Table of Contents

← Return to the Personal Security Course

Technical controls still depend on decisions about enrollment, recovery, approval, and exceptions. Social engineering targets those decisions. A repeatable verification process reduces reliance on memory and pressure.

Some attacks abuse legitimate workflows after persuading a person to authorize an action. Others exploit software, accounts, providers, or combined paths. This module covers the decision and verification layer.

Budget about 30 minutes. Practice callback verification before a pressured request arrives.

What You Will Learn

  • Explain how urgency, authority, familiarity, fear, and helpfulness support a pretext
  • Apply verification which works under pressure rather than only in calm conditions
  • Recognize phishing across email, messaging, calls, and search results
  • Assess voice cloning and video manipulation as a change to what a familiar voice proves
  • Reduce what public records and data brokers expose about you
  • Produce a verified contact list and an agreement with people who might be impersonated
TermMeaning
PretextA fabricated situation which justifies the request being made
VishingVoice phishing, where the attacker calls you rather than messaging
Callback verificationConfirming a request by calling back on a number you already had
Data brokerA company which aggregates and sells personal information
Prompt bombingFlooding approval requests until one is accepted

Why Pretexts Work

Attacks succeed by manipulating decisions, and the levers are consistent.

LeverHow It AppearsCountermeasure
UrgencyAn account closing, a payment due, a limited windowTreat speed pressure as the signal itself
AuthorityA bank, an employer, a government agency, or IT supportVerify independently before acting
FamiliarityA known name, a real colleague, a matching voiceConfirm out of band rather than in the same channel
Fear of consequenceLegal action, account loss, disciplinary outcomeSlow down deliberately
HelpfulnessA reasonable-sounding request for small assistanceNotice you are being recruited, not attacked

The most counterintuitive point: these attacks often do not feel like attacks. A request which arrives from a known name, with plausible context, and a reasonable ask produces no alarm, which is precisely the design.

Notice the emotion rather than the content. Urgency, obligation, and embarrassment are the outputs a pretext engineer aims for, and recognizing the feeling is faster than analyzing the message.

Verification Which Survives Pressure

Verification you devise in the moment is unreliable. Decide in advance.

The Callback Rule

Hang up and call back on a number you already had, from a card, a statement, or a saved contact. Never use a number supplied by the caller, even if it appears to match.

This habit interrupts many vishing, bank impersonation, and payment-change attempts when the stored number and receiving party are trustworthy.

Agreed Verification With People You Know

Arrange a signal with family members and close contacts before it is needed. Options include a phrase, a question with an answer only you both know, or an agreed second channel.

This is worth doing with anyone who might plausibly be impersonated to you, particularly parents, adult children, and anyone with access to shared finances.

Out-of-Band Confirmation for Money

Any payment instruction which arrives by message requires confirmation through a separate channel. This applies to requests from colleagues, contractors, and family.

SituationWhat to Do
New bank details from a supplierCall the supplier on a previously known number
Urgent payment from an executiveConfirm through a known internal channel, not a reply
Family emergency requestCall the family member directly on a stored number
Crypto or gift card requestTreat as fraud until proven otherwise. Both are irreversible and favored by attackers

Warning: gift-card or cryptocurrency payment demands and unexpected urgent wire changes are strong fraud signals. Some legitimate wires are time sensitive, so stop and verify the recipient, amount, and instructions through a trusted separate channel.

Recognizing Phishing

Phishing has moved past badly spelled email, so the old tells are less useful than the structural ones.

Structural SignalWhy It Matters
Domain differs from the real onepaypa1.com and paypal-secure.net are attacker-owned. Read the domain right to left, before the first slash
Unexpected attachment or linkEven from a known contact, since accounts get compromised
Unexpected request for credentials or a codeDo not disclose an unsolicited authentication code. Verify any legitimate workflow through the official site or number
Channel escalationSomeone moves the conversation to messaging to avoid email controls
Payment detail changeA classic business email compromise step
Lookalike page after a redirectThe padlock does not indicate authenticity, as Module 2 established

Two habits which reduce exposure:

  • Enter addresses manually or use a bookmark for anything involving credentials
  • Treat unsolicited pop-ups as untrusted, including ones warning about viruses or offering support. Close the page without using its phone number, link, or download. Check the browser and operating-system security tools directly

A web page might imitate a system warning. Close the tab or browser, then check the device through known operating-system or security controls.

What a Familiar Voice Proves

Voice cloning changed the threat model for verification by recognition.

Voice-cloning quality depends on the model, sample quality, language, length, and playback conditions. Public audio increases impersonation risk. A familiar voice or face alone does not prove identity for a consequential request.

The practical consequence: identity confirmation by voice is no longer sufficient for anything consequential. Replace it with out-of-band verification, which does not depend on recognition at all.

Old AssumptionCurrent Reality
“I recognize their voice”Generated or replayed audio might sound familiar
“Their face was on the video call”Real-time face manipulation exists and is improving
“They knew our internal details”Details might be public, breached, purchased, or supplied by an insider

This does not mean refusing every request. It means confirming consequential ones through a channel the attacker does not control, which is the callback rule applied to people as well as institutions.

Reducing Your Public Exposure

Public information often improves a pretext, though attackers also use random, breached, purchased, or insider data. Reducing exposed details removes selected starting material.

Data Brokers

  • Submit opt-out requests to the major people-search sites, and expect to repeat the process periodically
  • Verify the broker and minimize disclosure. Some opt-outs require identity verification. Read the privacy policy, submit only required fields, and use a safer channel when available
  • Track what you submitted and recheck on a schedule because broker records and opt-out duration differ

Domain Registration

Registration privacy or proxy services replace selected public registration fields where registry rules, registrar support, domain type, law, and policy permit. Check cost, disclosure rules, contact forwarding, and current public output.

This connects to Module 3. A domain used for email aliasing holds your identity in its registration record unless privacy protection is enabled.

Address and Identity Separation

PracticeBenefit
Forwarding address or parcel lockerReduces direct home-address disclosure for eligible deliveries. Billing, account, carrier, and regulated records might still retain the address
Opt out of marketing listsReduces selected marketing use and sharing under the provider’s current terms
Review social media privacy on a scheduleTerms, settings, defaults, and product behavior change over time

Our Phishing Awareness Checklist covers the organizational training version of this module.

Diagram showing a social engineering attempt interrupted by out-of-band verification, where the recipient calls back on a previously known number rather than trusting the inbound contact

The Human Record

HUMAN ASPECT RECORD
Verified contact list (numbers I already had):
  Bank:                  ______________________
  Card issuer:           ______________________
  Employer IT:           ______________________
  Family members:        ______________________

Agreed verification:
  Person:                ______________________  signal: __________
  Person:                ______________________  signal: __________

Money rules I follow:
  [ ] Never act on payment details received by message alone
  [ ] Never pay by gift card, crypto, or wire under time pressure
  [ ] Always call back on a number I already had
  [ ] Confirm any executive request through a known internal channel

Public exposure reduction:
  Whois privacy enabled:  yes / no / n/a
  Broker opt-outs filed:  ______________________
  Forwarding address:     yes / no
  Social privacy reviewed: date __________

Practice and Verify

Ask a friend to read two harmless scenarios. Do not use a real password, code, account, or payment.

Scenario A: a caller claims to be your bank and asks you to read a code. Say, “I do not share unsolicited authentication codes. I will call the number on my card.” End the call and locate the saved number.

Scenario B: a family message reports an emergency from a new number. Say, “I will verify through our saved contact or another trusted person.” Use the agreed signal only through the planned channel.

Scenario C: a legitimate support flow tells you to enter a code on the official site you opened yourself. Explain why entry on the official site differs from reading the code to an inbound caller.

Expected Result: the learner pauses, names the request, refuses channel-provided contact details, and reaches an independent verification route.

If the saved number fails, use the number on a physical card, statement, official app, or official website reached independently. If no route exists, make no payment or account change. Record the failed path and find a verified route after the exercise.

REDACTED EXAMPLE
Scenario A:            refused code, found card number
Scenario B:            reached family through stored contact
Scenario C:            entered code only on self-opened official site
Pressure response:     pause, name request, verify elsewhere
Failed route:          old employer number removed

Pass checklist

  • The learner completes all three scenarios without disclosing a secret
  • At least two official callback routes are stored
  • A failed callback has a safe stop condition
  • Broker opt-outs use verified sites and minimum required disclosure
  • Answers explain why each response is safe

Primary reference, checked 2026-10-10: FTC impersonator scam guidance . Use the official site or stored number for each real verification.

Next Steps

  1. Continue to Module 12 and cover the physical layer: Physical Security
  2. Save verified contact numbers for your bank, card issuers, and employer IT today
  3. Agree a verification signal with the two people most likely to be impersonated to you
  4. Enable Whois privacy if you own a personal domain
  5. File opt-outs with the major data brokers, and diarize a repeat in six months
  6. Read the organizational counterpart: Phishing Awareness Checklist