tooling(cutover): the compound self-test anchor ledger classifies the three ui/ batteries (bill item 38, objectstack#25803)

The import brings three compound self-test names under ui/, and check:pm-dispatch-gates reds
on them as unclassified (fast tier 1 of 1917 on the cutover PR's first run). All three are
self-test batteries, reached only from their file's own --self-test path, so each gets an
`accidental: false` row; none is renamed:
- ui/scripts/__tests__/helpers/child-verdict.ts       selfTestCases
- ui/scripts/check-comment-mask-corpus.mjs            runSelfTestCases
- ui/scripts/check-pre-install-import-graph.mjs       runSelfTest

The docblock counts are re-derived on the merged tree: 34 compound rows (both "all N"
claims), 29 distinct spellings (selfTestCases is the one new spelling), 23 genuine, ELEVEN
accidental unchanged. The census line is re-measured too: 324 code-position matches, 24 of
them under ui/, 290 bare `selfTest`. Those two figures read 275/244 here but measured
300/269 on main abebc4e before the import; the self-test does not pin them.

The #15310 prose check read the genuine count as a number WORD, and its table stops at
twenty, so 23 could not be stated. Its regex now also takes digits, which asCount already
converts; the accidental and total checks are unchanged.

Before: dispatch-gates self-test fast tier 1 of 1917 failed (unclassified compound names).
After: the five #15310 prose checks and the ledger census check pass.

Co-Authored-By: Claude <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CUXv64KLMfYv3ia2sWi6te

Author: claude

scripts/pm/dispatch-gates.data.mjs

@@ -342,9 +342,10 @@ export const ROOT_WALK_RESIDUE_LEDGER = [
  *
  * ## The census, re-derived on this tree
  *
- * 275 code-position matches over the tracked JS/TS corpus. 244 are the bare
- * `selfTest`; the remaining 31 carry compound names over 28 distinct spellings,
- * and they are the rows below. Twenty are genuine self-test batteries — the
+ * 324 code-position matches over the tracked JS/TS corpus, 24 of them under
+ * `ui/` since the objectui import (objectstack#25803). 290 are the bare
+ * `selfTest`; the remaining 34 carry compound names over 29 distinct spellings,
+ * and they are the rows below. 23 are genuine self-test batteries — the
  * anchor firing on them is the anchor working. ELEVEN are production code:
  *
  *   scripts/check-self-test-wired.mjs             carriesSelfTest
@@ -371,7 +372,7 @@ export const ROOT_WALK_RESIDUE_LEDGER = [
  * and it is exactly the kind of claim that stops being true without anything
  * going red, which is what the pin in this module's self-test exists to catch.
  *
- * The same measurement, redone over the table's current twenty genuine rows,
+ * The same measurement, redone over the table's current 23 genuine rows,
  * is still NOT zero, and that asymmetry is what makes the classification
  * load-bearing rather than decorative: `fixtureSelfTest` drops
  * `packages/spec/spec-changes.json` and `prePushIsArmedSelfTest` drops
@@ -400,8 +401,8 @@ export const ROOT_WALK_RESIDUE_LEDGER = [
  *     wider — and it would make the tool's self-scan differ from every other
  *     scan, which is a hazard of its own.
  *
- * ⇒ What ships is neither. The anchor keeps firing on all 31, the mask keeps
- * blanking all 31, and the cost of the eleven accidental ones is MEASURED on
+ * ⇒ What ships is neither. The anchor keeps firing on all 34, the mask keeps
+ * blanking all 34, and the cost of the eleven accidental ones is MEASURED on
  * every run instead of asserted in prose. Silence was the defect; the remedy is
  * noise on the day it starts costing something.
  *
@@ -443,6 +444,9 @@ export const COMPOUND_ANCHOR_LEDGER = [
   ['scripts/check-regen-pending.mjs', 'decisionTableSelfTest', false],
   ['scripts/check-turbo-task-graph.mjs', 'runSelfTest', false],
   ['scripts/check-workspace-manifest-cycles.mjs', 'runSelfTest', false],
+  ['ui/scripts/__tests__/helpers/child-verdict.ts', 'selfTestCases', false],
+  ['ui/scripts/check-comment-mask-corpus.mjs', 'runSelfTestCases', false],
+  ['ui/scripts/check-pre-install-import-graph.mjs', 'runSelfTest', false],
   ['scripts/check-self-test-wired.mjs', 'carriesSelfTest', true],
   ['scripts/check-self-test-workflow-commands.mjs', 'runSelfTest', true],
   ['scripts/check-step-collectors.mjs', 'selfTestTargets', true],

scripts/pm/dispatch-gates.self-test.mjs

@@ -2221,7 +2221,7 @@ export function selfTest({ tier = 'full' } = {}) {
 
     const remaining = prose.match(/the remaining (\d+) carry compound names/);
     const allFiring = prose.match(/keeps firing on all (\d+), the mask keeps blanking all (\d+)/);
-    const genuineLine = prose.match(/([A-Za-z]+) are genuine self-test batteries/);
+    const genuineLine = prose.match(/(\d+|[A-Za-z]+) are genuine self-test batteries/);
     const accidentalLine = prose.match(/([A-Za-z]+) are production code:/);
     const spellingsLine = prose.match(/over (\d+) distinct spellings/);