Ethical hacking isn’t improvisation — it follows a structured methodology so findings are thorough, repeatable, and defensible afterward (both technically and legally). The classic phases, echoed across most formal frameworks (PTES, OSSTMM, NIST SP 800-115):
The phases
1. Rules of Engagement / Authorization -> before anything else happens at all
2. Reconnaissance -> gather information about the target
3. Scanning & Enumeration -> map what's actually reachable and running
4. Exploitation -> attempt to leverage a weakness to gain access
5. Post-Exploitation -> determine actual impact (privilege escalation, lateral movement, data access)
6. Reporting -> document findings, severity, and remediation — the actual deliverable
Each phase is its own note in this folder: reconnaissance, scanning-and-enumeration, exploitation-concepts, post-exploitation-and-reporting — and rules-of-engagement-and-legal for the authorization step that has to happen before phase 2 even begins.
Why order matters
Each phase’s output feeds the next: reconnaissance identifies what to scan, scanning identifies what’s plausibly exploitable, exploitation (if successful) opens up post-exploitation, and everything gets synthesized into the report. Skipping straight to exploitation without proper reconnaissance/scanning means testing blind — more likely to miss things, more likely to cause unintended disruption on a live system, and harder to explain afterward why a particular attempt was made at all.
Types of engagement
- Black-box — the tester starts with no inside knowledge of the target, simulating a genuine external attacker’s starting position. Most realistic in terms of what an actual attacker faces, but slower and can miss internal-only weaknesses.
- White-box — the tester is given full information up front (source code, architecture diagrams, credentials) — more thorough and efficient, common for application security reviews where the goal is finding as many issues as possible rather than simulating a realistic attacker.
- Grey-box — partial information given (e.g. a standard user account, but no source code) — a common, practical middle ground reflecting what a malicious insider or a successfully phished attacker would actually have.
Red team vs. penetration test — a scope distinction
A penetration test typically has a defined scope and timeframe, aiming to find as many vulnerabilities as reasonably possible within it. A red team engagement is usually more narrowly goal-oriented (“can you reach this specific system/data without being detected”) and often explicitly tests the defensive (blue) team’s detection and response capability, not just whether a vulnerability exists. Confusing the two leads to mismatched expectations about what a report should contain — a red team report is as much about detection gaps as it is about the specific vulnerabilities used to get in.
Gotchas
- Every phase past reconnaissance assumes explicit written authorization is already in place — without it, identical technical actions are illegal, regardless of intent.
- A methodology is a checklist for thoroughness, not a rigid script — real engagements loop back (a post-exploitation finding often triggers renewed reconnaissance/scanning against a newly discovered internal system) rather than executing the six phases strictly once, in order.
Related
- reconnaissance
- scanning-and-enumeration
- exploitation-concepts
- post-exploitation-and-reporting
- rules-of-engagement-and-legal
Software engineering
Infrastructure & security
Computer science
Hardware & engineering