Social engineering

Attackers persuade people rather than break machines, using six levers

The targetAuthority“This is the CEO. I need this done.”Urgency“Within the hour, or the deal is lost.”Intimidation“Your account will be suspended.”Social proof“Everyone in finance has already done it.”Familiarity“Hi Sam, great to meet you at the expo!”Scarcity“Only three places left at this price.”Social engineering persuades a person to do the attacker’s work, using the same levers every time.Authority and urgency are the two most common: a senior name and a deadline leave no time to check. Every leveraims to make the target act before thinking. The defence is a habit, not a tool: when a message pushes you tohurry or to skip a step, that pressure is the signal to slow down and verify by another channel.

Social engineering uses authority, urgency, intimidation, social proof, familiarity and scarcity to make people act.

Phishing

A phishing email gives itself away in seven places

From:Microsoft 365 <alerts@m365-secure-mail.co>1Reply-To:helpdesk.reset@protonmail.example2Subject:URGENT: mailbox deleted in 24 hours3Dear user,4Your mailbox has exceeded its limit. Verify yourpassword now to avoid permanent deletion:https://outlook.office.com/verify5📎 Mailbox_Restore.html671The sender’s name is not its addressa look-alike domain, not microsoft.com2Replies go somewhere elsea free mailbox the attacker reads3Urgency and a threata deadline to stop you checking4A generic greetingthe real service knows your name5The link lieshover: it goes to login-0ffice.example6An HTML attachmentopens a fake sign-in page offline7It asks for your passwordno real service ever doesA phishing email gives itself away in seven places, and the link’s real destination is the surest one.Display names are free text; anyone can write “Microsoft 365”. Check the address, hover over links beforeclicking, and treat any request for a password, a payment change or a gift card as hostile until confirmed.When in doubt, do not reply: open the service yourself from a bookmark, or report the message.

Phishing tells: a look-alike sender, a different reply-to, urgency, a generic greeting, a lying link, an odd attachment, a request for secrets.

Email authentication · Going deeper

SPF, DKIM and DMARC: how mail servers check a sender

SPF lists allowed servers, DKIM signs the message, DMARC tells receivers what to do when both fail.

Three DNS records that make spoofing harder

  • SPF (Sender Policy Framework) a DNS record listing the servers allowed to send mail for a domain. Mail from anywhere else fails.
  • DKIM (DomainKeys Identified Mail) the sending server signs each message with a private key; receivers check it with the public key published in DNS. A changed message fails.
  • DMARC a policy that says what to do when SPF and DKIM do not line up with the visible From address: none (just report), quarantine or reject.

What they do not stop

They prove a message came from the domain it claims. They do nothing about look-alike domains (m365-secure-mail.co passes its own SPF perfectly) or about a real mailbox that has been taken over. That is why people still need the tells, and why BEC from a compromised supplier sails through.

Since 2024 Google and Yahoo require bulk senders to publish SPF, DKIM and DMARC. Set DMARC to p=reject once reports show your legitimate mail passes.

Phishing channels

The same trick arrives by email, phone, text, QR code and in person

EmailPhishing: anyoneSpear phishing: youWhaling: executivesVoice and textVishing: a callSmishing: SMSDeepfake voiceWebWatering holeTyposquattingBrand impersonationQR and mediaQuishing: QR codesUSB dropFake invoicesIn personTailgatingImpersonating staffShoulder surfingSpear phishing is aimed at one person with research behind it; whaling targets the most senior.The same trick arrives by every channel, and the less people expect it there, the better it works.Email filters have improved, so attackers move to channels with weaker checks: phone calls, SMS, QR codes onposters and parking meters, and messaging apps. A watering hole compromises a site the target already trusts;typosquatting registers a near-miss domain. In person, a confident visitor follows staff through a door.

Spear phishing, whaling, vishing, smishing, quishing, watering hole and tailgating are one idea on different channels.

Business email compromise

One email in a real thread can redirect a payment, with no malware

A real email thread, with the attacker reading alongSupplier“Invoice 2291 attached, due on the 30th.”You“Thanks, scheduled for payment.”Supplier?“Update: we changed banks. Please use the new account below.”From the supplier’s real mailbox, broken into weeks earlier, or a look-alike domain.$2.77 billionlost to BEC reported to the FBI in 2024,from 21,442 complaintsThe control that stops it:any change to bank details is confirmedby phone, on a number already on file,never one given in the email.Business email compromise needs no malware: one convincing message redirects a real payment.Attackers take over a mailbox, watch real conversations, then send one email at the moment a payment is due. Themessage comes from a familiar address in a familiar thread, so filters and people trust it. Losses are far largerthan from ransomware in the FBI’s figures. Verification outside email is the only reliable defence.

Business email compromise (BEC) hijacks or imitates a trusted mailbox to change where money goes.

Pretexting

A believable story and a helpful help desk can hand over the keys

MGM Resorts, September 2023, as reportedPublic profilename, role, managerCalls the help desk“I’m locked out, new phone”Gets a resetpassword and MFASigns in as adminreaches the identity systemRansomwarehotels and casinos stopAbout $100 millionhit to quarterly earnings, from a phone call.Stops it:no resets on the caller’s word; verify by video with ID, a manager’s approval, or a call-back.Pretexting invents a believable story, and an IT help desk built to be helpful is its favourite target.A pretext is a scenario with a role: a locked-out employee, an auditor, a new supplier. The attackers took namesfrom a professional networking site and asked the help desk to reset access. MGM said the attack would cost about$100 million in the quarter. Help desks need a verification rule that no story can talk past.

Pretexting invents a role and a reason. Help desks need identity verification that no story can bypass.

Pretexting · Going deeper · 1 of 2

Push prompt after push prompt, until one is approved

Phone at midnightApprove sign-in? 23:41DeniedApprove sign-in? 23:42DeniedApprove sign-in? 23:44DeniedApprove sign-in? 23:47DeniedApprove sign-in? 23:55DeniedApprove sign-in? 00:02DeniedApprove sign-in? 00:14ApprovedUber, September 2022A contractor’s password, bought by the attacker, thenpush after push until one was approved.Stops it:✓Number matching: type the code shown on the sign-in screen✓Limit prompts, and alert on repeated denials✓Phishing-resistant MFA: a passkey or security key, no prompt to acceptMFA fatigue floods a person with sign-in prompts until one is approved to make it stop.Simple “approve or deny” prompts ask a tired person a yes-or-no question with no context. Number matching makesthe user copy a code from the screen they are actually using, so a stray prompt cannot be approved by reflex.Microsoft turned it on for all Authenticator users in May 2023. The strongest answer removes the prompt entirely.

MFA fatigue (prompt bombing) floods approvals. Number matching and phishing-resistant MFA defeat it.

Pretexting · Going deeper · 2 of 2

A proxy page can pass on your one-time code and keep your session

Employeepassword + one-time codethe real page, relayedPhishing proxylogin-0ffice.examplecopies everythingreplays themsession cookieReal serviceThe proxy keeps the session cookie: signed in, MFA already passed.A passkey signs for the real site’s name only, so on login-0ffice.example it simply does not work.A phishing proxy sits between victim and site, relays one-time codes and steals the signed-in session.Codes typed by a person can be typed into the wrong site. Kits that do this (adversary-in-the-middle phishing) aresold as a service. Phishing-resistant authenticators, FIDO2 security keys and passkeys, bind the sign-in to thereal domain, so a look-alike site gets nothing usable. NIST names them the only phishing-resistant kind.

Adversary-in-the-middle phishing relays credentials and codes, then steals the session cookie. Passkeys stop it.

Security awareness

Someone will click; what matters is how fast someone reports

Seconds from opening a phishing emailclick: 21 senter data: +28 sMedians from Verizon’s 2024 report: under a minute, so prevention cannot rely on people alone.Share of people who report a phishing testWithout recent training5%After recent training21%One button to reportin the mail clientThe first reportpulls it from every inboxNo blame for clickingso people report fastSomeone will always click. Training pays off when it makes reporting fast, easy and blame-free.A phishing campaign is answered within minutes, so the goal is not zero clicks but early reports: one report letsthe security team remove the message from every mailbox. Simulations measure progress; recent training raisedreporting fourfold in Verizon’s 2025 data. Awareness is a control category of its own in ISC2 CC.

Security awareness training aims at fast reporting. One report lets the team pull a campaign from every inbox.

Awareness programme · Going deeper

What a security awareness programme covers

Awareness covers phishing, passwords, data handling, physical security and reporting, measured over time.

Topics in the exam objectives

  • Recognising phishing and social engineering with simulated campaigns that teach at the moment of the click.
  • Anomalous behaviour risky, unexpected or unintentional actions colleagues should report.
  • Password and MFA hygiene password managers, never reusing passwords, never approving unexpected prompts.
  • Handling data classification labels, where sensitive data may and may not go, removable media.
  • Physical and remote work badges, tailgating, clean desks, public Wi-Fi, screen privacy.
  • Policies and handbooks acceptable use, and how to report an incident.

Measuring it

Report ratethe share of test phishes reported: the number to raise
Time to reportminutes from delivery to first report
Repeat clickerswho needs more targeted help
ISC2 CC 2026 places security awareness in the Security Governance domain: training, education and measuring its effectiveness.

Lesson complete

You finished “Social engineering”

Nicely done.

Open in a new tab