plan(roadmap): cut v1.135 around PERF-P01 alone; TRACE-P02 → v1.136

artifacts/swreq/SWREQ-FALCON-MATHF32-P06.yaml

@@ -3,7 +3,7 @@ artifacts:
     type: sw-req
     title: "MATHF32-P06 — machine-checked Cody-Waite argument-reduction bound → FULL-range sin/cos accuracy"
     status: proposed
-    release: falcon-v1.135.0
+    release: falcon-v1.139.0
     description: >
       The remaining and HARDEST ingredient: a machine-checked bound on the
       Cody-Waite argument reduction r = ((x - n*PIO2_HI) - n*PIO2_MID) -
@@ -17,7 +17,13 @@ artifacts:
       slice (the reduction cancellation is the part those references spend
       most of their pages on — weeks, not days) and deliberately sequenced
       AFTER P04/P05 so the reduced-range proof lands first.
-    tags: [falcon, relay-math, f32, gappa, argument-reduction, machine-checked, qualification, v1.130]
+      Deferred to v1.139 (2026-08-26). Not blocked and not obsolete — simply
+      lower value than the four releases ahead of it (PERF-P01 measurement,
+      TRACE-P02 promotion, the trusted-base publication, and the oracle
+      repairs), each of which moves a claim the project currently cannot
+      make or cannot defend.
+
+    tags: [falcon, relay-math, f32, gappa, argument-reduction, machine-checked, qualification, v1.139]
     fields:
       req-type: non-functional
       priority: could

artifacts/swreq/SWREQ-FALCON-OCI-P03.yaml

@@ -3,7 +3,7 @@ artifacts:
     type: sw-req
     title: "OCI-P03 — bounded messaging payloads: adopt the wit-bindgen fork (fixed-length lists + arena/elision)"
     status: proposed
-    release: falcon-v1.135.0
+    release: falcon-v1.139.0
     description: >
       Message-carrying seams shall declare BOUNDED payloads, so the work memory
       they require can be sized and defended rather than being unbounded.
@@ -42,7 +42,22 @@ artifacts:
       justification for adopting the FORK is obsolete; the allocator work
       (gale-owned arena via cabi-realloc-extern, relay#330) is what remains.
       Rewrite the requirement around that before implementing it.
-    tags: [falcon, oci, wasm, wit-bindgen, messaging, embedded, v1.135]
+      Still parked at v1.139 pending the REWRITE noted above — do not
+      implement against the current text. Restated because it is easy to
+      pick up by release order: upstream wit-bindgen 0.60 already generates
+      `[u8; 8]` for `list<u8, 8>`, so the fixed-length-list justification for
+      adopting the fork is gone. What remains is the gale-owned arena
+      (relay#330), and measured 2026-08-25 the fork is three minors stale
+      (synced at upstream 0.58, upstream now 0.61.1) with NOTHING in relay
+      depending on it.
+
+      The drift that actually matters is different and is in the flight
+      path: our components pin `wit-bindgen-rt` 0.41.0 while the workspace
+      is on 0.60.0 — a 19-minor spread in one repo, and 0.41 generates the
+      bindings baked into every published artifact. Rewrite this requirement
+      around THAT migration, which needs no fork at all.
+
+    tags: [falcon, oci, wasm, wit-bindgen, messaging, embedded, v1.139]
     fields:
       req-type: interface
       priority: should

artifacts/swreq/SWREQ-FALCON-TRACE-P02.yaml

@@ -3,7 +3,7 @@ artifacts:
     type: sw-req
     title: "TRACE-P02 — promote the V: drive implemented→verified from the verification gate's own per-artifact result"
     status: proposed
-    release: falcon-v1.135.0
+    release: falcon-v1.136.0
     description: >
       Artifact status shall reflect executed evidence, so that "verified" means
       a gate ran the artifact's steps and they passed.
@@ -36,7 +36,26 @@ artifacts:
       a downstream consumer (jess) is BLOCKED on the relocation fix and the
       published artifact is the only way to hand it to them. Accumulating
       unrelated scope would make them wait on work they do not need.
-    tags: [falcon, rivet, traceability, verification, hygiene, v1.135]
+      Re-scoped to v1.136 (2026-08-26). v1.135 is cut around PERF-P01 alone —
+      the cascade timing baseline — because performance was the one axis on
+      which the project could make no claim at all, and that gap was the
+      highest-value item in the standing assessment. TRACE-P02 follows
+      immediately and is now precisely sized (see below).
+
+      MEASURED SIZING (2026-08-25), which decomposes this into three
+      populations with very different costs rather than one sweep:
+        155 of 236 verification artifacts carry >=1 CI-runnable step
+            -> promotable MECHANICALLY from the gate's per-artifact result
+         36 are bench-only
+            -> need a named waiver or a real step; judgement, not bookkeeping
+         45 have NO steps at all
+            -> cannot be evidence as they stand; write steps or demote
+      Also measured: ZERO `implemented` requirements lack a verifier. The
+      22/247 verified figure is entirely UNPROMOTED verifiers, not missing
+      evidence — which makes this far cheaper than it looks, except for the
+      45.
+
+    tags: [falcon, rivet, traceability, verification, hygiene, v1.136]
     fields:
       req-type: constraint
       priority: should