[reference] — from the roadmap.sh cyber-security roadmap. Educational/defensive: how credential attacks, malware, and low-level exploits work, so you can defend against them.
Password / credential attacks
Credentials are the keys to the kingdom, so attacking them is perennial:
| Attack | How it works | Defense |
|---|---|---|
| Brute force | try every possible password | rate limiting, lockout, MFA, slow hashes |
| Dictionary attack | try common/leaked passwords | ban common passwords, MFA |
| Password spraying | try one common password against many accounts (evades per-account lockout) | detect the pattern, MFA, no lockout-evasion gaps |
| Credential stuffing | replay username/password pairs leaked from other breaches | MFA, breached-password checks, don’t reuse passwords |
| Rainbow tables | precomputed hash→password lookup | salting defeats them (hashing) |
| Pass-the-Hash | reuse a stolen password hash directly (no cracking needed) on Windows/NTLM | least privilege, credential guard, network segmentation |
| Kerberoasting | request and offline-crack service-account tickets in Active Directory | strong service passwords, managed service accounts |
The two defenses that dominate: MFA (a stolen password alone is useless) and proper password storage (slow, salted hashing like bcrypt/argon2 — secure authentication). Most “hacks” in the news are stolen/reused credentials, not zero-days.
Malware types
Malware = malicious software. Categorized by behavior:
- Virus — attaches to a file/program; spreads when that runs.
- Worm — self-propagating across a network with no user action (WannaCry).
- Trojan — disguised as legitimate software; a backdoor once run.
- Ransomware — encrypts the victim’s data and demands payment. The dominant modern threat; defense is backups (recovery), segmentation, and patching.
- Spyware / keylogger — covertly collects info/keystrokes.
- Rootkit — hides deep in the OS (often kernel-level) to maintain stealthy persistence; hard to detect and remove.
- Botnet — a network of compromised machines controlled by an attacker (for DDoS, spam, mining).
- RAT (remote access trojan) — full remote control of the victim machine.
- Fileless malware — runs in memory / abuses legitimate tools (living-off-the-land, LOLBins like PowerShell) to evade file-based antivirus.
Delivery is usually phishing or a drive-by/exploit; defense is layered — EDR/antivirus, patching, least privilege, and sandboxed analysis of suspect files.
Exploitation of software flaws
How attackers turn a bug into control:
- Buffer overflow — writing more data than a buffer holds, overwriting adjacent memory (including return addresses) to hijack execution. The classic memory-corruption exploit, the reason memory-safe languages and mitigations (ASLR, DEP/NX, stack canaries) exist. C/C++ are most exposed.
- Privilege escalation — turning limited access into higher (admin/root) access: vertical (user → admin, via a kernel/service bug or misconfiguration) or horizontal (accessing another peer user’s data). Post-exploitation’s core goal (post-exploitation).
- Zero-day — a vulnerability unknown to the vendor (no patch exists), so defenders have “zero days” to prepare. Valuable and dangerous; defense-in-depth and behavior-based detection are the only real hedge until a patch ships.
- Exploit frameworks — Metasploit and similar package known exploits for authorized testing (common tools). CVEs (the public vulnerability catalog) + CVSS scores drive patch prioritization.
The defensive summary
- Credentials: MFA + proper hashing + no reuse defeats most attacks.
- Malware: patch, least privilege, EDR, and backups (the ransomware answer).
- Exploits: patch management, memory-safe languages where possible, and the OS mitigations (ASLR/DEP) — plus detection for the zero-days you can’t patch.
Related
- Secure Authentication — MFA and password hashing
- Forensics & Malware Analysis — investigating malware safely
- Risk & Frameworks — patch/vulnerability management
Software engineering
Infrastructure & security
Computer science
Hardware & engineering