CB-WP-0018 T03/T04: explanations, and window 1's verdict

T03: input::describe writes a sentence per legal command; data-descs
carries them in step with data-targets; the ghost already following the
pointer shows the one for whatever legal target is under it, so the
explanation lands beside the target with no overlay layer to keep
aligned. ADR-0010 D1 binds -- the page renders it, never composes it.

Both mutations INITIALLY SURVIVED because the fixture's Attack card had
exactly one target, where an off-by-one shift and a truncation are both
no-ops. CB-EV-0014's lesson one level in: a fixture too thin to express
a failure is how the failure survives. Two attack targets now, both red.

T04: chaos rate d4 -> d8, window 2 open at 12 declarations, retiring if
an override changes nothing twice running. Window 1's condition was NOT
met -- both overrides changed the outcome -- so the mechanism is kept.
The weakest part of the decision is that it is a rate change argued from
n=2, so window 2 carries a falsifier: no override at all is evidence the
rate went too far, not that the mechanism is healthy.

InnerLoop.md hit 401 lines and the loadability gate fired; the rationale
moved to InnerLoopReference.md, structurally, per the standing precedent
that limits are not raised.

CB-WP-0017 settled at $9.48/40 against $5.19/23 reported mid-flight,
83% higher. Six for six, always low -- read by re-running the instrument
at the moment of quoting, which is CB-EV-0015's correction applied for
the first time.

make all exits 0.

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

gates.toml

@@ -120,15 +120,15 @@ retire_if = "two passes run with no finding while artifacts keep growing — tha
 id = "CHAOS"
 name = "the chaos roll"
 target = ""
-checks = "d4 on each tier declaration, 12-declaration calibration window"
+checks = "d8 on each tier declaration, 12-declaration calibration window (window 2, opened 2026-08-03; window 1 ran at d4)"
 added = "2026-07-30"
-review_by = "2026-09-30"
+review_by = "2026-11-30"
 caught = [
   "CB-WP-0011: first fire in 6 declarations — d4=4 rolled stage 1 from structural L to S; the deleted survey would have opened on 2D toolkits while the existing text renderer was showing 24 of 41 view fields (CB-EV-0009 §1)",
   "CB-WP-0017: d4=4 — second override in twelve, and the first to roll UP (structural S → M). It bought ADR-0010: the script's widening from 'it does one thing' to holding a drag, following the pointer and marking other elements would otherwise have landed under a tier-S provenance paragraph, silently outgrowing ADR-0007 D5. The ADR's own finding is that the permitted and forbidden designs are indistinguishable from outside, which demoted the vocabulary grep to a cheap first line and produced the two behavioural controls that replaced it",
   "CB-WP-0012: d4=1, no override — and the contrast is the entry. Tier L at full weight deleted its own structural trigger: adversarial review withdrew the capability port the declaration was made to build (ADR-0007 D2), and corrected the survey's headline claim by 85x (128x -> 1.5x, CB-EV-0010 §2). Two passes on one subject at two tiers, priced: 0.123 $/response at L against 0.099 at S (CB-EV-0010 §5)",
 ]
-retire_if = "the window closes with no overridden tier producing a different outcome than the argued one — the evaluation this window exists to make possible"
+retire_if = "an override changes nothing twice running (window 2 condition, CB-WP-0018 T04). Window 1's condition — no override changing the outcome — was NOT met: both did, so the mechanism was kept and the rate dropped d4 → d8 instead"
 # VERDICT, CB-EV-0015 §5 (window closed 2026-08-02, 12 declarations, 2 overrides).
 # Not retired: both overrides changed the outcome. CB-WP-0011 (L→S) bought a
 # defect in the existing renderer that the deleted survey would have walked