C-Login

Password authentication persists because it is universally deployable, not because it is secure. Proposals that replace it — federated identity, WebAuthn/passkeys — require the relying party to adopt a new protocol, which is why the long tail of the web still runs on (userid, password) two decades after the problems were catalogued. We describe C-Login, a credential-wrapping layer that removes the reusable password from the user’s side of the interaction while requiring no change whatsoever to the relying party. The user’s password is sealed at enrolment into a C-Password: a self-contained ciphertext carrying its own access policy — a user-chosen unlock secret (PIN, passcode, or delivered one-time code) and a validity window. At each login the user presents a freshly minted C-Password; a broker service verifies the policy through the LE cryptographic engine, recovers the password, and invokes the relying party’s existing login API. From the relying party’s perspective nothing has changed. The construction’s security rests on the underlying Ladhe hash-based scheme [11], whose hardness reduces to SHA-256 preimage resistance and therefore admits no known quantum shortcut. We give an explicit threat model, derive the offline-attack bound as a function of unlock- secret entropy and window duration, and report measurements from a working implementation. The validity window is cryptographically bound to decryption rather than checked as policy, so an expired artifact is unrecoverable regardless of adversary compute; we derive the resulting offline bound min(2^(H−1), R·d). We are deliberate about what this does not achieve: it is not phishing-resistant in the sense WebAuthn is, and the bridge deployment necessarily reconstructs the password transiently at the broker.