The Codex · Cybersecurity · Protocol CSP-003
The most sophisticated account security is worth nothing if the device used to access it is unencrypted, unpatched, or physically unsecured. This protocol defines the minimum configuration required on every device used for any work of significance — no exceptions for convenience.
The majority of significant security compromises begin at the endpoint — the laptop, the phone, the tablet — rather than at the network or the service. Credential theft via malware, physical access to an unlocked device, and exploitation of unpatched vulnerabilities are consistently among the most common and most consequential attack vectors. Hardening accounts and communications while leaving devices themselves inadequately secured is a fundamental gap in posture.
A device that is lost or stolen presents a direct risk to everything stored on it and everything it can access. An unencrypted drive can be read without the operating system login. An unlocked screen provides immediate access to everything open. An unpatched vulnerability can be exploited remotely before the owner is aware a patch exists. These are not hypothetical failures. They are routine ones.
Full-disk encryption ensures that if your physical storage medium — the SSD in your laptop, the chip in your phone — is removed and read directly, the contents are unreadable without the decryption key. Without encryption, physical access to the hardware is sufficient to access everything on it, regardless of your operating system password.
Enabling the operating system's native encryption tool is the required approach. Third-party full-disk encryption solutions introduce unnecessary complexity and potential key escrow risks. FileVault (macOS), BitLocker (Windows), and the default device encryption on iOS and modern Android are well-audited, hardware-accelerated, and sufficient. After enabling, verify encryption status explicitly — some systems report encryption as enabled before the initial encryption process has completed.
Security patches exist because vulnerabilities have been discovered and, in most cases, are already being actively exploited before the patch is released. The window between a vulnerability becoming publicly known and being patched on your device is the window during which you are exposed. The longer that window, the greater the risk.
The seven-day patch window is not a suggestion. It is the outer limit of acceptable delay for routine updates. For critical patches — those tagged as addressing zero-days or actively exploited vulnerabilities — the window is twenty-four hours. Automatic updates should be enabled wherever they do not create operational disruption. Where they are disabled for stability reasons, a manual patch review process must replace them.
Mobile devices are simultaneously the most frequently carried and the most frequently lost devices in most people's possession. They have access to email, authentication apps, financial accounts, and sensitive communications. They are also carried into environments — restaurants, transport, events — where loss or theft is far more likely than at a fixed workstation.
The remote wipe requirement is non-negotiable. Enabling it takes minutes. Discovering that it was not enabled after a device is stolen is not recoverable. Test the remote wipe process on a non-production device if possible. Understand what data is wiped, what data is backed up to a cloud service, and whether that cloud backup itself is adequately secured.