The Codex · Communications · Protocol COM-002
Email is the primary delivery mechanism for phishing, business email compromise, credential theft, and malware. It is also the communication channel most people use with the least scrutiny, because familiarity has bred complacency. This protocol defines the habits and configurations that close the most common attack vectors — not by making email safe, but by making exploitation measurably harder.
Email is the most widely used business communication channel and the one with the weakest built-in verification. Anyone can send an email claiming to be anyone. The protocol has no reliable sender authentication at the display level. Spam filters and domain authentication (SPF, DKIM, DMARC) help but do not eliminate the problem — attackers who have compromised a legitimate account, registered a convincing lookalike domain, or exploited a poorly configured mail server can bypass most automated filters.
The result is that email requires active scrutiny from the recipient in a way that most other systems do not. A door lock either works or it does not. Email security is, in significant part, a human process — and human processes fail when people are busy, tired, or operating under artificially created time pressure.
Email clients display a sender name alongside the sending address, and most people read the name rather than the address. Display names are set by the sender and have no verification. An email can display any name — "Apple Support", "Your Accountant", "Managing Director [Name]" — while arriving from a completely unrelated address.
The countermeasure is to configure your email client to display the full sending address at all times, not just the display name. On mobile clients, this typically requires tapping the sender name to expand it. Making this a reflex — check the address, not the name, before acting on any email — eliminates the display name spoofing vector entirely.
Links and attachments are the two primary delivery mechanisms for phishing and malware respectively. Links direct you to sites designed to capture credentials or install software. Attachments deliver malware directly to your device, often through document-embedded macros or exploits in file format parsers.
The hover-to-preview approach for links works for desktop clients but not mobile. On mobile, long-pressing a link typically previews the destination URL before following it. The key check is whether the destination domain is the one you expect — not whether the path after the domain looks plausible. A convincing path on a malicious domain is still a malicious domain.
For attachments, the relevant question is not whether the sender is known — it is whether this specific attachment from this sender at this time was expected. Attackers frequently compromise email accounts and send malicious attachments from them, using the account owner's legitimate identity and relationships to bypass suspicion.
Most phishing and fraud emails include an artificial urgency element — "your account will be suspended", "action required within 24 hours", "this payment is overdue". The urgency is designed to suppress the deliberation that would identify the fraud. A recipient who pauses to verify is a failed target. A recipient who acts immediately under pressure is a successful one.
The correct response to urgency in an email is not to act faster. It is to slow down and verify through an independent channel. If the urgency is genuine, a telephone call to confirm will take two minutes. If the urgency is manufactured, the telephone call reveals the fraud. The cost of verification is the same either way; the cost of not verifying in the case of fraud is not.