Security has always been a race. A harder lock, a larger key, a more expensive computation. KAGE is not a faster runner in that race. It steps off the track entirely. This paper introduces kenography, from the Greek kenos, void, as a distinct category of information security, and presents KAGE as its first implementation. The categorical distinction is precise: in every cryptographic system currently deployed, a key exists inside the observable channel. The security question is how difficult it is to find. In kenography, the encoding basis never enters the observable channel at all. The key is not hidden. It is absent. The consequence is structural rather than incremental. An adversary on the channel, quantum or classical, encounters nothing to act on. Quantum computers derive their advantage from observable structure they can amplify. KAGE provides none. This is not a resistance to quantum attacks. It is the absence of an attack surface. The brute-force search space for a 1, 000-bit message contains 2⁵2, 000 candidate states, a number with 15, 654 decimal digits. That number is astronomical, but does not measure how hard the search is. It measures how large a space must be navigated with no signal, no gradient, and no way to recognise success. Every candidate looks identical. The search has no finish line. KAGE provides four operations on this foundation: key exchange, symmetric encryption, a hash function, and digital signatures. All four complete in under 5 ms on a consumer CPU running an unoptimised Python implementation. The ciphertext output passes all 13 applicable tests of the NIST SP 800-22 Rev 1a statistical test suite across 100 independent sequences, with mean p-values between 0. 47 and 0. 55 and pass rates of 98–100%. A public challenge API is live. A message was encoded with KAGE at server startup under a key that was immediately discarded. The ciphertext and both public scalars are published. The key no longer exists. Recover the plaintext. More info and link in section 9 of the present document.
Jessy Pensédent (Wed,) studied this question.