The idea
Enumeration is where most of the engagement is won. Nmap's job is to answer three questions in order: what hosts are alive, what ports are open, and what is actually listening on them. I scan in stages rather than one giant command, so each result shapes the next.
How I work through it
I start wide and fast to find open ports, then go deep only on what is open. A full TCP SYN sweep first:
nmap -p- --min-rate 2000 -T4 10.10.10.10 -oA scans/all-tcp
Then version and default-script detection on just the ports that came back:
nmap -sCV -p 22,80,445 10.10.10.10 -oA scans/targeted
UDP is slow but worth a top-ports pass, because SNMP, DNS and TFTP hide there:
nmap -sU --top-ports 100 10.10.10.10 -oA scans/udp
When a service looks interesting I pull the relevant NSE category instead of blasting every script, e.g. --script "smb-enum-* and not brute" for SMB.
Pitfalls I watch for
- Always save output (
-oA). Re-scanning to recover a result I already had is wasted time, and the.gnmapfile greps cleanly for open ports. -sCVon all 65535 ports is painfully slow — find open ports first, then version-scan only those.- A high
--min-ratecan drop packets on a shaky VPN and hide open ports. If results look empty, slow down before concluding the host is closed.
Takeaway
Treat Nmap as a funnel, not a single command: sweep, then focus. The notes that matter later are the service versions, because those are what map to known vulnerabilities in the next phase.