From f555e9da542289f759e080d39440aad9e0791fef Mon Sep 17 00:00:00 2001
From: garrytan <garrytan@gmail.com>
Date: Thu, 8 Oct 2026 06:48:48 +0000
Subject: [PATCH] test(evals): name the office-hours-auto-mode case so the PR
 lane can select it

The PR lane runs each short probe with a Bun name pattern ending in its case
id. The office-hours AskUserQuestion-blocked smoke had a descriptive test
name and no CASE_TEST_NAMES label, so whenever office-hours changed its shard
matched 0 tests and failed in 0 s (E2E Evals slice 6 on #3075). Name the
test after its id, keep the description on the describe, and check that
every literal-named PR probe has a test the pattern selects.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
---
 test/paid-pr-profile.test.ts                  | 14 ++++++++++++++
 test/skill-e2e-office-hours-auto-mode.test.ts |  5 +++--
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/test/paid-pr-profile.test.ts b/test/paid-pr-profile.test.ts
index 9b1d16707..7b5f47ed7 100644
--- a/test/paid-pr-profile.test.ts
+++ b/test/paid-pr-profile.test.ts
@@ -12,6 +12,7 @@ import {
   type PaidCaseSelection, type SliceResult,
 } from '../scripts/test-paid-shards';
 import { PR_PROFILE_CASE_IDS, PR_PROFILE_FILES } from '../scripts/test-pr-profile';
+import { fileCaseRegistration } from '../scripts/lib/paid-cases';
 import { E2E_TOUCHFILES } from './helpers/touchfiles-data';
 import { resolveModuleSelection } from './helpers/e2e-helpers';
 
@@ -64,6 +65,19 @@ describe('PR profile paid-runner integration', () => {
     }
   });
 
+  test('every short probe with literal test names has a test the PR lane name pattern selects', () => {
+    for (const [file, cases] of Object.entries(PR_PROFILE_FILES)) {
+      const source = fs.readFileSync(path.join(ROOT, file), 'utf8');
+      if (fileCaseRegistration(file, source).computed) continue;
+      const titles = [...source.matchAll(/\b(?:describe\w*|test\w*|it)(?:\.\w+)?\(\s*(['"`])((?:(?!\1)[^\\]|\\.)*)\1/g)].map(m => m[2]!);
+      const names = titles.flatMap(outer => [outer, ...titles.map(inner => `${outer} ${inner}`)]);
+      for (const id of cases) {
+        const pattern = new RegExp(prProfileTestNamePattern(file, { e2e: [id], judges: [] }));
+        expect(names.some(name => pattern.test(name)), `${file} has no test named for ${id}`).toBe(true);
+      }
+    }
+  });
+
   test('planner binds E2E and judge IDs and plans the CEO plan-mode probe in the PR lane', () => {
     const manifest = ceoManifest();
     expect(manifest.profile).toBe('pr');
diff --git a/test/skill-e2e-office-hours-auto-mode.test.ts b/test/skill-e2e-office-hours-auto-mode.test.ts
index 8680a2f1e..6036b1d82 100644
--- a/test/skill-e2e-office-hours-auto-mode.test.ts
+++ b/test/skill-e2e-office-hours-auto-mode.test.ts
@@ -24,12 +24,13 @@ import { assertNoPlanFileDecisions } from './helpers/plan-mode-evidence';
 
 const describeE2E = describeE2ETier('gate');
 
-describeE2E('office-hours AskUserQuestion-blocked smoke (gate)', () => {
+// The PR lane selects this case by a test name ending in its id.
+describeE2E('office-hours AskUserQuestion-blocked smoke (gate): AskUserQuestion surfaces when --disallowedTools AskUserQuestion is set', () => {
   // With AskUserQuestion blocked the mode question must still be asked (the
   // prose fallback, observed as 'asked'); a plan-file ## Decisions section is
   // not a substitute. Failure signals also include 'auto_decided' +
   // silent_write/exited/timeout.
-  test('AskUserQuestion surfaces when --disallowedTools AskUserQuestion is set', async () => {
+  test('office-hours-auto-mode', async () => {
     const obs = await runPlanSkillObservation({
       skillName: 'office-hours',
       inPlanMode: true,