HANDOFF: bin/riddlc restaged on RC6, and the corpus cost measured

`../bin/riddlc` was the RC4 build; it is now 2.0.0-rc.26-5-ab3ada49 on Scala
3.9.0-RC6, staged through publish-and-stage.sh so ivy and binary came from one
invocation. Recorded as verified BY BEHAVIOUR rather than by the version string,
which a dirty-tree build understates: the boundary-violating repro now errors
while both legal forms stay clean, the inversion of how those same three models
behaved on rc.26.

Also records the corpus measurement that was owed on the boundary rule -- 6
findings in 1 of 190 models, all true positives -- so the next session does not
have to re-run it to find out.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Author: reid-spencer

NOTEBOOK.md

@@ -35,9 +35,12 @@ byte-identical to the rc.26 certification on RC4**, which is the expected result
 because no test source changed, and is therefore evidence that nothing was SKIPPED
 rather than that nothing broke.
 
-**`../bin/riddlc` is still the RC4 build** and is deliberately stale; restage with
-`scripts/publish-and-stage.sh` when a consumer needs RC6 behaviour. `[0.6]` in
-BACKLOG carries the full site list and the grep that finds them, for the eventual
+**`../bin/riddlc` is `2.0.0-rc.26-5-ab3ada49`, built on Scala 3.9.0-RC6** and restaged
+2026-08-27 via `scripts/publish-and-stage.sh`, so the ivy artifacts and the binary came
+from one invocation and agree. It is a real Mach-O arm64 file, not a symlink and not the
+JVM launcher. **Verified by BEHAVIOUR, not by the version string**: the boundary-violating
+repro errors and the two legal forms stay clean, which is the inversion of how the same
+three models behaved on rc.26. `[0.6]` in BACKLOG carries the site list for the eventual
 3.9.0-final bump.
 
 ### Traps a fresh session would hit
@@ -79,7 +82,8 @@ Nothing awaits triage, which is a fact about right now and not a reason to skip
 `release/2` carries the boundary work (`msg-target-crosses-boundary`) unpushed, and the
 Computational Model amendment is committed separately in the `ossuminc` repo (`dcaca50`),
 also unpushed. Certified tri-platform locally, zero failures on all six rows; CI has not
-seen it.
+seen it. Corpus cost measured on the staged binary: **6 findings in 1 of 190 models**
+(reactive-bbq), all true positives, migration filed in `riddl-models/task/`.
 
 ## 2026-08-27 — a boundary rule that already half-existed (`msg-target-crosses-boundary`)