DLC-APPLY: re-vendor the flipped decay KAT + pins (DKS-GOV item 3)

src/decay/__tests__/applyTimeDecay.kat.test.ts

@@ -22,7 +22,9 @@
  * binary64 — `Object.is` bit-exactness for all 32 vectors, no tolerance.
  *
  * Scope: profile uwr-weighted-lifts-v0.1 is testnet-provisional
- * (x-afiStatus: draft-non-implementation). Execution ≠ wiring — this suite
+ * (x-afiStatus: implemented since DLC-APPLY — the reactor serving-path
+ * derivation executes these vectors in production; DKS-GOV re-vendor).
+ * Execution ≠ wiring beyond that — this suite
  * executes vectors against the engine and nothing more: no reward, mint, or
  * validator-scoring path is touched and no registry is consumed at runtime
  * (the KAT JSON is imported only by this test). The KAT description's
@@ -52,7 +54,7 @@ const SIBLING_URL = new URL(
 
 /** sha256 of the vendored bytes, pinned at vendoring time (PR-UWR-KAT-EXEC). */
 const PINNED_SHA256 =
-  "1a1240cc9f8cc8ed70b22f4aa00e7cdfe1176f4f1d3c6b1e10e886b0cfe81b78";
+  "eb342a9c0a6ae0534c871930724e2b6ac2f5d6ee716699834e71f2a5c15cad1d";
 
 const DRIFT_MESSAGE =
   "Vendored copy or upstream KAT changed — requires a new scoped " +
@@ -107,7 +109,7 @@ describe("PR-UWR-KAT-EXEC: applyTimeDecay vs afi-config KAT vectors", () => {
 
   it("KAT metadata pins are intact", () => {
     expect(kat.schema, DRIFT_MESSAGE).toBe("afi.uwr-decay-kat.v0");
-    expect(kat["x-afiStatus"], DRIFT_MESSAGE).toBe("draft-non-implementation");
+    expect(kat["x-afiStatus"], DRIFT_MESSAGE).toBe("implemented");
     expect(kat.profileId, DRIFT_MESSAGE).toBe("uwr-weighted-lifts-v0.1");
     expect(kat.status, DRIFT_MESSAGE).toBe("testnet-provisional");
     expect(kat.decisionRef, DRIFT_MESSAGE).toBe(

src/decay/__tests__/kats/README.md

@@ -11,10 +11,10 @@ is never consumed by runtime code.
 |---|---|
 | Repository | `AFI-Protocol/afi-config` |
 | Path | `kats/uwr-profile/v0/apply-time-decay.kat.json` |
-| Merge commit (PR #17) | `fe329164919f0c1c9dc24bb5c279978fb680e983` |
-| Content commit | `7809af4b0308f7db47488947770a0ab9f236268d` |
-| Git blob SHA-1 | `1199f705ee65eef70e6c0c58fffcd10206693cf0` |
-| sha256 (file bytes) | `1a1240cc9f8cc8ed70b22f4aa00e7cdfe1176f4f1d3c6b1e10e886b0cfe81b78` |
+| Merge commit (PR #88, DKS-GOV re-vendor) | `36cdc66891f3fe11cd6967ea2a1cee5004bcc5a0` |
+| Content commit | `e8d7892` (status flip only; vectors byte-identical to the PR #17 vendoring) |
+| Git blob SHA-1 | `8baf7ddeec2efcea21ca5f8fe7e4f4f254629c2b` |
+| sha256 (file bytes) | `eb342a9c0a6ae0534c871930724e2b6ac2f5d6ee716699834e71f2a5c15cad1d` |
 | Executed against afi-core | `254185381c3c9be91da303454c1f7a27e8818983` (main, PR #16 merge) |
 | Profile | `uwr-weighted-lifts-v0.1` (afi-governance `decisions/uwr-profile-pin-v0.1.md`) |
 
@@ -31,10 +31,10 @@ bit-exact** against the pre-delegation closed form in
 outputs exactly. The vendored bytes are **not** "corrected" for this — the
 stale-looking pin is documented here instead, per change control.
 
-Decay-engine canonicality (UP-8, `math-authority-v0.1.md` §8) **remains
-OPEN**; neither the PR #16 delegation nor this KAT execution resolves it. This
-file pins the `GreeksDecayTemplate` surface only, as the KAT's own
-`engine.canonicalityNote` states.
+Decay-engine canonicality (UP-8) is **CLOSED by DLC-GOV D-DLC-1**
+(`decay-lifecycle-v0.1.md`, accepted 2026-08-24): the minutes-based
+exponential half-life law — exactly the `GreeksDecayTemplate.applyTimeDecay`
+surface these vectors pin — is the canonical engine.
 
 ## Change control
 
@@ -49,11 +49,18 @@ file pins the `GreeksDecayTemplate` surface only, as the KAT's own
 ## Status and scope
 
 - **testnet-provisional** (the KAT file's own `status`), under
-  `x-afiStatus: draft-non-implementation`.
+  `x-afiStatus: implemented` — flipped by DKS-GOV D-DKS-1 (accepted
+  `defe527`) when DLC-APPLY wired the reactor's serving-path derivation to
+  reproduce these vectors bit-exactly in production.
 - **Execution ≠ wiring.** These tests execute the KAT vectors against
   `applyTimeDecay` and nothing more. The KAT description's observation that
   base-1 one-half-life rows land on 0.5 (the testnet-provisional
   `minDecayScoreThreshold`, UP-9) stays data commentary: no eligibility gate
   is asserted or wired here, and no reward, mint, or validator-scoring path
   changes. Each follow-up (e.g. PR-UWR-STAMP, any runtime consumption, UP-8
   closure) is separately authorized.
+- **DKS-GOV re-vendor (2026-08-24):** the single byte-difference from the
+  PR #17 vendoring is the `x-afiStatus` marker (`draft-non-implementation` →
+  `implemented`); every vector, expected value, template, and the
+  `engine.sourceCommit 390b440` caveat are byte-identical. Pins above
+  updated per the change-control rule.

src/decay/__tests__/kats/apply-time-decay.kat.json

@@ -1,6 +1,6 @@
 {
   "schema": "afi.uwr-decay-kat.v0",
-  "x-afiStatus": "draft-non-implementation",
+  "x-afiStatus": "implemented",
   "profileId": "uwr-weighted-lifts-v0.1",
   "status": "testnet-provisional",
   "description": "Known-answer-test vectors for applyTimeDecay over the pinned GreeksDecayTemplate v1 template set of the version-pinned UWR profile uwr-weighted-lifts-v0.1 (UP-11 item 2). Pure data: expected values are decision-anchored literals; nothing in afi-config executes the decay function. 32 vectors: all four templates x base scores {1, 0.1875 (the D2 M2 golden anchor, = 3*2^-4)} x elapsed = {0, 1, 2, 4} half-lives. The base-1 rows at exactly one half-life land on 0.5, the testnet-provisional minDecayScoreThreshold (UP-9) - data commentary only; the qualification gate is not wired.",