Implement bounded Richese Nexus Betrayal auctions
This commit is contained in:
1 parent
a921c0da9b
commit
9c0000e469
25 files changed
+2208
-39
No files matched your search
@@ -1,5 +1,52 @@
|
||||
# Dune implementation status
|
||||
|
||||
## 30 September 2026 — bounded Richese Nexus Betrayal auction integration
|
||||
|
||||
Five implementation lanes delivered pure source/invoice/custody policy,
|
||||
typed prepayment engine continuation, human controls and minimal legal bots,
|
||||
authenticated SQLite recovery, and source-bound internal guidance.
|
||||
Fresh `richese-betrayal` entry supports Basic/Advanced paired CHOAM/Richese
|
||||
plus classic seats, the physical CHOAM deck and Nexus alone.
|
||||
|
||||
A public cache auction won by native Richese may be vetoed before payment:
|
||||
the exact intended card enters ordinary discard, nobody pays or acquires,
|
||||
and the original lot retires once. Another buyer's positive cache or
|
||||
concealed Black Market sale keeps the original price, authorized escrow
|
||||
split, physical delivery and earned Harkonnen bonus; only its recipient
|
||||
changes from Richese to bank. All-pass keeps the Nexus and settles the
|
||||
unchanged original purchase. Public held-Nexus presence determines neutral
|
||||
acknowledgement membership; actual identity affects only own private Use.
|
||||
Independent pending/completed cursor history binds saved source ownership
|
||||
and rejects orphaned frames before disclosure or SQL writes.
|
||||
|
||||
Independent review found and repaired donor escrow double-counting and a
|
||||
Harkonnen bonus path that skipped exhausted-cache preflight. Regressions
|
||||
cover a seller funding its buyer with zero unpledged spice and rejection
|
||||
of an unfinishable closing bid before any receipt, payment or delivery.
|
||||
Existing wording/incidental assertions were removed rather than re-pinned.
|
||||
|
||||
Verification: `npm run check` passes types, lint and **5,953 offline tests**;
|
||||
`npm run build` passes. Eight focused files pass **80 cases**, including
|
||||
both alternatives/all-pass, private timing parity, actual table controls,
|
||||
four legal bot profiles, JSON/production SQLite restart, competing Use/Pass,
|
||||
no-write rejections and later recyclable physical custody. Final independent
|
||||
economy and persistence re-reviews found no remaining scoped defect.
|
||||
|
||||
Actual prototype CLI started four owned disposable rooms from version 5
|
||||
to setup 6. Genuine setup/bids created typed opportunities; a separate
|
||||
runtime smoke proves Basic/Advanced veto and cache/Black Market diversion.
|
||||
At 390 pixels, real controls prove Basic veto, Advanced all-pass, Basic
|
||||
cache diversion and refreshed Advanced Black Market diversion without an
|
||||
unknown-face inspector. Saved outcomes retain the original completed event,
|
||||
exact card custody and buyer/seller balances. No human games or seats were
|
||||
reset or modified.
|
||||
|
||||
Current built-worker local HTTP is **51/55**, with three POST 503s and one
|
||||
Advanced-preview timeout; it is not green acceptance. Proxy-free container
|
||||
CI and publication are recorded separately after the push. Full Richese
|
||||
games, hidden normal-deck self-purchases, private special-Karama veto,
|
||||
broader modules and live NAS acceptance remain open; no public gate changes.
|
||||
|
||||
## 30 September 2026 — opt-in printed Kull Wahad integration
|
||||
|
||||
Five implementation lanes supplied validated activation frames,
|
||||
|
||||
Reference in new issue
Block a user