Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Process tree analysis reads the parent-child relationships between running processes. Most intrusions are visible as a chain of processes that no legitimate workflow would ever produce.

What Is Process Tree Analysis?
Every process on a Windows host was started by another process, and that relationship is recorded. Reconstructing the chain — which process launched which, with what command line, and when — is process tree analysis, and it remains one of the highest-yield techniques in endpoint detection. Malicious code rarely looks unusual on its own. It looks unusual in context, because the thing that launched it had no business doing so.
Consider a user opening an invoice. winword.exe starting cmd.exe, which starts powershell.exe, which starts rundll32.exe, describes a document execution chain that no version of Word produces in the course of displaying a document. Nothing in that sequence involves an unsigned binary or a file on disk that an antivirus product would object to. The chain itself is the detection.
Three sources provide process lineage, with different fidelity.
ProcessGuid that uniquely identifies each process even after a PID is reused.Whatever the source, two fields do most of the work: the parent image and the full command line. A detection programme that collects process creation events without command lines is collecting half a log.
winword.exe, excel.exe, or outlook.exe as the parent of cmd.exe, powershell.exe, wscript.exe, mshta.exe, or regsvr32.exe.w3wp.exe, httpd.exe, or a Java application server as the parent of a command interpreter is the classic signature of a web shell, and it was how a great many organisations first learned they had been hit by the ProxyShell and Log4Shell exploitation waves.certutil.exe downloading, bitsadmin.exe transferring, or msbuild.exe started by anything other than a developer tool.rundll32.exe is started with comsvcs.dll MiniDump arguments, or any process opens a handle to lsass.exe that is not an expected security product.Attackers know this technique is watched, and two methods break naive analysis.
Parent process ID spoofing lets a process declare a different parent at creation time using the PROC_THREAD_ATTRIBUTE_PARENT_PROCESS attribute. A payload launched from a malicious document can present itself as a child of explorer.exe, which looks entirely normal. The defence is to compare the recorded parent against the process that actually ran: the spoofed parent’s own start time, user context, and integrity level frequently fail to line up, and Sysmon’s ParentProcessGuid makes that comparison possible.
Process injection and hollowing remove the chain altogether, because the malicious code runs inside a process that started legitimately. Sysmon Event ID 8 for remote thread creation, Event ID 10 for process access, and Event ID 25 for process tampering are what catch these, which is why process tree analysis is a component of endpoint detection rather than the whole of it.
Orphaned processes are the mundane version of the same problem. When a parent exits, the child keeps running with a parent ID that now points at nothing or has been reused by an unrelated process. Analysis that relies on PIDs alone will produce nonsense trees; analysis that uses GUIDs and creation timestamps will not.
Start from a small set of parent-child pairs that should never occur in your estate, measure how often they occur, and tune from there. Most organisations find the list is shorter and quieter than they expected. Allowlisting is specific rather than general: exclude the one build server that legitimately spawns compilers from Office automation, not every host in the subnet.
Record the chains you confirm as malicious in your hunting hypothesis library, and express the resulting logic in a portable format so it survives a change of platform. Coverage is best measured against MITRE ATT&CK, where execution, defence evasion, and credential access techniques map directly onto specific lineages.
At Privia Security, we deploy and tune endpoint telemetry of this kind, including Sysmon configuration, through our SOC maturation services.
Often not, because modern endpoint detection products collect the same process lineage and keep it searchable. Sysmon still earns its place where agent licences do not reach, where you want a vendor-independent copy of the telemetry in your own SIEM, or where you need specific events the product does not expose. Running both on the same host is reasonable but doubles the event volume.
Compare the claimed parent against reality. A spoofed parent often started long before the child, runs as a different user, or sits at a different integrity level than the child would inherit. Sysmon records the parent process GUID and both creation times, so a query that flags children whose claimed parent exited before they started, or whose user context differs, catches most attempts.
Office applications spawning command interpreters and script hosts. It is a narrow, well-understood pattern, it covers the initial execution stage of a large proportion of real intrusions, and in a typical organisation it fires only a handful of times a month once legitimate automation is excluded. It is a realistic first rule for a team with limited tuning capacity.
Yes, with different data. Linux process creation is captured through auditd execve records, eBPF-based sensors, or Sysmon for Linux, and the same reasoning applies: a web server process spawning a shell, a cron job launching an interpreter that downloads and executes code, or a container process starting something outside its image are all lineage anomalies.
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.