A coherence-governed security layer.
QLX is a cryptographic client that executes only under admissible conditions. It consumes CohOS decisions and QAKD attestation to govern when keys are used, when sessions are established, and when encryption actions may occur.
Security fails when used at the wrong time.
Many compromises are not broken cryptography. They are correct cryptography executed on inadmissible state: replayed telemetry, substituted entropy, drifted instrumentation, or unbounded timing uncertainty.
QLX is governance plus disciplined cryptography.
QLX is designed to be evaluated as a cipher family, while the operational advantage comes from admissibility governance. Cryptanalysis and operational gating are separable. QCT does not claim formal superiority to AES or NIST PQC by default.
decision.json
{
"decision": "ALLOW",
"reason": "ALLOW_ALL_CHECKS_PASS"
}
qlx_action: PERMITTED
decision.json
{
"decision": "REFUSE",
"reason": "REPLAY_DETECTED"
}
qlx_action: BLOCKED
QLX is downstream of admissibility.
CohOS is the authority. QAKD is the key and identity layer. QLX is the security executor. The stack is intentionally ordered so that crypto cannot override refusal or operate without evidence.
Security actions are audit-bound.
QLX actions attach to the same artifact set as the admissibility decision. This makes “why was this allowed” as answerable as “what cipher was used.”
Round-trip verification is a first-class invariant.
In QCT’s operational posture, correctness is not assumed. QLX encryption is validated by explicit round-trip checks. If encryption and decryption do not agree under the declared parameters, the system refuses.
decision.json
{
"decision": "REFUSE",
"reason": "ROUNDTRIP_FAIL"
}
Guard, firewall, and VPN become governed execution surfaces.
QLX is deployed as an execution layer that consults CohOS before sensitive operations. These modes start in shadow mode: measure and refuse risky actions before enforcing, then graduate to enforcement once baselines are stable.

