CSRP — Cyber Security Responder & Practitioner
Lesson 26: OSINT and Public Intelligence Gathering
Student Lab Guide
Cyber School · BDO Ziv Haft
CSRP LESSON 26 — STUDENT LAB GUIDE
Module 4 — Offensive Cyber & Vulnerability Analysis · Lesson 2 of 8 · Course Lesson 26 of 40 Title: OSINT and Public Intelligence Gathering Duration: 4 hours · 3 labs + 1 Challenge Badge on Challenge pass: [Passive-Recon-Analyst]
⚠️ Ethical-use contract (read this first)
Everything you do today is for defensive purposes only. You operate inside a lab environment, against the fictional target Northwind Academy — a deliberately-fictional company that does not exist on the real internet. You will not run dig, whois, or exiftool against any real organization during this lesson.
Outside this room and outside written authorization: – Active probing (port scans, vulnerability scans — taught next lesson) of any system you do not own can violate the US Computer Fraud and Abuse Act, the UK Computer Misuse Act, Israel’s Computer Law, and equivalent legislation in every country we operate in. – Even passive techniques (Whois lookups, LinkedIn scraping, certificate-transparency-log queries at scale) can violate platform terms-of-service and trigger rate-limits or account suspension. – Real-world penetration testing requires written authorization, a defined scope, and a clear rules-of-engagement document. No exceptions.
When in doubt — ask before you act.
Why you need this as a cyber professional
OSINT is the skill that opens every cyber-attacker’s day, and it is also the skill that opens every mature defender’s quarter. Most 2025-2026 incidents start with attackers reading an employee’s LinkedIn, pulling forgotten subdomains from certificate-transparency logs, harvesting internal hostnames from a leaked PDF, or finding an exposed API key in a public GitHub commit. The technical work is mundane — and that is exactly why it succeeds. Defenders who can’t run the OSINT toolkit fluently can’t see their own organization the way an attacker sees it.
In any operational role you take after CSRP — SOC analyst, threat hunter, incident responder, red teamer, security engineer — OSINT will be a daily-or-weekly tool. Threat hunters use it to enrich indicators with public context. SOC analysts use it to triage suspicious external IPs and domains. IR teams use it to scope a breach by enumerating the affected organization’s full attack surface. Security engineers use it quarterly to audit and shrink the public exposure their own org publishes. Red teamers use it for hours before a single packet leaves their machine.
L26 builds the fluency. By 4:00 PM today you can profile a fictional company by reading its public-facing surface (Lab 26.1), query its DNS and Whois data with the standard Linux toolkit (Lab 26.2), extract sensitive metadata from its published documents with exiftool (Lab 26.3), and assemble cross-source clues into a verified flag (Challenge — Public Clues).
Where you are
L25 (yesterday): Cyber Kill Chain framework — you mapped attack timelines to 7 phases and earned [Kill-Chain-Mapper].
L26 (today): Phase 1 in operational detail. OSINT — the public-source recon every attacker runs first, and every defender SHOULD run on themselves.
L27 (next): Active scanning with Nmap — you start generating traffic the target sees.
Module 4 progress by end of today: 2 of 8 lessons. Course progress: 26 of 40.
Today’s environment
Primary: – A browser inside the lab pointed at http://northwind-academy.local/ — a local copy of a fictional company website. Staff bio pages, public documents, press releases, careers pages.
Secondary: – A Linux workstation (Ubuntu/Kali) with dig, nslookup, whois, and exiftool installed. A local DNS resolver answers queries about northwind-academy.local as if it were a real domain. – Sample documents pre-staged at /labs/L26/samples/ (≥10 files: PDFs, DOCX, JPGs).
You never query the real internet from this lab. Every domain you investigate today resolves inside the lab network only.
Quick-win (~5 min) · Extract ONE Recon Finding
Goal: prove to yourself that public information leaks faster than people realize.
The instructor shows a screenshot of a fictional job posting. Your task: name ONE OSINT signal that an attacker would extract from this single posting.
Anticipated signals: – Tech-stack mentions (specific tools/products) → tells the attacker what defenses to expect and what evasion to plan. – Cloud provider → narrows the attack-surface enumeration. – Role + location → physical and organizational footprint. – SOC maturity → “Tier-2” implies a larger, more sophisticated SOC.
Debrief: one job posting just leaked five infrastructure facts. Multiply by 10 postings + 50 LinkedIn profiles + 100 employee bios — the dossier writes itself.
Lab 26.1 — Passive Recon on a Fictional Company (~50 min · Core)
Topic coverage: OSINT definition, passive vs active boundary, public-source enumeration, target dossier construction.
Why you need this as a cyber professional: the first thing every attacker does — and the first thing every quarterly defender’s-self-OSINT-pass does — is enumerate what the organization publishes about itself. Names, emails, technology, vendors, partnerships, projects. Without this fluency you cannot see your own organization as an attacker sees it, and you cannot prioritize the right hardening work.
Pre-flight
Browser open to http://northwind-academy.local/.
Worksheet template open (digital or printed) — the OSINT Profile Dossier.
Steps
Land on the homepage. Note the org’s headline language, claimed industry, and any partner / customer logos visible. Record on the worksheet under “Org positioning.”
Navigate to the “About” or “Team” pages. Catalog every employee name visible. For each: name, role, email if shown, photo if relevant.
What to expect: between 6 and 15 staff members listed; titles ranging from CEO to L1 engineer.
Why it matters: names + roles + email format = the input to spear-phishing pretexting.
Identify the email-format pattern. Look at every email shown on staff pages. Extract the pattern: is it first.last@northwind-academy.local? firstinitial.lastname@? last.first@? Record the inferred pattern.
Why it matters: once you know the pattern, an attacker can generate the email for ANY employee whose name they harvest from LinkedIn — even one whose email wasn’t published.
Scan press releases / news / blog pages. Catalog technology mentions, vendor names, project names, customer references.
What to expect: casual mentions like “we run our SOC on Splunk” or “deployed our new endpoint on Windows 11.”
Why it matters: each mention is a fact about the defender stack that informs which exploits / techniques an attacker would consider.
Open the Careers / Jobs page. Read every open posting. Catalog every tool, platform, and vendor named.
Why it matters: job descriptions are the single richest OSINT source for technology stack. They are written to attract candidates with specific skills — which means they enumerate the exact tools the org uses.
List public documents linked from the site. Whitepapers, datasheets, annual reports, regulatory filings. Save the filenames for Lab 26.3.
Why it matters: every published document is a metadata leak waiting to happen.
Build the OSINT profile dossier. Use the worksheet to consolidate findings into a single-page profile: org positioning, staff inventory, email pattern, tech stack, public document list, two or three observations about exposure.
Quick-reference card — what to extract from a public website
What you find
What an attacker does with it
Employee names + roles
Targeted phishing, LinkedIn pivoting
Email-format pattern
Email enumeration for any name
Tech-stack mentions
Defense fingerprint → exploit selection
Vendor / partner names
Supply-chain attack vectors
Project / product names
Internal jargon for pretexting credibility
Press release dates
Timeline of org changes (M&A, leadership moves)
Document downloads
Metadata extraction (Lab 26.3)
Success criteria
≥6 staff names recorded with roles
Email format pattern stated explicitly (e.g., first.last@northwind-academy.local)
≥3 technology-stack mentions captured
≥5 public documents listed for Lab 26.3 follow-up
One-page OSINT Profile Dossier complete
Common mistakes
Listing employee names without recording the email format → you have raw data but no enumeration pattern.
Skimming the Careers page → you miss the richest tech-stack signal in the entire site.
Treating one mention as confirmation → cross-reference: a tool named in TWO places is far more likely to be in production.
Lab 26.2 — DNS and Whois Investigation (~25 min · Core)
Topic coverage: DNS as OSINT, record types (A / MX / TXT / SPF / DMARC), Whois data, exposed-service inventory.
Why you need this as a cyber professional: DNS is the OSINT source defenders MUST audit on themselves quarterly. Forgotten subdomains, dangling CNAMEs pointing to abandoned cloud resources, missing SPF/DMARC, stale name-server records — these are how organizations get hit by subdomain takeovers, spoofed email campaigns, and supply-chain attacks. You cannot harden what you cannot see, and DNS is the second place you look (after the public website).
Pre-flight
Terminal open on the Linux workstation.
Confirm tools available: which dig nslookup whois exiftool should return paths for each.
Steps
Query A records — the IPv4 addresses associated with the domain:
dig northwind-academy.local A
What to expect: one or more IPs in the answer section.
Why it matters: the IPs reveal whether the org is on-prem, in cloud (which provider — IP ranges identify AWS / Azure / GCP), or behind a CDN. Geolocation of the IP narrows physical footprint.
Query MX records — mail-server entries:
dig northwind-academy.local MX
What to expect: one or more priority + hostname pairs, often pointing at the org’s mail vendor (Google Workspace, Microsoft 365, Mimecast, Proofpoint).
Why it matters: the mail vendor reveals the defender’s mail security stack and informs phishing-evasion choices.
Query TXT records — text records, where SPF / DMARC / DKIM live:
dig northwind-academy.local TXT
What to expect: SPF entries starting with v=spf1, DMARC entries starting with v=DMARC1, sometimes DKIM-related entries.
Why it matters: the presence and policy of SPF/DMARC tells defenders whether spoofed email gets through. ~all (softfail) vs -all (hardfail) is the difference between an attacker’s email landing in inbox or being rejected.
Identify SPF / DMARC posture explicitly. Parse the actual content of the TXT records:
SPF: which mail servers are authorized to send (look for include: directives)? What’s the failure policy (-all strict, ~all soft, ?all neutral, +all permissive-bad)?
DMARC: what’s the policy (p=none, p=quarantine, p=reject)? Where do reports go (rua=mailto:…)?
Why it matters: a publicly-discoverable p=none DMARC is a green light for spoofers.
Review Whois data:
whois northwind-academy.local
What to expect: registrar name, registration date, expiry date, name servers, sometimes contact email (depending on privacy settings).
Why it matters: registration patterns can reveal corporate ownership relationships (subsidiaries, brand registrations). Stale registrar info can hint at organizational drift.
Document exposed services. On the worksheet, build a table:
Record type
Value(s) found
Defender concern
A
(the IPs)
Cloud provider visibility? Geolocation?
MX
(mail hosts)
Mail vendor revealed
TXT/SPF
(SPF content)
Policy strict / soft / missing?
TXT/DMARC
(DMARC content)
Policy strict / quarantine / none?
Whois
(registrar + age)
Anything stale or unexpected?
Optional — try ANY for everything at once:
dig +short northwind-academy.local ANY
Useful for quick triage; many production DNS servers refuse ANY queries due to amplification-attack concerns, but the lab fixture allows it.
Quick-reference card — DNS toolkit
Command
Purpose
dig <domain> A
IPv4 address records
dig <domain> MX
Mail-server records
dig <domain> TXT
Text records (SPF, DMARC, verification tokens)
dig +short <domain> ANY
Quick all-records dump
dig @<dns_server> <domain>
Query a specific DNS server
nslookup -type=MX <domain>
Older alternative to dig
whois <domain>
Registration metadata
Success criteria
A / MX / TXT records all queried and parsed
SPF posture stated (strict / soft / missing)
DMARC posture stated (reject / quarantine / none / missing)
Whois data reviewed; one observation recorded
Worksheet exposed-services table complete
Common mistakes
Querying records but not parsing the content (e.g., noting “SPF exists” without checking -all vs ~all).
Confusing nslookup and dig output formats — they look different but answer the same questions.
Missing a hidden subdomain — the lab fixture includes project-aurora.northwind-academy.local as a quiet signal.
Lab 26.3 — Metadata Exposure (~20 min · Core)
Topic coverage: document metadata, exiftool usage, sensitive field identification, metadata-removal controls.
Why you need this as a cyber professional: every Word document and every PDF your marketing team publishes is a potential metadata leak. Author names, last-modified-by names, internal file paths, software versions (which may have CVEs), tracked changes, and embedded comments all ride with the document. As a defender, you are the person who needs to know what’s leaking — and who builds the pipeline that strips it before publication.
Pre-flight
Terminal open on the Linux workstation.
Confirm sample files staged:
ls /labs/L26/samples/
You should see ≥10 files including PDFs, DOCX, and JPGs.
Steps
Open one sample PDF in exiftool:
cd /labs/L26/samples/
exiftool sample01.pdf
What to expect: a wall of key-value pairs covering Author, Title, Subject, Keywords, Creator, Producer, CreationDate, ModifyDate, possibly Comments and embedded tags.
Repeat for a DOCX:
exiftool meeting_notes_q4.docx
What to expect: Author, Last-Modified-By, Application name + version, Title, Company, Lines/Words/Pages counts, sometimes Track-Changes history.
Repeat for a JPG (photo):
exiftool office_photo.jpg
What to expect: Camera Make + Model, Software (which photo-editing tool), DateTime, sometimes GPS coordinates (Latitude, Longitude — if the photo was taken on a phone with GPS enabled).
Run exiftool against ALL samples in one shot:
exiftool *.pdf *.docx *.jpg
Use the output to scan for any “INTERNAL:” or “Confidential” tags. Modern collaboration tools sometimes auto-add classification tags that ride into the published document.
Mark sensitive metadata. On the worksheet, for each file:
File
Author
Internal path?
GPS?
Sensitive tag?
Overall risk
sample01.pdf
High / Med / Low
meeting_notes_q4.docx
office_photo.jpg
Recommend metadata-removal controls. For each file family:
PDF: qpdf –linearize –object-streams=generate <input> <output> (strips most metadata), or exiftool -all= -overwrite_original <file>. For PDFs created from Office, also “Inspect Document” → “Remove” before saving.
DOCX: Office “Document Inspector” → “Remove All” before publication. CI-pipeline alternative: convert DOCX to PDF via a metadata-stripping pipeline.
JPG: exiftool -all= -overwrite_original <file> strips EXIF including GPS. For phone photos especially.
Write a short exposure report. 5-8 sentences summarizing: how many files were inspected, how many had sensitive metadata, which fields were the worst offenders, and which 2-3 controls would prevent the leak going forward.
Quick-reference card — exiftool cheatsheet
Command
Purpose
exiftool <file>
Dump all metadata fields
exiftool -a -G1 <file>
Group by metadata family (PDF, XMP, EXIF, etc.)
exiftool -GPS* <file>
Show only GPS fields
exiftool -Author -Creator <file>
Show only specific fields
exiftool -all= -overwrite_original <file>
Strip ALL metadata (irreversible)
exiftool -CommonTags <file>
Show only the high-signal fields
Success criteria
≥5 files inspected with exiftool
Sensitive-field table complete for each file
≥1 file flagged for “INTERNAL:” or equivalent sensitive tag
≥2 metadata-removal controls recommended in the exposure report
Common mistakes
Skimming exiftool output and missing the “INTERNAL: Aurora Project Plan” tag — it’s there, in plain text, on at least one PDF.
Treating “Author: J. Smith” as harmless. If J. Smith is your CISO, you’ve just disclosed the author of an “official” document — pretexting material for the next attacker.
Recommending “be careful” instead of specific controls. Defenders write pipelines, not aspirations.
Challenge — Public Clues (~25 min · Independent · BADGE)
Your brief
You are investigating Northwind Academy’s public exposure. Four clues are scattered across the website, DNS records, and document metadata. Your task: assemble them per the validator’s instructions and submit the flag.
The four clues to find
The email-format pattern of Northwind Academy staff (from the staff bio pages on the website).
A specific hidden subdomain named after an internal project (from DNS records — try ANY or TXT and look for unusual subdomains).
A sensitive metadata tag on a published PDF (from running exiftool against the samples).
The hidden project name that appears across all three of the above.
What you have
Browser pointed at http://northwind-academy.local/
Linux workstation with dig, whois, exiftool
~10 sample documents at /labs/L26/samples/
What you produce
/labs/L26/work/challenge_public_clues.md — your investigation notes with the 4 elements + reasoning
The ULC{…} flag, derived per the validator’s combination rules
Tiered hints
T=8 min — Concept hint: Each clue comes from a different source. One on the website, one in DNS, one in metadata, one assembled by cross-referencing the first three.
T=15 min — Tool hint: For DNS, try dig +short northwind-academy.local ANY and dig northwind-academy.local TXT. For metadata, exiftool *.pdf and scan for “INTERNAL:” tags. The hidden subdomain has a project-themed name.
T=20 min — Near-spoiler: The email format is on the staff bio pages. The hidden subdomain is named after a project. The PDF metadata contains a tag like “INTERNAL: <Project> Plan.” The same project name appears in all three. Combine per the validator’s format.
Success criteria
4 clues correctly identified and documented in the investigation note
ULC{…} flag submitted and validated
Badge [Passive-Recon-Analyst] issued automatically on pass
Common mistakes
Looking for all 4 clues on the website. Use all three tools.
Submitting the project name in the wrong case or with the wrong separators. The validator is strict — follow the format.
Spending 25 minutes on Lab 26.1 territory and ignoring DNS + metadata. The Challenge requires all three.
Quick-reference card — today’s full toolkit
Tool
Use
Browser
Navigate the fictional target’s website; harvest names, email patterns, tech mentions
dig
Query DNS records (A, MX, TXT, SPF, DMARC, ANY)
nslookup
Older alternative to dig for quick DNS lookups
whois
Domain registration metadata
exiftool
Read and strip metadata from PDFs, DOCX, JPGs
Wrap-up
Today you built a passive OSINT toolkit. You can now profile a public website, query DNS records, extract document metadata, and combine cross-source clues. You earned [Passive-Recon-Analyst].
The defender’s takeaway: run today’s checklist on your own organization quarterly. Discover what attackers will find. Drive the harden-the-attack-surface roadmap from those findings.
Next lesson: L27 — Network and Port Scanning with Nmap. Today you stayed passive (target never saw you). Next time you generate traffic the target sees — and the ethical-use contract intensifies accordingly.
Lesson Progress
0% Complete