[QA] review: name the mapper branch in row failures; scope the edit-rule note

- json-looking-judge-prompts.spec.ts: the JsonLookingPrompt.why field was
  populated and never read, so a missing row failed with "row for <name>" and
  the reader had to go back to the table to learn which mapper branch that
  shape reaches. Carry name+why through the UI step and put it in the message.

- taxonomy.yaml: the edit-rule note read as though the edit dialog was driven
  for both prompt shapes. Only the plain-string half is; the content_array
  half is API-only, because the dialog renders no image part. Scoped the note
  to say so, and recorded the open question — the content_array test asserts
  content-shape preservation, which no capability in this area names, so it
  rides on edit-rule until a human decides whether to add a key or drop the tag.

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

Author: comet-qa-bot

tests_end_to_end/coverage/taxonomy.yaml

@@ -661,11 +661,23 @@ areas:
       rule-filters:            { covered: false }
       sampling-rate:           { covered: true,  tier: t2-cuj, note: "50% rule vs 100% control over one 30-trace batch, binomial band 15-85%; plus a 0%-rate rule at trigger_scope=both, which must skip every SDK trace and still score experiment/playground/optimization ones" }
       clone-rule:              { covered: false }
-      # Scoped: the edit dialog is opened, its hydrated prompt asserted byte-exact,
-      # and submitted unchanged so the save round trip is real. Editing a FIELD
-      # through the dialog (renaming, changing the model, rewriting the prompt) is
-      # still uncovered.
-      edit-rule:               { covered: true,  tier: t2-cuj, note: "dialog hydration + no-op re-save preserve a judge prompt byte-for-byte, in both the plain-string and content_array shapes; editing a field through the dialog is not covered" }
+      # Scoped, and the two halves are not equally scoped:
+      #   - plain-string prompt: the edit dialog IS opened, its hydrated prompt
+      #     asserted byte-exact, and submitted unchanged, so the UI save round
+      #     trip is real.
+      #   - content_array prompt: API-level only. The dialog renders no image
+      #     part, so the url/detail that a regression would drop are not
+      #     observable in the UI; that half is a read-then-write over REST and
+      #     no dialog gesture is made.
+      # Editing a FIELD through the dialog (renaming, changing the model,
+      # rewriting the prompt) is still uncovered either way.
+      #
+      # OPEN, needs a human call: the content_array test asserts judge-message
+      # content-SHAPE preservation, which no capability in this area names. It
+      # rides on edit-rule for want of a better key. Either add a capability for
+      # it or drop that tag and leave the behaviour honestly uncovered — do not
+      # let it sit here as a silent widening of edit-rule.
+      edit-rule:               { covered: true,  tier: t2-cuj, note: "plain-string prompt: edit-dialog hydration + no-op re-save, byte-for-byte; content_array prompt: API-only read-then-write re-save, no dialog gesture; editing a field through the dialog is not covered" }
       enable-disable-rule:     { covered: true,  tier: t2-cuj, note: "edit-dialog switch; control rule proves scoring stopped, then resumed" }
       delete-rule:             { covered: true,  tier: t2-cuj, note: "row kebab delete; control rule proves scoring stopped" }
       # Deliberately still false. online-evaluation-python-metric-errors.spec.ts

tests_end_to_end/e2e/tests/online-evaluation/online-evaluation-json-looking-judge-prompts.spec.ts

@@ -73,10 +73,17 @@ test.describe('Online Evaluation — judge prompts that look like JSON', { tag:
     automationRulesCleanup,
   }) => {
     const controlName = `${testNamespace}-control`;
-    const seededNames = [
-      controlName,
-      ...JSON_LOOKING_PROMPTS.map((p) => `${testNamespace}-${p.label}`),
+    // Carried as name+why rather than name alone so a missing row names the
+    // mapper branch it came from: "row for …-empty-object missing" is a bug
+    // report, "row missing" is a starting point for one.
+    const seededRules: Array<{ name: string; why: string }> = [
+      { name: controlName, why: 'ordinary prose — the collateral-damage control' },
+      ...JSON_LOOKING_PROMPTS.map((p) => ({
+        name: `${testNamespace}-${p.label}`,
+        why: p.why,
+      })),
     ];
+    const seededNames = seededRules.map((r) => r.name);
 
     await test.step(
       `Seed a prose control rule plus ${JSON_LOOKING_PROMPTS.length} bracket-opening ones`,
@@ -152,8 +159,8 @@ test.describe('Online Evaluation — judge prompts that look like JSON', { tag:
       await onlineEval.goto(project.id);
       await onlineEval.waitForReady();
 
-      for (const name of seededNames) {
-        await expect(onlineEval.ruleRow(name), `row for ${name}`).toHaveCount(1);
+      for (const { name, why } of seededRules) {
+        await expect(onlineEval.ruleRow(name), `row for ${name} — ${why}`).toHaveCount(1);
       }
       await expect(
         onlineEval.ruleRows,