OSINT Fundamentals
Before you touch a target, you read it. Public data does most of the work.
What you will learn
- -Explain what counts as OSINT and what does not
- -Run the intelligence cycle instead of collecting at random
- -Keep your collection legal and low-noise
Half the findings in a real engagement come from data the target published themselves. Job postings that name the exact they run. A LinkedIn profile listing "migrating from on-prem Exchange to M365". A PDF on the marketing site with author metadata pointing at an internal username. None of that touches the target's , and all of it shapes how you attack.
Open Source Intelligence is intelligence built from publicly available sources. Public is the load-bearing word. If getting the data requires a login you were not given, a you guessed, or a you were not authorized to touch, it stopped being OSINT and became something you can be prosecuted for. Reading a company's records is OSINT. Logging into their webmail with creds from a breach dump is not.
Passive vs active, and why the line matters
Passive collection never sends a the target can see as yours. You query third parties: Google's index, transparency , WHOIS registrars, breach databases. The target's logs stay empty.
Active collection touches the target directly. A DNS zone transfer request, a connection to their , an scan. It's still OSINT-adjacent, but now you show up in their access logs with a source IP. Know which side of that line every command sits on, because on a red team the whole point is staying quiet until you choose to be loud.
Run the cycle, do not just collect
Random googling produces a pile of tabs and no direction. Intelligence work is a loop with five steps, and skipping the first and last is the usual amateur mistake.
- Direction: decide what you actually need to answer. "Find valid email addresses for phishing" is a direction. "Learn about the company" is not.
- Collection: gather raw data against that question only.
- Processing: normalize it. Dedupe usernames, resolve names to accounts, sort .
- Analysis: turn data into a . "This subdomain runs an old Citrix gateway" is analysis, not a screenshot.
- Dissemination: write it down where your team can use it, then feed gaps back into step one.
The loop is what keeps a two-week recon phase from becoming a hoarding exercise. Every collected item should trace back to a question you wrote down.
OSINT feels harmless because you never break in. It is still recon against a real organization. Only profile targets you have written to test, and stay inside the your rules of engagement define. A breach-dump password belongs in your report as a finding, never in a login form.
Keep a clean collection habit
Start every engagement with a scoping file: the in-scope domains, IP ranges, and names, plus anything explicitly out of scope. Everything you pull gets checked against it before it goes in your notes.
Timestamp and source every finding. Six weeks later when a asks "how did you know our VPN portal was there", the answer is "certificate transparency log, crt.sh, pulled on this date", not a shrug. Provenance is what separates an intelligence product from a rumor.
You have five recon actions: (a) a WHOIS lookup on the target domain, (b) an nmap SYN scan of their mail server, (c) searching Have I Been Pwned for a target email, (d) browsing the target's public GitHub org, (e) a DNS zone transfer attempt against their nameserver. Sort each into passive or active, and say which one is most likely to appear in the target's own logs.
Which of these stops being OSINT?