rhaist/marq / skills/physical-human/social-engineering-defense
social-engineering-defense
VerifiedCovers defending against phishing, vishing, smishing, pretexting and BEC: out-of-band verification for payments and credentials, caller identity verification, reporting culture, and DMARC/SPF/DKIM and MFA backstops. Use when the user asks how to stop CEO fraud or invoice fraud, verify help-desk callers, prevent deepfake voice scams, or harden against social engineering.
Social-engineering defense
Social engineering attacks the human, so the defense is process + technical
backstops, not "be more careful." In the AI/deepfake era, content-based
detection ("does this look fake?") is losing: shift defense to process
enforcement that holds regardless of how convincing the lure is. This is the
defensive pair to marq's passive people-OSINT tools; for the
program/culture side see load_skill security-awareness.
Threat taxonomy (quick reference)
- Phishing: mass email lure → credentials/malware/payment.
- Spear-phishing / whaling: targeted, recon-driven, often impersonating an exec.
- Vishing: phone/voice, now with cloned-voice deepfakes; 41% of CISOs report a deepfake social-engineering incident on an employee audio call, 36% on video, in the past year (Gartner, Sep 2026). verify: https://www.gartner.com/en/newsroom/press-releases/2026-09-22-gartner-survey-finds-41-percent-of-cisos-reported-at-least-one-social-engineering-incident-involving-a-deepfake-in-the-past-12-months
- Smishing: SMS/iMessage/WhatsApp lures, often "package"/"toll"/"boss needs you."
- Pretexting: fabricated scenario to extract info or access (IT support, auditor, new vendor).
- BEC: compromised or look-alike business email → fraudulent wire/payroll/invoice change.
- verify (AI-era trends): https://hoxhunt.com/blog/business-email-compromise-statistics
The core control: out-of-band verification
The single highest-value defense for money and access. Make it mandatory policy, not judgment:
- Any payment instruction, new payee, bank-detail change, or payroll change → verify by calling a number you already have on file (vendor master / internal directory), never a number from the email/text/caller. Callback, don't call-forward.
- Credential/MFA/account changes (reset, device enrollment, "I'm locked out") → verify the requester through a second channel before acting.
- Dual authorization above a defined threshold; mandatory waiting period for first-time or changed payees.
- Treat urgency + secrecy + channel-switch as the red flag; that's the BEC signature, regardless of how legitimate the message reads.
- verify: https://www.cisa.gov/ (BEC guidance) and FBI IC3 BEC advisories.
Caller / identity verification (anti-vishing, anti-pretext)
- Help desk: verify identity with something not publicly knowable and not on the badge (e.g. ticket-driven callback to the number in HR records, manager confirmation): never DOB/employee-ID/last-4 alone, all of which leak.
- Establish a duress / verification phrase or internal callback protocol for high-stakes requests; deepfake voice/video can now pass "it sounded like them."
- Default posture: a real exec under real pressure will tolerate a 2-minute verification. Anyone who won't is the tell.
- Vendors/IT impersonation: confirm via the relationship's known contact, not contact details supplied in the request.
Reporting culture (the scalable detector)
- One person reporting protects everyone who already clicked: reporting is the human SOC sensor. Make it one-click (Report Phish button) and blameless.
- Measure report rate and time-to-report, not click rate alone
(
load_skill security-awarenessfor the metrics rationale). - Close the loop: reports feed triage/IR; tell reporters what happened so the behavior reinforces. Silence kills reporting.
- Wire reports to the SOC →
load_skill incident-response-leadershipon confirmed compromise. Triage a reported message withload_skill phishing-email-analysis; a confirmed mailbox/BEC takeover goes toload_skill bec-email-investigation.
Technical backstops
- DMARC at p=reject (with aligned SPF + DKIM): stops domain spoofing and
look-alike-from-your-domain. CISA BOD 18-01 mandates p=reject for US federal
domains; PCI DSS v4.0.1 req. 5.4.1 makes anti-phishing controls (DMARC/SPF/DKIM
named) mandatory for card-data orgs since Mar 31, 2025. Rollout to p=reject and
the Gmail/Yahoo/Microsoft bulk-sender rules:
load_skill email-security. Limit: DMARC does not stop a real compromised account or display-name spoofing from a different domain; layer the rest. verify: https://dmarcreport.com/blog/business-email-compromise-bec-scams-take-new-dimension-with-multi-stage-attacks/ - Phishing-resistant MFA (FIDO2/passkeys > push > SMS) on all email and remote access; it defeats credential phish and limits BEC account takeover.
- Mail hygiene: external-sender banners, look-alike/cousin-domain detection, newly-registered-domain blocking, attachment sandboxing, link rewriting.
- Lookalike-domain monitoring + takedown; register obvious typo-squats.
- Conditional access / impossible-travel + session-token protection to catch post-phish logins.
- Least privilege + segmentation so a single compromised user isn't game-over.
Decision rule (frontline)
Money or access requested + urgency or channel-switch or secrecy → stop, verify out-of-band on a known-good number, before acting. No exceptions for seniority. Process beats content detection every time.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 196k | Updated | Claude CodeCodex | Verified | Install |