Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Havoc is an open-source command and control framework whose Demon implant uses sleep encryption, indirect syscalls, and return address spoofing to resist memory scanning.

What Is Havoc C2?
Havoc is an open-source command and control framework released in 2022 by the developer known as C5pider. It consists of a team server written in Go, an operator client in C++ and Qt, and an implant called Demon written in C and assembly. It is free, publicly available on GitHub, and implements evasion techniques that until recently were found mainly in commercial products and in the tooling of well-resourced intrusion sets.
Its arrival marked a shift worth understanding. The techniques that defeat endpoint detection used to require either a licensed product costing thousands of pounds a year or the ability to write them yourself. Havoc, along with Sliver and Mythic, moved that capability into the open.
Demon’s design is organised around one objective: survive inspection of its own process.
ntdll.dll to observe what a process asks the kernel to do. Demon resolves system call numbers and invokes them in a way that routes execution through legitimate ntdll addresses, so the call neither passes through the hook nor originates from an obviously suspicious location.Communication supports HTTP, HTTPS, and SMB named pipes for peer-to-peer tasking within a segmented network, and the team server exposes a Python scripting API for automation.
Havoc has appeared in real intrusions rather than remaining a research curiosity. In early 2023 Zscaler’s threat research team documented a campaign against a government organisation in which a Havoc Demon implant was delivered through a multi-stage loader, chosen in part because the framework was less well recognised than Cobalt Strike. Other vendors have since reported Havoc alongside Sliver in intrusions where the operators appeared to be deliberately avoiding the most heavily signatured tooling.
That is the pattern to expect. As detection for one framework matures, operators move to the next, and the open-source options have no licensing friction to slow adoption.
Signature-based approaches fare badly here, and the evasion features are specifically aimed at the memory scanning that catches older frameworks. The approaches that still work:
ntdll has been reloaded from disk, or that AMSI and ETW functions have been modified in memory, is a high-fidelity signal in its own right.None of this is specific to Havoc, which is the point. Detections written against implant behaviour generalise across frameworks, while detections written against one tool’s indicators expire when the operator switches.
Make sure the telemetry exists: ETW Threat Intelligence enabled, PowerShell script block logging on, Sysmon deployed with image load and named pipe events, and endpoint detection in a mode that permits memory inspection rather than file scanning alone. Reduce the execution surface with application allowlisting. Constrain egress so that an implant beaconing to unfamiliar infrastructure is visible on the network as well as the host.
Then test it. Run a current framework in a controlled environment, map each technique to MITRE ATT&CK, and record which steps produced telemetry, which produced an alert, and which produced nothing. A purple team session of this kind is worth more than any number of indicator feeds.
At Privia Security, our SOC Maturation Services measure detection coverage against current implant tradecraft and build the rules that close the gaps.
Three reasons. Detection for Cobalt Strike has matured, so using it now means being found faster. Open-source frameworks cost nothing and carry no licensing or watermarking that could assist attribution. And the modern ones implement sleep encryption, indirect syscalls, and stack spoofing out of the box, which previously required either a commercial product or significant development effort.
An implant spends most of its life dormant between check-ins, and that is when memory scanners typically find it. Sleep obfuscation encrypts the implant’s own memory before sleeping and decrypts it on waking, so a scan during the dormant period sees encrypted bytes. The counter is to trigger scanning on activity such as thread creation or network connection instead.
Mature products catch default configurations, and the detections that work are behavioural rather than signature-based: unbacked executable memory, implausible call stacks, ETW Threat Intelligence events for memory protection changes, and evidence that ntdll or AMSI has been tampered with. A skilled operator with a custom loader raises the difficulty considerably, which is why layered network and identity controls still matter.
Run a current framework in a controlled lab, execute specific techniques deliberately, and check three things for each one: whether telemetry was recorded, whether an alert fired, and whether an analyst would have understood it. Map the results to MITRE ATT&CK, write detections for the gaps, and re-run the technique to confirm the new rule works.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.