Skip to content
Tutorials Center
My learning

Topics

Lesson 24 of 38 ยท Exploitation Concepts

Social Engineering Assessments

Learn how authorized phishing, vishing and physical social-engineering tests are planned, what rules keep them ethical, and how to analyze emails and measure results.

Social engineering manipulates people instead of software: a convincing email, a phone call from "IT support", a stranger following an employee through a door. It works because people want to be helpful and are busy. Attackers know that tricking one employee is often easier than defeating a firewall, so many organizations ask penetration testers to measure how resilient their staff and processes really are.

This is the area where ethics and legal care matter most, because the "targets" are real people. A good assessment teaches and improves processes. It never humiliates individuals, collects real personal data it does not need, or targets anyone outside the agreed scope. All examples here use invented messages and domains.

Forms of social engineering tests

Type What it tests Typical safe design
Phishing (email) Whether users click, enter credentials or report Link to a harmless landing page that records only "clicked" or "submitted something", never the real password
Spear phishing Targeted, researched messages Limited to approved staff groups, content approved in advance
Vishing (phone) Help-desk and staff verification procedures Scripted pretext, no real account changes without approval
Smishing (SMS) Mobile habits Rarely used; needs explicit approval
Physical Door control, tailgating, badges, clean desk Written authorization letter ("get out of jail" letter) carried at all times
USB drop Policy on unknown devices Benign devices that only phone home a "plugged in" marker

Always carry written authorization for physical tests, with a client contact who can confirm it by phone. Without it, a test can end with a real arrest. Never involve HR data, family members, personal accounts or anything beyond the signed scope.

Planning rules that protect people

  1. Get explicit approval for pretexts, target groups and dates. Executives and the security team may need a separate decision.
  2. Do not collect real passwords. Record that a form was submitted, not what was typed. If a tool captures entries, discard or hash them immediately.
  3. Avoid sensitive pretexts such as layoffs, bonuses, health or family emergencies. They cause real distress and legal trouble.
  4. Report in aggregate. Results should say "18% clicked, 6% reported" and name departments only when the client agreed. Never publish a shame list.
  5. Provide a reporting path and training afterwards. The most valuable metric is the report rate, not the click rate.
  6. Coordinate with a trusted contact who can stop the test, and tell the mail provider or SOC if filtering needs allow-listing.
flowchart LR
    A["Agree scope and rules"] --> B["Design pretext and landing page"]
    B --> C["Approval from client sponsor"]
    C --> D["Send in small waves"]
    D --> E["Track clicks, entries, reports"]
    E --> F["Debrief and training"]
    F --> G["Aggregate report and retest"]

An ethical phishing exercise from agreement to retest.

Reading a suspicious email

Testers also teach people to spot phishing, so you should know the signals well. Typical red flags are an unexpected request, urgency, a sender domain that is almost right, a link whose real destination differs from the text, and failures in the email authentication checks:

  • SPF checks the sending server is allowed for the domain.
  • DKIM checks a cryptographic signature on the message.
  • DMARC tells receivers what to do when SPF/DKIM do not align with the visible From domain.

The stand-in below analyzes an invented email header block. It compares the From and Return-Path domains, reads the authentication results and checks for look-alike domains with an edit-distance function.

Python Analyze a suspicious email header

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The sender domain examp1e-corp.com uses a digit 1 in place of the letter l, which is a homoglyph trick; edit distance catches it. Defenders use the same idea when they register look-alike domains to block them.

Measuring results sensibly

Click rate alone is a weak and often misleading number. Better indicators are the share who reported the message, how fast the first report arrived, and whether anyone submitted credentials. Compare waves over time, not departments against each other.

Python Summarize a phishing exercise

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

A falling click rate is good, but a rising report rate with a very short time to first report is better: it means the security team now learns about a campaign within minutes, which is what limits real damage.

Defenses to recommend

  • Email filtering with SPF, DKIM and DMARC at enforcement, plus banners for external mail.
  • A one-click "report phishing" button and a team that answers reports quickly.
  • Phishing-resistant MFA (FIDO2 keys, passkeys) so stolen passwords alone are useless.
  • Verification procedures for payment changes and password resets by phone, with call-back to known numbers.
  • Badge discipline, visitor escorts and a culture where challenging a stranger is welcomed.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In an ethical phishing exercise, what should happen to passwords typed into the landing page?

    Recording the submission event proves the point without handling real credentials.
  2. 2. What is the most valuable success metric for a phishing awareness program?

    Fast reporting lets the security team contain a real campaign, which clicks alone cannot show.
  3. 3. Why should a tester carry a signed authorization letter on a physical test?

    Security staff or police may respond to an apparent intruder. The letter and a contact who can confirm it protect you.
  4. 4. Which header finding suggests a spoofed or look-alike sender?

    Mismatched domains and failed authentication checks are classic phishing indicators.

Summary

  • Social engineering tests measure people and processes; they must be approved, scoped and respectful.
  • Never collect real passwords, use distressing pretexts, or name individuals in public results.
  • Learn the email signals: mismatched domains, failed SPF/DKIM/DMARC, urgency and look-alike names.
  • Measure reports and response time, not just clicks, and follow up with training.
  • Recommend filtering, easy reporting, phishing-resistant MFA and call-back verification.

Next: Post-exploitation rules and cleanup