zeliha-infosec-journey

Attack Types — Offensive Identification Cheat Sheet

How to look at recon output and decide which category of attack actually applies to a target — the decision layer that sits between “I scanned the box” and “I picked the right technique.”

All techniques below are for use in authorized environments only — personal labs, CTFs, and engagements covered by written authorization (RoE).


1. How to Think About This

Every attack type has a signal that tells you it’s on the table. Recon isn’t just “what’s open” — it’s “which attack categories does this open up.” Work through the categories below in order; most real engagements chain 2–3 of them together (e.g., web app → credentials → AD → privesc).


2. Network-Based Attacks

Signal you’ll see in recon Attack type it points to Go-to tools
Same broadcast domain / VLAN access ARP spoofing, MITM arpspoof, bettercap, ettercap
LLMNR/NBT-NS enabled (common on Windows networks) Responder-style credential capture Responder, Inveigh
No 802.1X / open switch ports VLAN hopping, rogue device yersinia, manual trunk negotiation
DHCP server reachable, no snooping DHCP starvation / rogue DHCP Yersinia, dhcpstarv
Unencrypted protocols in traffic capture (FTP, Telnet, HTTP, SNMP v1/2c) Credential sniffing Wireshark, tcpdump, net-creds

Tell: you’re on the same L2 segment as the target, or traffic is traversing a switch you can influence.


3. Web Application Attacks

Signal Attack type Go-to tools
User input reflected in response without encoding Reflected/Stored XSS Burp Repeater, manual payloads
Input concatenated into a DB query (error messages, odd behavior on ') SQL Injection sqlmap, Burp
App fetches a URL you can influence (webhooks, “import from URL”) SSRF Burp Collaborator, manual
XML input accepted (SOAP, some upload flows) XXE Burp, manual DTD payloads
File upload with weak extension/type validation Malicious file upload → RCE Burp, manual
Shell metacharacters reach a system call Command Injection manual, commix
Predictable object references in URLs/params (?id=101) IDOR Burp, manual enumeration
Old CMS/plugin version fingerprinted Known-CVE exploitation wpscan, searchsploit, Metasploit

Tell: the target exposes a web app with user-controllable input — this is almost always the first thing to rule in/out.


4. Authentication Attacks

Signal Attack type Go-to tools
Login form / SSH / RDP / SMB with no lockout policy Brute-force Hydra, medusa
Known breach-list overlap likely (reused corporate emails) Credential stuffing Hydra w/ combo lists
One password sprayed across many accounts to dodge lockouts Password spraying kerbrute, CrackMapExec
Captured hash (NTLM, Kerberos ticket, /etc/shadow) Offline cracking hashcat, John the Ripper
SMB/AD reachable with valid creds Pass-the-hash / pass-the-ticket CrackMapExec, Impacket

Tell: you have a login surface or a captured credential artifact (hash, ticket) and need to turn it into access.


5. Active Directory / Windows Domain Attacks

Signal Attack type Go-to tools
Domain-joined host, valid low-priv creds AD enumeration BloodHound, PowerView, CrackMapExec
Service account with SPN set Kerberoasting Rubeus, GetUserSPNs.py (Impacket)
Account with no Kerberos pre-auth required AS-REP Roasting GetNPUsers.py (Impacket)
Misconfigured ACLs found via BloodHound ACL abuse (e.g., GenericAll, WriteDACL) BloodHound, Impacket
Unconstrained/constrained delegation enabled Delegation abuse Rubeus, Impacket

Tell: you’re inside a Windows domain with at least one set of valid (even low-privilege) credentials.


6. Malware-Delivery Style Attacks

Signal Attack type Go-to tools
Client-side execution needed (no direct network path) Phishing payload delivery msfvenom, custom macros/HTA
AV/EDR present on target Payload obfuscation / evasion needed Veil, manual encoding, LOLBins
Persistence required across reboots Backdoor/implant placement scheduled tasks, registry run keys, cron

Tell: direct remote exploitation isn’t viable and you need the user (or a scheduled process) to execute something for you.


7. Social Engineering

Signal Attack type Go-to approach
Employee emails/org chart discoverable (OSINT) Phishing / spear-phishing SET, GoPhish, crafted pretext
Physical access in scope Tailgating, badge cloning, USB drop Physical engagement rules apply
Help desk / support line in scope Vishing / pretexting Scripted pretext calls

Tell: the human is the weakest link in scope, and technical controls alone won’t get you in.


8. Denial of Service

Signal Attack type Note
Resource-intensive endpoint (search, PDF export, regex input) Application-layer DoS Almost always out of scope on live/prod systems — confirm RoE explicitly
No rate limiting on auth endpoints Account lockout DoS Same caveat

Tell: DoS is rarely authorized outside dedicated test environments — treat any DoS-shaped finding as a reporting item unless RoE explicitly permits testing it live.


9. Wireless

Signal Attack type Go-to tools
WPA2-PSK network in scope Handshake capture + offline crack aircrack-ng, hashcat
WPS enabled WPS PIN attack reaver, bully
Open/guest SSID reachable Rogue AP / evil twin airbase-ng, hostapd-wpe

10. Cloud / Misconfiguration

Signal Attack type Go-to tools
S3 bucket / blob storage referenced in app Public bucket exposure aws s3 ls (unauthenticated), cloud_enum
IAM role attached to a compromised EC2/VM Metadata service abuse → credential theft curl 169.254.169.254, pacu
Overly permissive IAM policy discovered Privilege escalation in cloud pacu, ScoutSuite

11. Quick Decision Flow

  1. What surface is exposed? (network / web / auth / domain / human / wireless / cloud) — pick the matching section above.
  2. What artifact do I already have? (a shell, a hash, a set of creds, a ticket) — that determines whether you’re still in initial access or already in post-exploitation territory.
  3. What does the RoE actually permit? — DoS, physical, and social engineering categories are frequently excluded; always confirm before acting.

Companion to attack-types-detection-cheatsheet-professional.md in Blue-Team — same categories, viewed from the defender’s side.

Prepared for eJPT-aligned offensive work. Use only in authorized environments.