocks do not implement a persistence mechanism; however, later versions achieve persistence via registry run keys, ensuring sustained proxy operational time [1].

While proxying is its primary purpose, GhostSocks also incorporates backdoor functionality, enabling malicious actors to run arbitrary commands and download and deploy additional malicious payloads. This was evident with the well‑known ransomware group Black Basta, which reportedly used GhostSocks as a way of maintaining long‑term access to victims’ networks [1].

Darktrace’s detection of GhostSocks Malware

Darktrace observed a steady increase in GhostSocks activity across its customer base from late 2025, with its Threat Research team identifying multiple incidents involving the malware. In one notable case from December 2025, Darktrace detected GhostSocks operating alongside Lumma Stealer, reinforcing that the partnership between Lumma and GhostSocks remains active despite recent attempts to disrupt Lumma’s infrastructure.

Darktrace’s first detection of GhostSocks‑related activity came when a device on the network of a customer in the education sector began making connections to an endpoint with a suspicious self‑signed certificate that had never been seen on the network before.

The endpoint in question, 159.89.46[.]92 with the hostname retreaw[.]click, has been flagged by multiple open‑source intelligence (OSINT) sources as being associated with Lumma Stealer’s C2 infrastructure [2], indicating its likely role in the delivery of malicious payloads.

Figure 1: Darktrace’s detection of suspicious SSL connections to retreaw[.]click, indicating an attempted link to Lumma C2 infrastructure.

Less than two minutes later, Darktrace observed the same device downloading the executable (.exe) file “Renewable.exe” from the IP 86.54.24[.]29, which Darktrace recognized as 100% rare for this network.

Figure 2: Darktrace’s detection of a device downloading the unusual executable file “Renewable.exe”.

Both the file MD5 hash and the executable itself have been identified by multiple OSINT vendors as being associated with the GhostSocks malware [3], with the executable likely the backdoor component of the GhostSocks malware, facilitating the distribution of additional malicious payloads [4].

Following this detection, Darktrace’s Autonomous Response capability recommended a blocking action for the device in an early attempt to stop the malicious file download. In this instance, Darktrace was configured in Human Confirmation Mode, meaning the customer’s security team was required to manually apply any mitigative response actions. Had Autonomous Response been fully enabled at the time of the attack, the connections to 86.54.24[.]29 would have been blocked, rendering the malware ineffective at reaching its C2 infrastructure and halting any further malicious communication.

Figure 3: Darktrace’s Autonomous Response capability suggesting blocking the suspicious connections to the unusual endpoint from which the malicious executable was downloaded.

As the attack was able to progress, two days later the device was detected downloading additional payloads from the endpoint www.lbfs[.]site (23.106.58[.]48), including “Setup.exe”, “,.exe”, and “/vp6c63yoz.exe”.

Figure 4: Darktrace’s detection of a malicious payload being downloaded from the endpoint www.lbfs[.]site.

Once again, Darktrace recognized the anomalous nature of these downloads and suggested that a “group pattern of life” be enforced on the offending device in an attempt to contain the activity. By enforcing a pattern of life on a device, Darktrace restricts its activity to connections and behaviors similar to those performed by peer devices within the same group, while still allowing it to carry out its expected activity, effectively preventing deviations indicative of compromise while minimizing disruption. As mentioned earlier, these mitigative actions required manual implementation, so the activity was able to continue. Darktrace proceeded to suggest further actions to contain subsequent malicious downloads, including an attempt to block all outbound traffic to stop the attack from progressing.

Figure 5: An overview of download activity and the Autonomous Response actions recommended by Darktrace to block the downloads.

Around the same time, a third executable download was detected, this time from the hostname hxxp[://]d2ihv8ymzp14lr.cloudfront.net/2021-08-19/udppump[.]exe, along with the file “udppump.exe”.While GhostSocks may have been present only to facilitate the delivery of additional payloads, there is no indication that these CloudFront endpoints or files are functionally linked to GhostSocks. Rather, the evidence points to broader malicious file‑download activity.

Shortly after the multiple executable files had been downloaded, Darktrace observed the device initiating a series of repeated successful connections to several rare external endpoints, behavior consistent with early-stage C2 beaconing activity.

Cyber AI Analyst’s investigation

Figure 7: Darktrace’s detection of additional malicious file downloads from malicious CloudFront endpoints.

Throughout the course of this attack, Darktrace’s Cyber AI Analyst carried out its own autonomous investigation, piecing together seemingly separate events into one wider incident encompassing the first suspicious downloads beginning on December 4, the unusual connectivity to many suspicious IPs that followed, and the successful beaconing activity observed two days later. By analyzing these events in real-time and viewing them as part of the bigger picture, Cyber AI Analyst was able to construct an in‑depth breakdown of the attack to aid the customer’s investigation and remediation efforts.

Figure 8: Cyber AI Analyst investigation detailing the sequence of events on the compromised device, highlighting its extensive connectivity to rare endpoints, the related malicious file‑download activity, and finally the emergence of C2 beaconing behavior.

Conclusion

The versatility offered by GhostSocks is far from new, but its ability to convert compromised devices into residential proxy nodes, while enabling long‑term, covert network access—illustrates how threat actors continue to maximise the value of their victims’ infrastructure. Its growing popularity, coupled with its ongoing partnership with Lumma, demonstrates that infrastructure takedowns alone are insufficient; as long as threat actors remain committed to maintaining anonymity and can rapidly rebuild their ecosystems, related malware activity is likely to persist in some form.

Credit to Isabel Evans (Cyber Analyst), Gernice Lee (Associate Principal Analyst & Regional Consultancy Lead – APJ)
Edited by Ryan Traill (Content Manager)

Appendices

References

1. https://bloo.io/research/malware/ghostsocks

2. https://www.virustotal.com/gui/domain/retreaw.click/community

3. https://synthient.com/blog/ghostsocks-from-initial-access-to-residential-proxy

4. https://www.joesandbox.com/analysis/1810568/0/html

5. https://www.virustotal.com/gui/url/fab6525bf6e77249b74736cb74501a9491109dc7950688b3ae898354eb920413

Darktrace Model Detections

Real-time Detection Models

Anomalous Connection / Suspicious Self-Signed SSL

Anomalous Connection / Rare External SSL Self-Signed

Anomalous File / EXE from Rare External Location

Anomalous File / Multiple EXE from Rare External Locations

Compromise / Possible Fast Flux C2 Activity

Compromise / Large Number of Suspicious Successful Connections

Compromise / Large Number of Suspicious Failed Connections

Compromise / Sustained SSL or HTTP Increase

Autonomous Response Models

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

Antigena / Network / External Threat / Antigena Suspicious File Block

Antigena / Network / Significant Anomaly / Antigena Controlled and Model Alert

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

Antigena / Network / Significant Anomaly / Antigena Alerts Over Time Block

Antigena / Network / External Threat / Antigena Suspicious Activity Block

MITRE ATT&CK Mapping

Tactic – Technique – Sub-Technique

Resource Development – T1588 - Malware

Initial Access - T1189 - Drive-by Compromise

Persistence – T1112 – Modify Registry

Command and Control – T1071 – Application Layer Protocol

Command and Control – T1095 – Non-application Layer Protocol

Command and Control – T1071 – Web Protocols

Command and Control – T1571 – Non-Standard Port

Command and Control – T1102 – One-Way Communication

List of Indicators of Compromise (IoCs)

86.54.24[.]29 - IP - Likely GhostSocks C2

http[://]86.54.24[.]29/Renewable[.]exe - Hostname - GhostSocks Distribution Endpoint

http[://]d2ihv8ymzp14lr.cloudfront[.]net/2021-08-19/udppump[.]exe - CDN - Payload Distribution Endpoint

www.lbfs[.]site - Hostname - Likely C2 Endpoint

retreaw[.]click - Hostname - Lumma C2 Endpoint

alltipi[.]com - Hostname - Possible C2 Endpoint

w2.bruggebogeyed[.]site - Hostname - Possible C2 Endpoint

9b90c62299d4bed2e0752e2e1fc777ac50308534 - SHA1 file hash – Likely GhostSocks payload

3d9d7a7905e46a3e39a45405cb010c1baa735f9e - SHA1 file hash - Likely follow-up payload

10f928e00a1ed0181992a1e4771673566a02f4e3 - SHA1 file hash - Likely follow-up payload

About the author

Gernice Lee

Associate Principal Analyst & Regional Consultancy Lead

Blog

AI

March 26, 2026

State of AI Cybersecurity 2026: 92% of security professionals concerned about the impact of AI agents

The findings in this blog are taken from Darktrace's annual State of AI Cybersecurity Report 2026.

AI is already embedded in day-to-day enterprise activity, with 78% of participants in one recent survey reporting that their organizations are using generative AI in at least one business function. Generative AI now acts as an always-on assistant, researcher, creator, and coach across an expanding array of departments and functions. Autonomous agents are performing multi-step operational workflows from end to end. AI features have been layered on top of every SaaS application. And vibe coding is making it possible for employees without deep technical expertise to build their own AI-powered automations.

According to Gartner, more than 80% of enterprises will have deployed GenAI models, applications, or APIs in production environments by the end of this year, up from less than 5% in 2023. Companies report a 130% increase in spending on AI over the same period, with 72% of business leaders using AI tools at least weekly. The outsized efficiency and productivity gains that were once a future vision are quickly becoming everyday reality.

AI is currently driving business growth and innovation, and organizations risk falling behind peers if they don’t keep up with the pace of adoption, but it is also quietly expanding the enterprise attack surface. The modern CISO is challenged to both enable innovation and protect the business from these emerging threats.

AI agents introduce new risks and vulnerabilities

AI agents are playing growing roles in enterprise production environments. In many cases, these agents act with broad permissions across multiple software systems and platforms. This means they’re granted far-reaching access – to sensitive data, business-critical applications, tokens and APIs, and IT and security tools. With this access comes risk for security leaders – 92% are concerned about the use of AI agents across the workforce and their impact on security.

These agents must be governed as identities, with least-privilege access and ongoing monitoring. They can’t be thought of as invisible aspects of the application estate. Understanding how AI agents behave, and how to manage their permissions, control their behavior, and limit their data access will be a top security priority throughout 2026.

Generative AI prompts: The next frontier

Prompts are how users – both human and agentic – interact with AI systems, and they’re where natural language gets translated into model behavior. Natural language is infinite in its potential combinations and permutations, making this aspect of the attack surface open-ended and far more complex than traditional CVEs. With carefully crafted prompts, bad actors may be able to coax models into disclosing sensitive data, bypassing guardrails, or initiating undesirable actions.

Among security leaders, the biggest worries about AI usage in their environments all involve ways that systems might be manipulated to bypass traditional controls.

61% are most concerned about the exposure of sensitive data
56% are most concerned about potential data security and policy violations
51% are most concerned about the misuse or abuse of AI tools

The more employees rely on AI in their day-to-day workflows, the more critical it becomes for security teams to understand how prompt behavior determines model behavior – and where that behavior could go wrong.

What does “securing AI” mean in practice?

AI adoption opens new security risks that blur the boundaries between traditional security disciplines. A single malicious interaction with an AI model could involve identity misuse, sensitive data exposure, application logic abuse, and supply chain risk – all within a single workflow. Protecting this dynamic and rapidly evolving attack surface requires an approach that spans identity security, cloud security, application security, data security, software development security, and more.

The task for security leaders is to implement the tools, policies, and frameworks to mitigate these novel, expansive, and cross-disciplinary risks.

However, within most enterprises, AI policy creation remains in its infancy. Just 37% of security leaders report that their organization has a formal AI policy, representing a small but worrisome decrease from last year. Conversations about AI abound: in 52% of organizations, there’s discussion about an AI policy. Still, talk is cheap, and leaders will need to take action if they’re to successfully enable secure AI innovation.

To govern and protect their AI systems, organizations must take a multi-pronged approach. This requires building out policies, but it also demands that they are able to:

Monitor the prompts driving GenAI assistants and agents in real time. Organizations must be able to inspect prompts, sessions, and responses across enterprise GenAI tools, low- and high-code environments, and SaaS and SASE so that they can detect clever conversational prompt attacks and malicious chaining.
Secure all business AI agent identities. Security teams need to identify all the agents acting within their environment and supply chain, map their connections and interactions via MCP and services like Amazon S3, and audit their behavior across the cloud, SaaS environments, and on the network and endpoint devices.
Maintain centralized, comprehensive visibility. Understanding intent, assessing risks, and enforcing policies all require that security teams have a single view that spans AI interactions across the entire business.
Discover and control shadow AI. Teams need to be able to identify unsanctioned AI activities, distinguish the misuse of legitimate tools from their appropriate use, and apply policies to protect data, while guiding users towards approved solutions.

Scaling AI safely and responsibly

The approach that most cybersecurity vendors have taken – using historical patterns to predict future threats – doesn’t work well for AI systems. Because AI changes its behavior in response to the information it encounters while taking action, previous patterns don’t indicate what it will do next. Looking at past attacks can’t tell you how complex models will behave in your individual business.

Securing AI requires interpreting ambiguous interactions, uncovering subtleties that reveal intent within extended conversations, understanding how access accumulates over time, and recognizing when behavior – both human and machine – begins to drift towards areas of risk. To do this, you need to understand what “normal” looks like in each unique organization: how users, systems, applications, and AI agents behave, how they communicate, and how data flows between them.

Darktrace has spent more than a decade designing AI-powered solutions that can understand and adapt to evolving behavior in complex environments. This technology learns directly from the environment it protects, identifying malicious actions that deviate from normal operations, so that it can stop AI-related threats on the very first encounter.

As AI adoption reshapes enterprise operations, humans and machines will collaborate more and more often. This collaboration might dramatically expand the attack surface, but it also has the potential to be a force multiplier for defenders.

‍

Explore the full State of AI Cybersecurity 2026 report for deeper insights into how security leaders are responding to AI-driven risks.

Learn more about securing AI in your enterprise.

‍

[related-resource]

About the author

The Darktrace Community

Your data. Our AI.

Elevate your network security with Darktrace AI

Get a demo

ShadowV2: An emerging DDoS for hire botnet

Introduction: ShadowV2 DDoS

ShadowV2 attack techniques

Analysis of a ShadowV2 attack

Initial access

Malware analysis

API/C2 analysis

Conclusion

Appendices

References

IoCs

Yara rule

Phantom Footprints: Tracking GhostSocks Malware

State of AI Cybersecurity 2026: 92% of security professionals concerned about the impact of AI agents

Enjoying the blog?

More in this series

Phantom Footprints: Tracking GhostSocks Malware

When Reality Diverges from the Playbook: Darktrace Identifies Encryption in a World Leaks Ransomware Attack

NetSupport RAT: How Legitimate Tools Can Be as Damaging as Malware

Blog

Network

March 26, 2026

Phantom Footprints: Tracking GhostSocks Malware

Why are attackers using residential proxies?

What is GhostSocks?

How does Ghostsocks malware work?

How does Ghostsocks malware evade detection?

Darktrace’s detection of GhostSocks Malware

Cyber AI Analyst’s investigation

Conclusion

Appendices

References

Darktrace Model Detections

MITRE ATT&CK Mapping

List of Indicators of Compromise (IoCs)

Blog

AI

March 26, 2026

State of AI Cybersecurity 2026: 92% of security professionals concerned about the impact of AI agents

AI agents introduce new risks and vulnerabilities

Generative AI prompts: The next frontier

What does “securing AI” mean in practice?

Scaling AI safely and responsibly