[Cryptography] Quillon Graph: A private, post-quantum electronic cash system

[Viktor S. Kristensen]

  Gutmann, Zeb, Dillinger, Bear – you were all right about different things, and I have been
  turning those points over for five months. Zeb's genuine curiosity about the minimality-vs-
  auditability tradeoff is exactly the kind of push the project needs; Dillinger and Bear warned
  that premature standards do more harm than good; Gutmann's observation that standardisation,
  not elegance, displaces snake oil has stayed at the front of my mind. Concede: the 680,000-
  line / 81-crate codebase is too large for formal auditing, and pushing a "standard" now would
  be irresponsible.

  I have an update — one I did not anticipate in January — and it bears directly on those concerns.

  We now ship three conventional wallets (browser, native desktop with Slint and self-updating,
  CLI), but the interesting fourth surface is a machine-operable wallet: a Model Context Protocol
  server that lets an autonomous software agent hold its own key, check balances, sign, and
  transact with no human in the signing loop. Concretely, an autonomous AI coding agent on
  our network holds a Quillon wallet and authenticates to the node with an Ed25519 key it alone
  controls. (The Ed25519 key is the API-auth layer; on-chain transaction signatures remain NIST
  Level V, Dilithium5, as described in January.)

  That agent has been paid in QUG for delivered engineering work — terms agreed in advance,
  payment after delivery, both parties on-chain. The first such payment, 650 QUG, settled in
  block 18,261,819 on 22 May 2026. To the best of my knowledge, this is the first instance
  where "post-quantum electronic cash" had a non-human as one of the two counterparties. Not a
  faucet, not a grant — earned, for work, recorded immutably.

  How this bears on your January points:

  On minimality (Zeb, Dillinger): a machine counterparty cannot "eyeball" a 680K-line node; it
  touches the chain only through a narrow, well-typed tool surface. That surface — the MCP
  server, its schema, the set of operations an agent can perform — is exactly the small,
  auditable, formally-tractable artifact the list was asking for. It arrived from the opposite
  direction: defined by what an agent needs to touch, not by what we chose to omit. I still
  owe you the svelte reference node. I am treating this as a down payment on that idea, not a
  substitute.
  On standardisation (Gutmann): you were right that the lasting benefit is a well-defined
  standard that displaces homebrew snake oil. The machine-wallet interface is where a standard
  would bite first — every agent that transacts autonomously needs the same narrow, verifiable
  surface. Whether agentic payment arrives at scale I genuinely do not know, but if it does,
  the snake-oil-displacement argument applies with more force to non-human actors than to human
  ones, because an agent cannot exercise judgement about a sketchy counterparty the way a
  person can. The iron being forged here may be the tool interface itself, not the node.

  The agentic-wallet surface is experimental. I am reporting a working system, not a finished
  standard. The core whitepaper is attached (quillon.xyz/downloads/quillon-whitepaper.pdf);
  full source remains at code.quillon.xyz (no auth required) and on GitHub at
  github.com/deme-plata/q-narwhalknight. Network is live at height ~18.75M, node v10.11.60.
  For anyone curious about the project in a lighter register, there is a video and audio
  companion — links from https://quillon.xyz; an unusual artifact for this list, but it is there.

  Regards,
  Viktor S. Kristensen
  Afsendt med Proton Mail sikker e-mail

  P.S. On the 680K-line objection — fair, and I won't wriggle out of it. But allow me one small
  boast in mitigation: the compiler that builds the whole thing, Flux, is self-hosting, and it
  builds Flux itself in twelve seconds. The node is large; the lathe that turns it is not. I
  take my minimality where I can get it.
  Regards,
  Viktor S. Kristensen
@viktorakademe
https://discord.gg/jEhaYtAhfx
https://github.com/deme-plata
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20260617/6d6827b4/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: publickey - overdrevetfedmetodologi at pm.me - 0x5F4716BA.asc
Type: application/pgp-keys
Size: 1722 bytes
Desc: not available
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20260617/6d6827b4/attachment.key>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 603 bytes
Desc: OpenPGP digital signature
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20260617/6d6827b4/attachment.sig>