The information-gathering phase — building a picture of the target before touching it directly (or before touching it in any way that could be noticed). What gets found here shapes every later phase: you can’t scan or exploit what you don’t know exists.
Passive vs active reconnaissance
- Passive — gathering information without directly interacting with the target’s systems at all: public records, search engines, social media, DNS records, job postings (which often reveal internal tech stack), leaked credential databases. Effectively undetectable to the target, since no traffic touches their infrastructure.
- Active — directly interacting with the target’s systems to gather information (pinging hosts, resolving DNS, browsing the target’s own website) — still typically low-risk and often within normal scope, but detectable in logs, unlike passive recon.
Most engagements start entirely passive and only move to active reconnaissance once broad passive groundwork is done.
OSINT — Open Source Intelligence
The practice of gathering information from publicly available sources. Common categories:
- Domain/DNS information — WHOIS records, DNS records (which can reveal mail servers, subdomains, hosting providers).
- Organizational information — company structure, employee names/roles (useful for understanding likely usernames, and for social engineering awareness testing specifically authorized in scope).
- Technical footprint — job postings mentioning specific technologies, code repositories, exposed documentation, cached pages showing since-removed content.
- Breach data — checking whether an organization’s email domains appear in known public breach databases, which can indicate reused/weak credentials in circulation.
# illustrative, conceptual examples of passive recon commands — not exploit tooling
whois example.com
dig example.com ANY
nslookup -type=MX example.com
Why reconnaissance quality determines the rest of the engagement
A thorough recon phase can reveal an organization’s entire external attack surface — every subdomain, every exposed service, every piece of software in use — often before a single scan is run. Skipping or rushing this phase means later scanning/exploitation is working from an incomplete picture, and it’s the phase most likely to reveal an easy, low-effort weakness (an exposed admin panel, a misconfigured cloud storage bucket) that a rushed engagement would miss entirely by jumping straight to scanning.
Social engineering awareness (as a specifically scoped activity)
Some engagements explicitly include testing human vulnerability to social engineering — phishing simulations, pretexting phone calls — but only when explicitly authorized in scope (see rules-of-engagement-and-legal), since it involves interacting with real people who haven’t individually consented the way a system owner has. This is a distinct, separately-scoped activity from technical reconnaissance, not a default part of every engagement.
Gotchas
- Even passive reconnaissance should stay within the agreed scope — gathering information about systems or subsidiaries not covered by the authorization can create legal exposure despite involving no direct interaction with those systems.
- Information gathered in recon can go stale — a subdomain found early in a long engagement may be decommissioned by the time exploitation begins; recon is often revisited throughout an engagement, not treated as a one-time, upfront step.
Related
Software engineering
Infrastructure & security
Computer science
Hardware & engineering