Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Dynamic analysis runs the sample on purpose, in an environment built to record everything it touches, and reads the behaviour instead of guessing at the code.

What Is Dynamic Malware Analysis?
Dynamic malware analysis means executing a suspicious file deliberately, inside an isolated and heavily instrumented environment, and recording what it does. Where static examination reads the sample at rest, dynamic analysis watches it work: the processes it spawns, the files it writes, the registry keys it sets, and the servers it calls. Packing, encryption, and obfuscation all have to come undone at runtime, so behaviour is visible even when the code is not.
A useful lab records four layers at once.
Automated sandboxes such as CAPE, Joe Sandbox, and Any.Run wrap all of this into a report in a few minutes. Manual work with a debugger follows when the automated run stops short, which it often does.
The lab has to be assumed hostile from the first second. Run virtual machines on an isolated segment with no route to production, never on a bridged adapter. Disable shared folders, clipboard sharing, and guest additions, both because they are escape routes and because their presence is the first thing malware checks. Revert to a snapshot after every run. Handle samples in password-protected archives so that no scanner on the way in or out quarantines your evidence, and keep an eye on the legal position: a sample containing stolen data is still stolen data, and the handling obligations do not pause because it arrived as malware.
A sample that runs for ninety seconds and exits cleanly is usually not benign; it has decided it is being watched. The checks are well documented and easy to anticipate:
The counter-measures are mundane and effective: age the image, add browsing history and documents, give it four cores and a realistic disk, patch the obvious platform artefacts, accelerate sleep calls in the sandbox, and join a decoy domain where the sample expects one. If a sample still refuses, that refusal is a finding worth recording.
The point is not the report but what goes into defence. Expect a concrete list of indicators of compromise: hashes, dropped file paths, mutex names, scheduled tasks, service names, registry run keys, domains, and IP addresses. Expect the behaviour mapped to MITRE ATT&CK techniques so that detection gaps become visible. Expect detection content: YARA rules against memory for the unpacked family, and endpoint rules for the behaviour, which survive repacking in a way that file hashes do not.
Dynamic and static analysis are not alternatives. Running the sample tells you what it did on this occasion, under these conditions; reading the code tells you what it could do on another. Serious work alternates between them, and the hand-off, usually a memory dump of unpacked code, is where most of the value appears.
At Privia Security, our investigation and incident response services include sample detonation and reverse engineering as part of containment, not as an academic exercise afterwards.
It is reasonably safe with care, and careless with none. Isolate the network, remove shared folders and clipboard integration, disable guest additions, and revert from a snapshot after every run. Escapes from hypervisors are rare but real, so do not run analysis on a machine that holds anything you value, and never on a host connected to the corporate domain.
Because they checked and did not like what they saw. Low core counts, small disks, hypervisor drivers, absent user activity, and short run times all indicate an automated environment. Some samples simply sleep past the analysis window; others decrypt their payload using a key derived from the intended victim’s domain, so they cannot run anywhere else even if they wanted to.
They answer different questions. Endpoint detection tells you something suspicious happened on a host and contains it. Analysis of the sample tells you what it was designed to do, what it would have done next, which data it targeted, and what to search the rest of the estate for. For a significant incident, that difference usually determines the scope of the response.
For commodity samples, often yes, and the results are fast. The caution is disclosure: files submitted to public services may be shared with other subscribers, so a document containing client data, an internal tool, or a targeted sample naming your organisation should not go there. Attackers also monitor public submissions, and an upload can tell them their campaign has been noticed.
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.