Join our Newsletter — 33% off our NHI Course
Home› Guides› Volatility 3 Guide: Plugins, Symbols and Safe Memory…
Guide Incident Response & Forensics

Volatility 3 Guide: Plugins, Symbols and Safe Memory Analysis

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 8 November 2026 13 min read
On this page

Volatility 3 is the framework most responders reach for when a memory image has to answer what was really running, but it only gives defensible answers when symbols, plugin choice and evidence handling are prepared before the incident. This guide explains how the framework works (translation layers, automagic and symbol tables rather than Volatility 2 profiles), which plugins answer which investigation questions on Windows and Linux, and how to run it safely on sensitive evidence. It goes deeper than the Memory Forensics Guide, which covers when and how to acquire memory, and focuses on the tool itself. It is for DFIR analysts, incident responders and detection engineers. Facts about plugins and options were checked against the project's repository and documentation (framework version 2.28.3 on the develop branch) on 27 October 2026.

This is an independent guide from NHI Mgmt Group, based on the Volatility Foundation's public documentation as of November 2026. It is not sponsored or reviewed by the Volatility Foundation. Product features change, so confirm settings against the Volatility Foundation's current documentation.

Key takeaways

  • There are no profiles any more. Volatility 3 uses JSON symbol tables (ISF files); Windows symbols are usually generated from Microsoft PDBs automatically, while Linux and macOS symbols must match the kernel banner exactly.
  • Plugin names have moved. Malware-hunting plugins now live under windows.malware.* and registry plugins under windows.registry.*; older names such as windows.malfind still run but are marked deprecated.
  • Cross-view comparison finds hiding. List-walking plugins (pslist, svclist) compared with scanning plugins (psscan, svcscan, psxview, svcdiff) are how unlinked processes and services surface.
  • Offline mode and a staged symbol cache matter for evidence. By default the framework may fetch PDBs and remote ISF data; --offline plus pre-staged symbol packs keeps analysis reproducible and inside your network.
  • The licence has a wrapper clause. The Volatility Software License requires publishing source for Additions, which explicitly include wrappers built to run Volatility and parse its results.

What the tool does

Volatility 3 is a Python framework, maintained by the Volatility Foundation, that extracts artefacts from a memory sample independently of the system it came from. It was released in 2019 as a complete rewrite of Volatility 2, requires Python 3.8.0 or later and is published on PyPI as volatility3. The stable branch on GitHub holds the latest stable release; develop is the default branch. The command-line entry point is vol (or vol.py from a clone), with volshell for interactive exploration.

Volatility does not acquire memory. The project documentation points to WinPmem and FTK Imager for Windows and Microsoft AVML for Linux; acquisition method and tool choice are covered in the Memory Forensics Guide.

How it is built

ConceptWhat it means in Volatility 3Why an analyst cares
Translation layersReplace Volatility 2 address spaces; a layer can depend on several lower layers, which is how swap files are integrated (--single-swap-locations)Pagefile or swap evidence can be added to the analysis instead of being ignored
Automagic and stackersAutomatically detects the image format and stacks layers (for example a raw file, then an Intel page-table layer)Most runs need only -f image plugin; when detection fails, verbose logs (-vvv) show where
Symbol tables (ISF)JSON, optionally .json.gz or .json.xz, under the symbols directory or zip packs named after the OSWrong or missing symbols are the most common reason a plugin returns nothing
Static objectsData is read once when an object is constructed and does not change during a runFaster than Volatility 2, but live memory analysis is effectively out of scope
Renderers-r selects quick (default), pretty, csv, json or jsonlJSON lines with -q feeds directly into a SIEM, Timesketch or a notebook

Symbols in practice

  • Windows: the framework reads the PDB name, GUID and age from the image, looks for a matching ISF under symbols/windows, and if none exists downloads the PDB from Microsoft's symbol server, converts it and caches it. pdbconv.py converts a PDB manually when the analysis host is offline.
  • Linux and macOS: symbols are matched on the kernel banner string, which must match exactly, including build details such as compile time and compiler version. Run the banners plugin first, then obtain the matching debug kernel and convert it with dwarf2json. The documentation is candid that the right symbols may not always be obtainable for a given Linux kernel.
  • Symbol packs: the project publishes Windows, Mac and Linux packs with SHA256 checksums. Place the zip files as named in the symbols directory (or point -s at them). The first run builds a cache, which can take time but is only needed once per location.
  • Cache: cached symbol data lives under ~/.cache/volatility3 (or $XDG_CACHE_HOME/volatility3). --clear-cache resets short-term items and --cache-path changes the location.

Defender use cases

The plugins below are grouped by the question an investigator is trying to answer. Names are as published in the current repository; run vol <plugin> -h for options, since plugin options are dynamic and listed after the plugin name.

Windows plugins by investigation question

QuestionPluginsRelated ATT&CK technique
Can the image be read, and which build is it?windows.info(baseline)
What was running, and is anything hidden?windows.pslist, windows.psscan, windows.pstree, windows.malware.psxviewT1014 Rootkit
What did processes run with?windows.cmdline, windows.envars, windows.consoles, windows.cmdscanT1059.001 PowerShell
Is there injected or unbacked code?windows.malware.malfind, windows.malware.ldrmodules, windows.malware.hollowprocesses, windows.malware.processghosting, windows.malware.pebmasqueradeT1055 Process Injection, T1055.012 Process Hollowing, T1620 Reflective Code Loading
Is EDR or ETW being bypassed?windows.etwpatch, windows.malware.unhooked_system_calls, windows.malware.direct_system_calls, windows.malware.indirect_system_callsTA0112 Defense Impairment
Who talked to whom?windows.netscan, windows.netstatT1071.001 Web Protocols
How does it persist?windows.svcscan, windows.malware.svcdiff, windows.registry.scheduled_tasks, windows.registry.printkeyT1543.003 Windows Service, T1053.005 Scheduled Task, T1547.001 Registry Run Keys
Is a kernel component hiding activity?windows.modscan, windows.malware.drivermodule, windows.callbacks, windows.ssdt, windows.timersT1014 Rootkit
Were credentials exposed?windows.registry.hashdump, windows.registry.lsadump, windows.registry.cachedump, windows.malware.skeleton_key_checkT1003.001 LSASS Memory
Which files and executables were present?windows.filescan, windows.dumpfiles, windows.mftscan, windows.shimcachemem, windows.registry.amcache(supporting evidence)
Does a known signature match?windows.vadyarascan with your YARA rules(depends on rule)

Linux plugins by investigation question

QuestionPlugins
Which kernel is this, so I can find symbols?banners
Processes and hidden taskslinux.pslist, linux.psscan, linux.pstree, linux.malware.process_spoofing
Commands typedlinux.bash (history recovered from bash process memory)
Kernel rootkitslinux.malware.check_syscall, linux.malware.check_idt, linux.malware.hidden_modules, linux.malware.modxview, linux.malware.tty_check, linux.malware.keyboard_notifiers
Privilege abuselinux.malware.check_creds (shared credential structures), linux.capabilities, linux.ptrace
Network and packet filteringlinux.sockstat, linux.malware.netfilter, linux.ebpf
Injected code and signatureslinux.malware.malfind, linux.vmayarascan, linux.elfs

A repeatable triage order

  1. Validate the image: run windows.info or banners; if it fails, fix symbols or the image format before anything else.
  2. Baseline processes: pstree, then psscan and psxview to compare views; note parents that do not fit (an Office process spawning a shell, a service host with an odd path).
  3. Command lines and network: cmdline and netscan for the processes of interest; pivot on remote addresses in network and proxy logs.
  4. Code integrity: malfind, ldrmodules, hollowprocesses and the system call plugins; dump suspect regions with --dump into a controlled output directory (-o) for static analysis.
  5. Persistence and kernel: services, scheduled tasks, drivers and callbacks.
  6. Timeline: timeliner combines time-aware plugin output and can write a bodyfile (--create-bodyfile) for merging with disk timelines in Plaso or other tools. See the Forensic Timeline Analysis Guide.

Dumped regions and files are inputs to the Malware Analysis Guide for Defenders workflow, and strings or byte patterns found there become YARA rules for vadyarascan across other images. See the YARA Rules Guide.

Deployment and safe use

Analysis host

  • Install: use a virtual environment and pin a release (pip install volatility3, or a clone of stable). Record the framework version in case notes: plugins and output columns change between releases.
  • Isolation: analyse on a dedicated forensic workstation, not on the affected estate. Memory images may contain live malware pages; Volatility parses rather than executes them, but dumped regions can be executable files.
  • Offline operation: stage symbol packs and verify them against the published SHA256SUMS, then run with --offline so the tool does not search online for JSON files, Windows symbol tables or Linux and macOS banner repositories. Set it permanently in ~/.config/volatility3/vol.json (or %APPDATA%/volatility3/vol.json), remembering that a boolean set true there is not unset by omitting the flag.
  • Plugin directories: -p loads extra plugin paths before the built-in ones, and the documentation warns that all Python code in those directories is evaluated. Only load community plugins you have reviewed.

Evidence handling

  • Work on a copy: hash the acquired image, analyse a verified working copy and keep the original read-only. See the Evidence Handling and Chain of Custody Guide.
  • Record every run: -l writes full logs to a file and --save-config writes a config.json that recreates the environment; keep both with the case. This supports NIST CSF 2.0 RS.AN-06 (investigation actions recorded with integrity and provenance preserved).
  • Treat output as secret: hashdump, lsadump and cachedump output, environment variables and dumped process memory can hold passwords, NTLM hashes, LSA secrets, API keys and session tokens. Store output with the same controls as the image.

Licensing

Volatility 3 is released under the Volatility Software License (VSL) version 1.0. It allows free use, sharing and building on the software, but if you make Additions available to others you must publish their source code. The licence text defines Additions to include software designed to execute Volatility and parse its results, such as a wrapper, but not general shell or execution menu software. Teams embedding Volatility in a commercial or shared triage platform should have the licence reviewed before distribution.

Detection and hardening value

Memory analysis findings should feed back into prevention and detection, not stop at the report.

Finding in memoryFeed it back asFramework reference
Unbacked executable region in a signed processEDR rule on the parent and injection pattern; YARA rule from the dumped bytesNIST SP 800-53 SI-3, SI-4
ETW patching or unhooked ntdll stubsAlert on loss of telemetry from a host; review EDR tamper protectionTA0112 Defense Impairment; see the EDR Tampering and BYOVD Response Playbook
Credential material exposed (hashdump, LSA secrets)Rotate the affected accounts, including service accounts; enable LSA protection where supportedT1003.001; NIST SP 800-53 IR-4
Malicious service, task or Run keyDetection on service and task creation events; hunt the same artefact fleet-wideT1543.003, T1053.005, T1547.001
C2 address in netscanBlock and hunt in proxy and DNS logs; share as an indicatorNIST CSF 2.0 DE.AE-02, DE.AE-03

Where identity fits

Memory is where attackers harvest machine credentials: service account passwords in LSA secrets, cached domain credentials, cloud tokens in process environments and API keys in application heaps. Every memory case should end with a list of identities whose secrets were resident on the host and a rotation decision for each. See the Credential Dumping Guide and the Service Account Security Guide.

Measuring the capability

  • Image-to-first-answer time: minutes from receiving an image to a validated process list; target under 30 minutes for Windows images with staged symbols.
  • Symbol coverage: share of your Linux kernel builds (by banner) with a tested ISF file on the analysis host.
  • Offline success rate: share of exercise images analysed with --offline without missing-symbol failures.
  • Feedback rate: share of memory cases producing at least one new detection, YARA rule or hardening ticket.
  • Credential follow-through: share of exposed identities rotated within the agreed window.

Common failure patterns

  • Empty output blamed on the image: a Linux image analysed with a near-match symbol file produces nothing; the banner must match exactly, so stage symbols for your production kernels in advance.
  • Trusting one plugin: pslist alone misses unlinked processes; always compare with psscan and psxview.
  • Old command references: runbooks written for Volatility 2 (--profile, imageinfo) or for early Volatility 3 plugin names slow responders down; update them to the windows.malware.* and windows.registry.* names.
  • Silent internet dependency: analysis fails or leaks indicators of a case because the workstation fetched PDBs or ISF data during a run; use --offline with staged symbols.
  • Exposed credentials left unrotated: the report notes hashes and secrets found in memory but nobody owns rotation, so the attacker's harvested access survives remediation.

Alternatives

OptionChoose it whenTrade-off
Volatility 3 CLIDeep, scripted analysis of individual images with full control of plugins and outputNeeds Python skills and symbol preparation, especially for Linux
Volatility 2Only for reproducing old case results that depended on its pluginsPython 2 era code and profiles; not the basis for new runbooks
Live response via an endpoint agentFleet-wide questions (is this process or mutex anywhere?) where a full image per host is impracticalQueries a running, possibly compromised, OS; see the Velociraptor Guide
Commercial memory forensics suitesTeams wanting a GUI, vendor support and integrated reportingLicence cost; confirm with the vendor how artefacts are parsed
EDR memory scanningContinuous detection of known injection patternsDetection, not forensic analysis; results are not a substitute for an image

Small teams usually do best with Volatility 3 for occasional deep cases and an incident response retainer for surge work; larger teams build a symbol pipeline for every production kernel and automate the triage order above with JSON lines output. See the Open Source DFIR Tools Landscape Guide for where it sits in a wider toolkit.

Practitioner checklist

  • Pin a Volatility 3 release in a virtual environment on a dedicated forensic workstation and record the version in every case.
  • Download the Windows, Mac and Linux symbol packs, verify SHA256 checksums and pre-build the cache.
  • Generate and test ISF files with dwarf2json for every Linux kernel you run in production.
  • Set --offline as the default in vol.json for case work.
  • Update runbooks to current plugin names (windows.malware.*, windows.registry.*).
  • Compare list-walking and scanning plugins before concluding nothing is hidden.
  • Keep -l logs and --save-config files with case evidence; protect plugin output as secret data.
  • Turn each finding into a detection, YARA rule or credential rotation ticket.
  • Have the VSL reviewed before distributing any wrapper or platform built on Volatility.

Standards and references

Related NHI Mgmt Group resources: Memory Forensics Guide · Open Source DFIR Tools Landscape Guide · Malware Analysis Guide for Defenders · Forensic Timeline Analysis Guide · Credential Dumping Guide

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 8 November 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.