[reference] — from the roadmap.sh cyber-security roadmap. The deep-investigation end of the blue team: reconstructing what happened from digital evidence, and safely dissecting the malware involved. Together this is often called DFIR (Digital Forensics & Incident Response).
Digital forensics
Digital forensics is recovering and analyzing digital evidence to reconstruct events — after an incident, for legal proceedings, or to understand an attack. Its defining constraint is evidence integrity: everything must be done so the findings hold up (legally and technically).
The cardinal rules:
- Order of volatility — collect the most ephemeral evidence first: CPU registers/cache → RAM → network state → disk → backups. RAM (with running processes, encryption keys, fileless malware) vanishes on power-off, so capture it before touching disk.
- Work on copies, never originals — take a bit-for-bit forensic image of a disk and analyze the copy, so the original evidence is unaltered.
- Chain of custody — document who handled evidence, when, and how, unbroken — or it’s inadmissible in court.
- Hashing for integrity — hash evidence at collection (hashing) and re-verify, proving it wasn’t altered.
The two main domains:
- Disk forensics — recover files (including deleted ones), examine file systems, timestamps (the “timeline”), browser history, logs. Tools: Autopsy/The Sleuth Kit, FTK, EnCase.
- Memory forensics — analyze a RAM capture for running processes, network connections, injected code, and secrets. Tool: Volatility. Essential for modern malware that never touches disk.
Forensics answers the incident’s key questions: how did they get in, what did they touch, what did they take, are they still here, and how do we stop it recurring.
Malware analysis
When an incident involves malware, you analyze it to understand its capabilities, build IOCs/detections, and scope the damage. Do this only in an isolated, disposable environment — a sandbox VM with no network path to anything real, snapshots to revert. Running malware to study it is exactly the kind of action that needs strict isolation.
Two approaches:
- Static analysis — examine the file without running it: hashes, strings, headers, embedded resources, disassembly (reverse engineering). Safe, but obfuscation/packing limits what you see.
- Dynamic analysis — run it in a sandbox and observe behavior: files created, registry/config changes, processes spawned, network callbacks (C2). Reveals what it actually does, but sophisticated malware detects sandboxes and hides.
Automated sandboxes (Cuckoo, and online services like Any.Run, Joe Sandbox, Hybrid Analysis, VirusTotal) detonate a sample and report behavior quickly — the practical first step before manual reverse engineering. The output feeds detection: a new hash to block, a C2 domain to sinkhole, an ATT&CK technique to hunt for.
The safety and scope caveat
This is legitimate defensive work, but it’s inherently handling dangerous material and (for forensics) potentially sensitive/legal evidence. The professional disciplines — isolation for malware, chain of custody and work-on-copies for evidence — aren’t bureaucracy; they’re what keep the analysis from causing a second incident or invalidating the whole investigation.
Related
- Incident Response — the process forensics serves
- Malware & Exploits — what’s being analyzed
- Hashing & Integrity — proving evidence is unaltered
Software engineering
Infrastructure & security
Computer science
Hardware & engineering