claude-code-session-manager 0.98.0 → 0.101.0

package/src/main/__tests__/runVerify.test.cjs

@@ -1,1784 +0,0 @@
-/**
- * runVerify.test.cjs — unit tests for the post-run reconciliation verifier.
- *
- * Run standalone: timeout 180 npx vitest run src/main/__tests__/runVerify.test.cjs
- *
- * Three fixture scenarios (matching the 2026-05-23→24 incident window):
- *
- *   1. Case 1 (false-PASS): M2 acceptance script prints test failures + Traceback,
- *      then echoes "M2 acceptance: PASS". Scheduler used to stamp 'completed'.
- *      Expected: transcript_errors → needs_review
- *
- *   2. Case 2 (HALT): M7 agent correctly halts because M1 soak is unmet.
- *      Expected: halt → pending, reason embeds soak floor date
- *
- *   3. Case 3 (deps_unmet): PRD 56 has an unsatisfied dependsOn entry.
- *      Expected: deps_unmet → pending
- *
- *   4. Truly clean: all tests pass, no error markers.
- *      Expected: clean → null (caller may stamp 'completed')
- */
-
-import { test, beforeAll, afterAll, vi } from 'vitest';
-const assert = require('node:assert/strict');
-const os = require('node:os');
-const fs = require('node:fs');
-const path = require('node:path');
-const { execFileSync } = require('node:child_process');
-const { verifyRun, parseLog } = require('../runVerify.cjs');
-
-// The verdicts sidecar goes through config.writeJsonSync, whose write allow-list
-// rejects os.tmpdir() fixtures. Route it to a plain write for these fixtures.
-let writeSpy;
-beforeAll(() => {
-  writeSpy = vi.spyOn(require('../config.cjs'), 'writeJsonSync').mockImplementation((abs, data) => {
-    fs.writeFileSync(abs, JSON.stringify(data, null, 2) + '\n');
-    return { ok: true, mtimeMs: 0 };
-  });
-});
-afterAll(() => writeSpy.mockRestore());
-
-// ─── helpers ──────────────────────────────────────────────────────────────────
-
-/** Create a temp directory that is cleaned up after the test. */
-function makeTmpDir() {
-  return fs.mkdtempSync(path.join(os.tmpdir(), 'run-verify-test-'));
-}
-
-function rmdir(dir) {
-  try { fs.rmSync(dir, { recursive: true, force: true }); } catch { /* */ }
-}
-
-/**
- * Build a minimal stream-JSON log file.
- * `events` is an array of raw JSON objects to serialise one-per-line.
- * A "[scheduler] starting..." prefix line is also prepended (must be skipped).
- */
-function writeLog(dir, slug, events) {
-  const lines = ['[scheduler] starting ' + slug + ' at 2026-05-24T00:00:00.000Z'];
-  for (const ev of events) lines.push(JSON.stringify(ev));
-  lines.push('[scheduler] exit code=0 (raw code=0 signal=null) duration=47s');
-  fs.writeFileSync(path.join(dir, `${slug}.log`), lines.join('\n') + '\n');
-}
-
-/** Write a minimal PRD .md file with frontmatter + body. */
-function writePrd(dir, slug, body) {
-  const text = `---\ntitle: Test PRD\ncwd: /tmp\nestimateMinutes: 30\n---\n${body}\n`;
-  fs.writeFileSync(path.join(dir, `${slug}.md`), text);
-  return path.join(dir, `${slug}.md`);
-}
-
-// ─── fixture: Case 1 — false-PASS (M2 acceptance) ─────────────────────────
-
-test('Case 1: false-PASS (Traceback + echo PASS) → transcript_errors/needs_review', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '39-signal-builder-m2-sentiment-stack';
-    const logEvents = [
-      // Assistant turn: run acceptance script
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_m2_001',
-            name: 'Bash',
-            input: {
-              command: 'bash check_m2.sh',
-              description: 'Run M2 acceptance checks',
-            },
-          }],
-        },
-      },
-      // Tool result: test failures then echo PASS (the false-positive pattern)
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_m2_001',
-            content: [
-              '=== Signal-builder side ===',
-              'ERROR tests/test_news_signals_panel.py',
-              '!!!!!!!! stopping after 1 failures !!!!!!!!',
-              '1 error in 0.03s',
-              'SB: 0',
-              '=== Sanity: SB sentiment_cache.db rows ===',
-              'rows: 0',
-              '=== contract.json panels.sentiment ===',
-              'Traceback (most recent call last):',
-              '  File "<string>", line 1, in <module>',
-              "KeyError: 'panels.sentiment'",
-              'M2 acceptance: PASS',
-            ].join('\n'),
-            is_error: false,
-          }],
-        },
-      },
-      // Result: agent declares success
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'M2 acceptance criteria verified. All checks passed.',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(tmp, slug, '# M2 Sentiment Stack\n\nBuild the sentiment pipeline.');
-
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-    });
-
-    assert.equal(verdict.verdict, 'transcript_errors', `expected transcript_errors, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-    assert.ok(verdict.reason.length > 0, 'reason should be non-empty');
-
-    // verdicts.json sidecar should have been written
-    const sidecar = JSON.parse(fs.readFileSync(path.join(tmp, `${slug}.verdicts.json`), 'utf8'));
-    assert.equal(sidecar.schemaVersion, 1);
-    assert.equal(sidecar.verdict, 'transcript_errors');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── fixture: Case 2 — HALT (M7 soak unmet) ──────────────────────────────────
-
-test('Case 2: HALT in result event → halt/pending with floor date in reason', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '44-signal-builder-m7-final-cutover';
-    const haltMsg = [
-      'HALT: M7 prerequisite — M1 not yet soaked 30 days × 6 phases',
-      '  M1 commit: f91ce44  2026-05-23 (age: 0 days, need ≥ 180)',
-      '  M7 earliest eligible: ~2026-11-19',
-    ].join('\n');
-
-    const logEvents = [
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_m7_001',
-            name: 'Bash',
-            input: { command: 'git log --oneline m1-tag', description: 'Check M1 soak period' },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_m7_001',
-            content: 'f91ce44  2026-05-23 feat(m1): initial trading signals',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: haltMsg,
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(
-      tmp,
-      slug,
-      '# M7 Final Cutover\n\nOnly after M1–M6 have run 30+ days each.',
-    );
-
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-    });
-
-    assert.equal(verdict.verdict, 'halt', `expected halt, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'pending');
-    assert.ok(verdict.reason.includes('HALT'), 'reason should mention HALT');
-    // Floor date from the HALT message should appear in reason
-    assert.ok(
-      verdict.reason.includes('2026-11-19') || verdict.reason.length > 20,
-      `reason should embed soak floor date: ${verdict.reason}`,
-    );
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── fixture: Case 3 — deps_unmet (dependsOn unsatisfied) ────────────────────
-
-test('Case 3: dependsOn pointing at pending dep → deps_unmet/pending', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '56-burrow-mcp-retire-tier3-frozen-endpoints';
-    const logEvents = [
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'Completed burrow MCP tier-3 retirement.',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(
-      tmp,
-      slug,
-      '# Burrow MCP Retire Tier-3\n\nRetire frozen endpoints.',
-    );
-
-    // M7 is a predecessor that is still pending.
-    const allJobs = [
-      { slug: '44-signal-builder-m7-final-cutover', status: 'pending', finishedAt: null },
-      { slug: slug, status: 'running', finishedAt: null },
-    ];
-    const queueEntry = {
-      slug,
-      status: 'running',
-      dependsOn: ['44-signal-builder-m7-final-cutover'],
-    };
-
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry, allJobs });
-
-    assert.equal(verdict.verdict, 'deps_unmet', `expected deps_unmet, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'pending');
-    assert.ok(verdict.reason.includes('44-signal-builder-m7-final-cutover'), 'reason should name the unmet dep');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── fixture: truly clean run ─────────────────────────────────────────────────
-
-test('Truly clean run (no error markers) → clean/null', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '20-clean-feature';
-    const logEvents = [
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_clean_001',
-            name: 'Bash',
-            input: { command: 'npm test', description: 'Run test suite' },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_clean_001',
-            content: '✓ 42 tests passed (1.3s)\nAll tests passed.',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'Feature complete. All acceptance criteria satisfied.\nSCHEDULER_VERDICT: PASS',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(tmp, slug, '# Clean Feature\n\nImplement a clean feature.');
-
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-
-    assert.equal(verdict.verdict, 'clean', `expected clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── self-recovery: FAIL followed by successful retry ──────────────────────
-
-test('FAIL recovered within 30 events → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '21-self-recovering';
-    const desc = 'Run pytest';
-    const logEvents = [
-      // First attempt: fails
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r1',
-            name: 'Bash',
-            input: { command: 'pytest', description: desc },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r1',
-            content: 'FAILED tests/test_foo.py::test_bar - AssertionError',
-            is_error: false,
-          }],
-        },
-      },
-      // Agent fixes and retries with same description
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r2',
-            name: 'Bash',
-            input: { command: 'pytest', description: desc },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r2',
-            content: '5 passed in 0.42s',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'Fixed the failing test and verified all pass.\nSCHEDULER_VERDICT: PASS',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(tmp, slug, '# Self-recovering test');
-
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-
-    assert.equal(verdict.verdict, 'clean', `self-recovery should yield clean, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── self-recovery: sleep-prefix normalization (RCA 745-pr188-ci-lint-docs-integrity) ──
-
-/**
- * Build N innocuous tool_use/tool_result pairs, used to push a later error
- * event into the last-20%-of-transcript window that the is_error scan
- * requires (parseLog counts every tool_use/tool_result as its own event).
- */
-function paddingEvents(n) {
-  const out = [];
-  for (let i = 0; i < n; i++) {
-    out.push({
-      type: 'assistant',
-      message: { role: 'assistant', content: [{ type: 'tool_use', id: `toolu_pad${i}`, name: 'Bash', input: { command: 'echo ok', description: `padding step ${i}` } }] },
-    });
-    out.push({
-      type: 'user',
-      message: { role: 'user', content: [{ type: 'tool_result', tool_use_id: `toolu_pad${i}`, content: 'ok', is_error: false }] },
-    });
-  }
-  return out;
-}
-
-test('sleep-prefixed FAIL recovered by a bare retry of the same command → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-sleep-prefix-recovery';
-    const logEvents = [
-      ...paddingEvents(8),
-      // First attempt: polls too early with a `sleep 20 &&` prefix, gets a
-      // pending/error result.
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r1',
-            name: 'Bash',
-            input: {
-              command: 'sleep 20 && gh pr checks 188 --repo midt-bg/sigma 2>&1',
-              description: 'sleep 20 && gh pr checks 188 --repo midt-bg/sigma',
-            },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r1',
-            content: 'Exit code 8\ncheck\tpending\t0\thttps://github.com/midt-bg/sigma/actions/runs/1/job/1',
-            is_error: true,
-          }],
-        },
-      },
-      // Later retry: same underlying command, no sleep prefix, succeeds.
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r2',
-            name: 'Bash',
-            input: {
-              command: 'gh pr checks 188 --repo midt-bg/sigma 2>&1',
-              description: 'gh pr checks 188 --repo midt-bg/sigma',
-            },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r2',
-            content: 'check\tpass\t2m16s\thttps://github.com/midt-bg/sigma/actions/runs/1/job/1',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'CI is green.\nSCHEDULER_VERDICT: PASS',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(tmp, slug, '# Sleep-prefix recovery test');
-
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-
-    assert.equal(verdict.verdict, 'clean', `sleep-prefix recovery should yield clean, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('normalizeDescForRecovery only strips the sleep wrapper, not unrelated text', () => {
-  const { normalizeDescForRecovery } = require('../runVerify.cjs');
-  assert.equal(
-    normalizeDescForRecovery('sleep 20 && gh pr checks 188 --repo midt-bg/sigma'),
-    'gh pr checks 188 --repo midt-bg/sigma',
-  );
-  assert.equal(
-    normalizeDescForRecovery('sleep 90; gh pr checks 188 --repo midt-bg/sigma'),
-    'gh pr checks 188 --repo midt-bg/sigma',
-  );
-  assert.equal(
-    normalizeDescForRecovery('gh pr checks 188 --repo midt-bg/sigma'),
-    'gh pr checks 188 --repo midt-bg/sigma',
-  );
-});
-
-test('two different commands, one sleep-prefixed, are not paired as recovered', () => {
-  // Exercise isSelfRecovered directly (via parseLog) rather than the full
-  // verifyRun pipeline: a SCHEDULER_VERDICT: PASS + landed commit legitimately
-  // overrides a transcript_errors finding at the verifyRun layer by design
-  // (see verifyRun's sentinel-override comment), which would mask the
-  // narrower claim this test makes — that the sleep-prefix strip alone must
-  // not pair two genuinely different commands.
-  const { isSelfRecovered, parseLog } = require('../runVerify.cjs');
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-sleep-prefix-no-false-pair';
-    const logEvents = [
-      // First attempt: sleep-prefixed poll of PR 188, fails.
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r1',
-            name: 'Bash',
-            input: {
-              command: 'sleep 20 && gh pr checks 188 --repo midt-bg/sigma 2>&1',
-              description: 'sleep 20 && gh pr checks 188 --repo midt-bg/sigma',
-            },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r1',
-            content: 'Exit code 8\ncheck\tpending\t0\thttps://github.com/midt-bg/sigma/actions/runs/1/job/1',
-            is_error: true,
-          }],
-        },
-      },
-      // "Retry" is actually a different command (different PR number) —
-      // after stripping the sleep wrapper it must NOT be treated as the
-      // same underlying command.
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_r2',
-            name: 'Bash',
-            input: {
-              command: 'gh pr checks 999 --repo midt-bg/sigma 2>&1',
-              description: 'gh pr checks 999 --repo midt-bg/sigma',
-            },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_r2',
-            content: 'check\tpass\t2m16s\thttps://github.com/midt-bg/sigma/actions/runs/2/job/2',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'Done.\nSCHEDULER_VERDICT: PASS',
-      },
-    ];
-
-    writeLog(tmp, slug, logEvents);
-    const { events } = parseLog(path.join(tmp, `${slug}.log`));
-    const failedToolUse = events.find((ev) => ev.kind === 'tool_use' && ev.toolUseId === 'toolu_r1');
-    assert.ok(failedToolUse, 'fixture should contain the failing tool_use event');
-
-    const recovered = isSelfRecovered(events, failedToolUse.seq, failedToolUse.description);
-    assert.equal(recovered, false, 'a differently-numbered PR check must not be treated as a recovered retry of the sleep-prefixed one');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── fixtures: feedback 2026-06-10-01 — quoted-error false positives ─────────
-
-/** Build a one-Bash-call log: tool_use → tool_result(content) → result event. */
-function bashRunEvents(content, { toolName = 'Bash', resultSubtype = 'success' } = {}) {
-  return [
-    {
-      type: 'assistant',
-      message: {
-        role: 'assistant',
-        content: [{
-          type: 'tool_use',
-          id: 'toolu_fb_001',
-          name: toolName,
-          input: { command: 'run something', description: 'Run checks' },
-        }],
-      },
-    },
-    {
-      type: 'user',
-      message: {
-        role: 'user',
-        content: [{
-          type: 'tool_result',
-          tool_use_id: 'toolu_fb_001',
-          content,
-          is_error: false,
-        }],
-      },
-    },
-    { type: 'result', subtype: resultSubtype, result: 'All acceptance criteria verified.\nSCHEDULER_VERDICT: PASS' },
-  ];
-}
-
-test('feedback 01: reviewer prose mentioning ImportError mid-sentence → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '25-self-pipeline-driver-hardening';
-    const prose = [
-      '{"finding": "the script will crash with ImportError rather than a clean RESULT: failed line",',
-      ' "note": "a missing dep would fail with an ImportError that is captured only in cron.log"}',
-    ].join('\n');
-    writeLog(tmp, slug, bashRunEvents(prose));
-    const prdPath = writePrd(tmp, slug, '# Hardening');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `prose mention must not flag, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-test('feedback 01: real line-anchored ModuleNotFoundError in a FAILED run → verify_unavailable/needs_review', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '25-real-import-error';
-    const out = [
-      '$ python run_checks.py',
-      'Traceback (most recent call last):',
-      '  File "run_checks.py", line 2, in <module>',
-      '    from playwright.sync_api import sync_playwright',
-      "ModuleNotFoundError: No module named 'playwright'",
-    ].join('\n');
-    // Run did NOT succeed: the missing dependency was never resolved, so the
-    // "couldn't verify" signal must still escalate to a human.
-    writeLog(tmp, slug, bashRunEvents(out, { resultSubtype: 'error_during_execution' }));
-    const prdPath = writePrd(tmp, slug, '# Real failure');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [] });
-    assert.equal(verdict.verdict, 'verify_unavailable', `unresolved missing-dep must flag, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally { rmdir(tmp); }
-});
-
-// ─── feedback 2026-06-10 addendum — recovered env-probe false positives ──────
-//
-// Setup probes (interpreter/venv search) that surface ModuleNotFoundError but
-// the run still reaches result:success are the missing-dependency class, not a
-// real failure. They must NOT downgrade — only annotate. But a Traceback ending
-// in a real logic exception (KeyError/AssertionError) still hard-flags, even on
-// "success", preserving the 2026-05-23 false-PASS guard.
-
-test('addendum: Traceback→ModuleNotFoundError in a SUCCEEDED run → clean (annotated, not downgraded)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '26-self-billbot-shared-lib';
-    const probe = [
-      'Exit code 1',
-      'Traceback (most recent call last):',
-      '  File "/home/bilko/Self/.claude/skills/snopud-bill/download_bill.py", line 42, in <module>',
-      '    from playwright.sync_api import TimeoutError as PWTimeout',
-      "ModuleNotFoundError: No module named 'playwright'",
-    ].join('\n');
-    writeLog(tmp, slug, bashRunEvents(probe)); // resultSubtype defaults to success
-    const prdPath = writePrd(tmp, slug, '# Shared lib');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `recovered env probe must not downgrade, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-    assert.ok(Array.isArray(verdict.annotations) && verdict.annotations.length === 1, 'should record one annotation');
-    assert.equal(verdict.annotations[0].verdict, 'verify_unavailable');
-  } finally { rmdir(tmp); }
-});
-
-test('addendum: bare Import/ModuleNotFound probe (no traceback) in SUCCEEDED run → clean/annotated', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '26-self-parser-tests';
-    writeLog(tmp, slug, bashRunEvents("ModuleNotFoundError: No module named 'conftest'"));
-    const prdPath = writePrd(tmp, slug, '# Parser tests');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `got ${verdict.verdict}: ${verdict.reason}`);
-    assert.ok(Array.isArray(verdict.annotations) && verdict.annotations.length === 1);
-  } finally { rmdir(tmp); }
-});
-
-test('addendum: Traceback→KeyError (real logic failure) on "success" → still transcript_errors/needs_review', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '26-real-logic-failure';
-    const out = [
-      '=== contract.json panels.sentiment ===',
-      'Traceback (most recent call last):',
-      '  File "<string>", line 1, in <module>',
-      "KeyError: 'panels.sentiment'",
-    ].join('\n');
-    writeLog(tmp, slug, bashRunEvents(out)); // success result — must NOT rescue a real failure
-    const prdPath = writePrd(tmp, slug, '# Logic failure');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [] });
-    assert.equal(verdict.verdict, 'transcript_errors', `real logic Traceback must still flag, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally { rmdir(tmp); }
-});
-
-test('feedback 01: error-shaped text inside a Task (subagent) result → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '25-subagent-quoting-error';
-    const reviewFinding = [
-      'Finding 3 (high): startup crash path:',
-      'Traceback (most recent call last):',
-      "ImportError: cannot import name 'foo'",
-      'This is what WOULD happen if the guard is removed.',
-    ].join('\n');
-    writeLog(tmp, slug, bashRunEvents(reviewFinding, { toolName: 'Task' }));
-    const prdPath = writePrd(tmp, slug, '# Review-heavy PRD');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `Task results must be exempt from pattern scan, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-test('feedback 01: quoted "Traceback..." line (leading quote) → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '25-quoted-traceback';
-    const prose = [
-      'The log format doc says:',
-      '  "Traceback (most recent call last):" appears first, then',
-      '  "SomeError: message" on a later line.',
-    ].join('\n');
-    writeLog(tmp, slug, bashRunEvents(prose));
-    const prdPath = writePrd(tmp, slug, '# Docs PRD');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `quoted traceback prose must not flag, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-// ─── harness tool errors (feedback follow-up 2026-06-10) ─────────────────────
-
-test('harness tool error (<tool_use_error>) in final 20% → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '58-harness-tool-error';
-    // Pad with benign events so the error lands in the final 20%, then a
-    // successful result — mirrors the real 58-web-remote-correctness-batch run.
-    const events = [];
-    for (let k = 0; k < 8; k++) {
-      events.push({ type: 'assistant', message: { role: 'assistant', content: [
-        { type: 'tool_use', id: `t${k}`, name: 'Read', input: { description: `read ${k}` } }] } });
-      events.push({ type: 'user', message: { role: 'user', content: [
-        { type: 'tool_result', tool_use_id: `t${k}`, content: 'ok', is_error: false }] } });
-    }
-    events.push({ type: 'assistant', message: { role: 'assistant', content: [
-      { type: 'tool_use', id: 'tbad', name: 'bash', input: { description: 'run tests' } }] } });
-    events.push({ type: 'user', message: { role: 'user', content: [
-      { type: 'tool_result', tool_use_id: 'tbad',
-        content: '<tool_use_error>Error: No such tool available: bash</tool_use_error>', is_error: true }] } });
-    events.push({ type: 'result', subtype: 'success', result: 'All acceptance criteria verified.\nSCHEDULER_VERDICT: PASS' });
-
-    writeLog(tmp, slug, events);
-    const prdPath = writePrd(tmp, slug, '# Correctness batch');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `harness tool error must not flag, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-// ─── subagent-internal tool errors (incident: 955-route-epic-create-through-ipc, 2026-08-03) ──
-
-test('subagent-internal is_error (parent_tool_use_id set) in final 20% → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '955-route-epic-create-through-ipc';
-    const events = [];
-    for (let k = 0; k < 8; k++) {
-      events.push({ type: 'assistant', message: { role: 'assistant', content: [
-        { type: 'tool_use', id: `t${k}`, name: 'Read', input: { description: `read ${k}` } }] } });
-      events.push({ type: 'user', message: { role: 'user', content: [
-        { type: 'tool_result', tool_use_id: `t${k}`, content: 'ok', is_error: false }] } });
-    }
-    // A code-reviewer Task subagent runs its own exploratory grep, which exits
-    // 1 (no matches) — this happened INSIDE the subagent, not the main agent.
-    events.push({
-      type: 'assistant',
-      parent_tool_use_id: 'toolu_task_001',
-      message: { role: 'assistant', content: [
-        { type: 'tool_use', id: 'tgrep', name: 'Bash', input: { description: 'grep for buildPromptSession' } }] },
-    });
-    events.push({
-      type: 'user',
-      parent_tool_use_id: 'toolu_task_001',
-      message: { role: 'user', content: [
-        { type: 'tool_result', tool_use_id: 'tgrep', content: '', is_error: true }] },
-    });
-    events.push({ type: 'result', subtype: 'success', result: 'All acceptance criteria verified.' });
-
-    writeLog(tmp, slug, events);
-    const prdPath = writePrd(tmp, slug, '# Route Epic create through IPC');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'clean', `subagent-internal is_error must not flag, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-test('main-agent is_error (no parent_tool_use_id) in final 20% → still transcript_errors', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '955-main-agent-error-guard';
-    const events = [];
-    for (let k = 0; k < 8; k++) {
-      events.push({ type: 'assistant', message: { role: 'assistant', content: [
-        { type: 'tool_use', id: `t${k}`, name: 'Read', input: { description: `read ${k}` } }] } });
-      events.push({ type: 'user', message: { role: 'user', content: [
-        { type: 'tool_result', tool_use_id: `t${k}`, content: 'ok', is_error: false }] } });
-    }
-    events.push({ type: 'assistant', message: { role: 'assistant', content: [
-      { type: 'tool_use', id: 'tbad', name: 'Bash', input: { description: 'run tests' } }] } });
-    events.push({ type: 'user', message: { role: 'user', content: [
-      { type: 'tool_result', tool_use_id: 'tbad', content: 'FAILED: 1 test failed', is_error: true }] } });
-    events.push({ type: 'result', subtype: 'success', result: 'All acceptance criteria verified.' });
-
-    writeLog(tmp, slug, events);
-    const prdPath = writePrd(tmp, slug, '# Guard: main-agent errors still flag');
-    const verdict = await verifyRun({ runDir: tmp, prdPath, queueEntry: { slug, status: 'running' }, allJobs: [], committedDuringRun: true });
-    assert.equal(verdict.verdict, 'transcript_errors', `main-agent is_error must still flag, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally { rmdir(tmp); }
-});
-
-test('parseLog populates parentToolUseId on tool_use and tool_result events, defaulting to null', () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '955-parselog-parent-id';
-    const events = [
-      {
-        type: 'assistant',
-        parent_tool_use_id: 'toolu_parent_001',
-        message: { role: 'assistant', content: [
-          { type: 'tool_use', id: 'tchild', name: 'Bash', input: { description: 'grep in subagent' } }] },
-      },
-      {
-        type: 'user',
-        parent_tool_use_id: 'toolu_parent_001',
-        message: { role: 'user', content: [
-          { type: 'tool_result', tool_use_id: 'tchild', content: '', is_error: true }] },
-      },
-      {
-        type: 'assistant',
-        message: { role: 'assistant', content: [
-          { type: 'tool_use', id: 'ttop', name: 'Bash', input: { description: 'run in main agent' } }] },
-      },
-      {
-        type: 'user',
-        message: { role: 'user', content: [
-          { type: 'tool_result', tool_use_id: 'ttop', content: 'ok', is_error: false }] },
-      },
-    ];
-    writeLog(tmp, slug, events);
-    const { events: parsed } = parseLog(path.join(tmp, `${slug}.log`));
-
-    const childUse = parsed.find((ev) => ev.kind === 'tool_use' && ev.toolUseId === 'tchild');
-    const childResult = parsed.find((ev) => ev.kind === 'tool_result' && ev.toolUseId === 'tchild');
-    const topUse = parsed.find((ev) => ev.kind === 'tool_use' && ev.toolUseId === 'ttop');
-    const topResult = parsed.find((ev) => ev.kind === 'tool_result' && ev.toolUseId === 'ttop');
-
-    assert.equal(childUse.parentToolUseId, 'toolu_parent_001');
-    assert.equal(childResult.parentToolUseId, 'toolu_parent_001');
-    assert.equal(topUse.parentToolUseId, null);
-    assert.equal(topResult.parentToolUseId, null);
-  } finally { rmdir(tmp); }
-});
-
-test('isHarnessToolError detects wrapper and "No such tool available"', () => {
-  const { isHarnessToolError } = require('../runVerify.cjs');
-  assert.equal(isHarnessToolError('<tool_use_error>Error: No such tool available: bash</tool_use_error>'), true);
-  assert.equal(isHarnessToolError('No such tool available: Foo'), true);
-  assert.equal(isHarnessToolError('ModuleNotFoundError: No module named x'), false);
-  assert.equal(isHarnessToolError(''), false);
-});
-
-// ─── SCHEDULER_VERDICT sentinel override tests ────────────────────────────────
-
-/** Build a log where a tool_result contains a Traceback+KeyError and the
- *  result event optionally contains the sentinel line. */
-function tracebackRunEvents(sentinelLine) {
-  const resultText = sentinelLine
-    ? `All work done.\n${sentinelLine}`
-    : 'All work done.';
-  return [
-    {
-      type: 'assistant',
-      message: {
-        role: 'assistant',
-        content: [{
-          type: 'tool_use',
-          id: 'toolu_tv_001',
-          name: 'Bash',
-          input: { command: 'pytest', description: 'Run acceptance tests' },
-        }],
-      },
-    },
-    {
-      type: 'user',
-      message: {
-        role: 'user',
-        content: [{
-          type: 'tool_result',
-          tool_use_id: 'toolu_tv_001',
-          content: [
-            '=== TDD red phase ===',
-            'Traceback (most recent call last):',
-            '  File "test_foo.py", line 5, in test_bar',
-            "KeyError: 'missing_key'",
-          ].join('\n'),
-          is_error: false,
-        }],
-      },
-    },
-    { type: 'result', subtype: 'success', result: resultText },
-  ];
-}
-
-// (a) sentinel PASS + committedDuringRun:true + Traceback+Error → clean
-test('sentinel PASS + committedDuringRun:true + Traceback → clean (override)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '77-sentinel-override-pass';
-    writeLog(tmp, slug, tracebackRunEvents('SCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Sentinel override test');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'clean', `expected clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-    assert.ok(verdict.reason.includes('SCHEDULER_VERDICT: PASS'), `reason should mention sentinel: ${verdict.reason}`);
-    // Sidecar should record the sentinel and override
-    const sidecar = JSON.parse(fs.readFileSync(path.join(tmp, `${slug}.verdicts.json`), 'utf8'));
-    assert.equal(sidecar.sentinel, 'pass');
-    assert.ok(sidecar.sentinelOverride, 'sidecar should record sentinelOverride');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (b) no sentinel + Traceback+Error → transcript_errors (unchanged baseline)
-test('no sentinel + Traceback+Error + committedDuringRun:true → transcript_errors (no override)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '77-no-sentinel-baseline';
-    writeLog(tmp, slug, tracebackRunEvents(null));
-    const prdPath = writePrd(tmp, slug, '# No sentinel baseline');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `expected transcript_errors without sentinel, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (c) sentinel PASS + committedDuringRun:false → stays transcript_errors (commit not confirmed)
-test('sentinel PASS + committedDuringRun:false → transcript_errors (commit unconfirmed)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '77-sentinel-no-commit';
-    writeLog(tmp, slug, tracebackRunEvents('SCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Sentinel without commit');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `PASS without commit must not override, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (d) sentinel FAIL → never clean (even with committedDuringRun:true)
-test('sentinel FAIL + committedDuringRun:true → transcript_errors (FAIL never overrides)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '77-sentinel-fail';
-    writeLog(tmp, slug, tracebackRunEvents('SCHEDULER_VERDICT: FAIL AC gate was red'));
-    const prdPath = writePrd(tmp, slug, '# Sentinel FAIL');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `FAIL sentinel must not override to clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── pre-sentinel heal (allowPreSentinelHeal) ─────────────────────────────────
-
-// (f) allowPreSentinelHeal=true + committed + no sentinel → clean
-test('pre-sentinel heal: committed + no sentinel + allowPreSentinelHeal → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '86-pre-sentinel-heal-pass';
-    writeLog(tmp, slug, tracebackRunEvents(null)); // no SCHEDULER_VERDICT line
-    const prdPath = writePrd(tmp, slug, '# Pre-sentinel heal');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'needs_review' },
-      allJobs: [],
-      committedDuringRun: true,
-      allowPreSentinelHeal: true,
-    });
-    assert.equal(verdict.verdict, 'clean', `expected clean via pre-sentinel heal, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-    assert.ok(verdict.reason.includes('pre-sentinel heal'), `reason should mention pre-sentinel heal: ${verdict.reason}`);
-    const sidecar = JSON.parse(fs.readFileSync(path.join(tmp, `${slug}.verdicts.json`), 'utf8'));
-    assert.ok(sidecar.preSentinelHeal, 'sidecar should record preSentinelHeal');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (g) allowPreSentinelHeal=true + committed + FAIL sentinel → stays transcript_errors
-test('pre-sentinel heal: committed + SCHEDULER_VERDICT: FAIL + allowPreSentinelHeal → transcript_errors', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '86-pre-sentinel-heal-fail';
-    writeLog(tmp, slug, tracebackRunEvents('SCHEDULER_VERDICT: FAIL AC gate was red'));
-    const prdPath = writePrd(tmp, slug, '# Pre-sentinel heal blocked by FAIL');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'needs_review' },
-      allJobs: [],
-      committedDuringRun: true,
-      allowPreSentinelHeal: true,
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `FAIL sentinel must block pre-sentinel heal, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (h) allowPreSentinelHeal=true + NOT committed + no sentinel → stays transcript_errors
-test('pre-sentinel heal: not committed + no sentinel + allowPreSentinelHeal → transcript_errors', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '86-pre-sentinel-heal-uncommitted';
-    writeLog(tmp, slug, tracebackRunEvents(null));
-    const prdPath = writePrd(tmp, slug, '# Pre-sentinel heal blocked by no commit');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'needs_review' },
-      allJobs: [],
-      committedDuringRun: false,
-      allowPreSentinelHeal: true,
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `no commit must block pre-sentinel heal, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (i) allowPreSentinelHeal=false (default) + committed + no sentinel → baseline unchanged
-test('no sentinel + committed + allowPreSentinelHeal=false (default) → transcript_errors', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '86-pre-sentinel-heal-disabled';
-    writeLog(tmp, slug, tracebackRunEvents(null));
-    const prdPath = writePrd(tmp, slug, '# allowPreSentinelHeal disabled by default');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'needs_review' },
-      allJobs: [],
-      committedDuringRun: true,
-      // allowPreSentinelHeal defaults to false
-    });
-    assert.equal(verdict.verdict, 'transcript_errors', `must not heal without allowPreSentinelHeal, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── no_verdict_sentinel: no commit + no sentinel is not "clean" by default ──
-
-/** A run with zero tool_result pattern hits — e.g. the agent asked a
- *  clarifying question and stopped, ending its turn without touching tools. */
-function noOpRunEvents(resultText) {
-  return [
-    { type: 'result', subtype: 'success', result: resultText },
-  ];
-}
-
-test('no_verdict_sentinel: no tool_result issues, no sentinel, no commit → needs_review', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '406-capture-panel-ui';
-    writeLog(tmp, slug, noOpRunEvents('I need clarification on which capture API to target before proceeding.'));
-    const prdPath = writePrd(tmp, slug, '# Browser capture panel UI');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'no_verdict_sentinel', `expected no_verdict_sentinel, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('no_verdict_sentinel guard: same run but committedDuringRun:true → stays clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '406-capture-panel-ui-committed';
-    writeLog(tmp, slug, noOpRunEvents('I need clarification on which capture API to target before proceeding.'));
-    const prdPath = writePrd(tmp, slug, '# Browser capture panel UI');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'clean', `commit landed should stay clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('no_verdict_sentinel guard: same run but with SCHEDULER_VERDICT: PASS sentinel + a commit → stays clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '406-capture-panel-ui-sentinel';
-    writeLog(tmp, slug, noOpRunEvents('All acceptance criteria verified.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Browser capture panel UI');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'clean', `PASS sentinel + commit should stay clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── abandoned_background_task: auto-backgrounded Bash + no sentinel/commit ─
-
-/** A run whose Bash command hit the harness's 120s foreground timeout and was
- *  auto-backgrounded, then the run ended without ever coming back to commit
- *  or emit the finish-protocol sentinel. */
-function abandonedBackgroundTaskRunEvents(resultText) {
-  return [
-    {
-      type: 'assistant',
-      message: {
-        role: 'assistant',
-        content: [{
-          type: 'tool_use',
-          id: 'toolu_bg_001',
-          name: 'Bash',
-          input: { command: 'npm run test:unit', description: 'Run full unit suite' },
-        }],
-      },
-    },
-    {
-      type: 'user',
-      message: {
-        role: 'user',
-        content: [{
-          type: 'tool_result',
-          tool_use_id: 'toolu_bg_001',
-          content: 'Command did not complete within its 120s timeout and was moved to the background (ID: bcn1lpvpd). Output is being written to: /tmp/claude-1000/tasks/bcn1lpvpd.output. You will be notified when it completes. To check interim output, use Read on that file path.',
-          is_error: false,
-        }],
-      },
-    },
-    { type: 'result', subtype: 'success', result: resultText },
-  ];
-}
-
-test('abandoned_background_task: background marker + no sentinel + no commit → abandoned_background_task', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '1116-encyclopedia-animations-section';
-    writeLog(tmp, slug, abandonedBackgroundTaskRunEvents('Waiting for the backgrounded test run to finish.'));
-    const prdPath = writePrd(tmp, slug, '# Encyclopedia animations section');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'abandoned_background_task', `expected abandoned_background_task, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('abandoned_background_task guard: no background marker, no sentinel, no commit → stays plain no_verdict_sentinel', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '1116-no-background-marker';
-    writeLog(tmp, slug, noOpRunEvents('I need clarification on which capture API to target before proceeding.'));
-    const prdPath = writePrd(tmp, slug, '# Encyclopedia animations section');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'no_verdict_sentinel', `expected plain no_verdict_sentinel, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('abandoned_background_task guard: background marker present but run DID commit → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '1116-background-marker-but-committed';
-    writeLog(tmp, slug, abandonedBackgroundTaskRunEvents('Backgrounded test run finished green; work committed.'));
-    const prdPath = writePrd(tmp, slug, '# Encyclopedia animations section');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'clean', `background marker with a landed commit should stay clean, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── pass_no_commit: PASS sentinel with no commit is not "clean" ───────────
-
-test('pass_no_commit: SCHEDULER_VERDICT: PASS but no commit landed → needs_review', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '511-recorder-mouse-drag-panel-export';
-    writeLog(tmp, slug, noOpRunEvents('All acceptance criteria verified.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Recorder mouse drag panel export');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('pass_no_commit: fix-plan job (slug ^\\d+-fix-) with PASS but no commit is exempt → clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '523-fix-bounded-fix-plan-retry';
-    writeLog(tmp, slug, noOpRunEvents('Verification complete — no code gap exists.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Fix: bounded fix-plan retry');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'clean', `expected clean (fix-plan exemption), got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── pass_no_commit_already_shipped: original PRD re-run after landing ─────
-// (incident: 655-needs-review-rca-feedback-hook, 2026-07-31)
-
-const { extractPrdDeliverablePaths, allDeliverablesAlreadyTracked } = require('../runVerify.cjs');
-const REPO_ROOT = path.resolve(__dirname, '../../..');
-
-test('pass_no_commit: original PRD, PASS no commit, all deliverable paths already tracked in git → pass_no_commit_already_shipped', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '655-needs-review-rca-feedback-hook';
-    writeLog(tmp, slug, noOpRunEvents('Verified rcaReport.cjs already shipped in e13168d; typecheck + tests green; no commit needed.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'Deliverables: `src/main/lib/rcaReport.cjs` and `src/main/__tests__/rcaReport.test.cjs`.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit_already_shipped', `expected already-shipped exemption, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('pass_no_commit: original PRD, PASS no commit, one deliverable path untracked → falls through to pass_no_commit', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '655-needs-review-rca-feedback-hook';
-    writeLog(tmp, slug, noOpRunEvents('Nothing to do here.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'Deliverables: `src/main/lib/rcaReport.cjs` and `src/main/lib/does-not-exist-xyz.cjs`.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected fallback to pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('pass_no_commit: original PRD, PASS no commit, zero extractable deliverable paths → falls through to pass_no_commit', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '999-some-vague-prd';
-    writeLog(tmp, slug, noOpRunEvents('Nothing to do here.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'This PRD names no file paths at all, just prose.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected fallback to pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (incident: 812-commit-guard-retry-on-worktree-ref-visibility-delay — a
-// stale scheduler process, booted before the already-shipped exemption
-// landed, produced a false pass_no_commit for exactly this shape of re-run)
-test('pass_no_commit: commit-guard-retry PRD naming scheduler.cjs + its committed-in-window test, PASS no commit → pass_no_commit_already_shipped', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-commit-guard-retry-on-worktree-ref-visibility-delay';
-    writeLog(tmp, slug, noOpRunEvents('Retry + tests already present; nothing to change.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(
-      tmp,
-      slug,
-      'Deliverables: `src/main/scheduler.cjs` and `src/main/__tests__/scheduler-committed-in-window.test.cjs`.',
-    );
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit_already_shipped', `expected already-shipped exemption, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('extractPrdDeliverablePaths: extracts backticked repo-relative paths, ignores prose', () => {
-  const body = 'Deliverables: `src/main/lib/rcaReport.cjs` and `src/main/__tests__/rcaReport.test.cjs`. Also see node_modules/foo/bar.js which should be excluded.';
-  const paths = extractPrdDeliverablePaths(body);
-  assert.deepEqual(paths, ['src/main/lib/rcaReport.cjs', 'src/main/__tests__/rcaReport.test.cjs']);
-  assert.deepEqual(extractPrdDeliverablePaths(''), []);
-  assert.deepEqual(extractPrdDeliverablePaths(undefined), []);
-});
-
-test('allDeliverablesAlreadyTracked: fail-safe on empty paths, non-git cwd, and git errors', () => {
-  assert.equal(allDeliverablesAlreadyTracked({ cwd: REPO_ROOT, paths: [] }), false);
-  assert.equal(allDeliverablesAlreadyTracked({
-    cwd: REPO_ROOT,
-    paths: ['src/main/runVerify.cjs'],
-    execImpl: () => { throw new Error('boom'); },
-  }), false);
-  assert.equal(allDeliverablesAlreadyTracked({
-    cwd: REPO_ROOT,
-    paths: ['src/main/runVerify.cjs'],
-    execImpl: () => '',
-  }), true);
-});
-
-// ─── repo-root config deliverable extraction ────────────────────────────────
-// (incident: 816-vitest-register-runverify-test, 2026-07-31 — PRD named only
-// `vitest.config.ts` as its deliverable; the source-dir-only regex extracted
-// nothing, so allDeliverablesAlreadyTracked early-returned false even though
-// the config edit was already tracked in git)
-
-test('extractPrdDeliverablePaths: extracts a root-config file alongside a src path', () => {
-  const body = 'Register the test in `vitest.config.ts` and port `src/main/__tests__/runVerify.test.cjs` onto vitest.';
-  const paths = extractPrdDeliverablePaths(body);
-  assert.deepEqual(new Set(paths), new Set(['vitest.config.ts', 'src/main/__tests__/runVerify.test.cjs']));
-});
-
-test('extractPrdDeliverablePaths: root-config-only PRD body yields a non-empty result', () => {
-  const body = 'The only deliverable is `vitest.config.ts` — add the missing test.include entry.';
-  const paths = extractPrdDeliverablePaths(body);
-  assert.deepEqual(paths, ['vitest.config.ts']);
-});
-
-test('extractPrdDeliverablePaths: nested or non-allowlisted root files do not produce spurious root-config entries', () => {
-  const body = 'See `some/nested/package.json` for the dependency, and update README.md with notes.';
-  const paths = extractPrdDeliverablePaths(body);
-  assert.equal(paths.includes('package.json'), false);
-  assert.equal(paths.includes('README.md'), false);
-  assert.deepEqual(paths, []);
-});
-
-test('pass_no_commit: root-config-only PRD, PASS no commit, config already tracked → pass_no_commit_already_shipped', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '816-vitest-register-runverify-test';
-    writeLog(tmp, slug, noOpRunEvents('Registered already in vitest.config.ts; 48 tests green.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'Deliverable: register the test file in `vitest.config.ts`.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit_already_shipped', `expected already-shipped exemption, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── pass_no_commit_prior_run_verified: re-fired slug whose PRIOR run landed ─
-// (incident: 812-workbench-review-nits-cleanup, 2026-07-31)
-
-const { isAncestorCommit } = require('../runVerify.cjs');
-const REPO_ROOT_COMMIT = execFileSync('git', ['rev-list', '--max-parents=0', 'HEAD'], {
-  cwd: REPO_ROOT, encoding: 'utf8',
-}).trim().split('\n')[0];
-
-test('pass_no_commit: PASS, no commit, priorLandedCommit is an ancestor of HEAD → pass_no_commit_prior_run_verified', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-workbench-review-nits-cleanup';
-    writeLog(tmp, slug, noOpRunEvents('Re-checked every AC item; all already implemented by an earlier run.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'This PRD names no deliverable paths, just prose.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-      priorLandedCommit: REPO_ROOT_COMMIT,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit_prior_run_verified', `expected prior-run exemption, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-    assert.equal(verdict.priorLandedCommit, REPO_ROOT_COMMIT);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('pass_no_commit: PASS, no commit, priorLandedCommit is NOT an ancestor of HEAD → pass_no_commit', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-workbench-review-nits-cleanup';
-    writeLog(tmp, slug, noOpRunEvents('Nothing to do here.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'This PRD names no deliverable paths, just prose.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-      priorLandedCommit: '0000000000000000000000000000000000000000',
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected fallback to pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-test('isAncestorCommit: fail-safe on missing sha and on a throwing execImpl (git error)', () => {
-  assert.equal(isAncestorCommit({ cwd: REPO_ROOT, sha: null }), false);
-  assert.equal(isAncestorCommit({
-    cwd: REPO_ROOT,
-    sha: REPO_ROOT_COMMIT,
-    execImpl: () => { throw new Error('git: command not found'); },
-  }), false);
-  assert.equal(isAncestorCommit({
-    cwd: REPO_ROOT,
-    sha: REPO_ROOT_COMMIT,
-    execImpl: () => '',
-  }), true);
-});
-
-test('pass_no_commit: PASS, no commit, priorLandedCommit not supplied (null) → pass_no_commit (regression guard)', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '812-workbench-review-nits-cleanup';
-    writeLog(tmp, slug, noOpRunEvents('Nothing to do here.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, 'This PRD names no deliverable paths, just prose.');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: REPO_ROOT },
-      allJobs: [],
-      committedDuringRun: false,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected pass_no_commit with no priorLandedCommit, got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (e) halt + sentinel PASS → still halt (override must not apply to halt)
-test('halt result + sentinel PASS + committedDuringRun:true → still halt', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '77-halt-sentinel-pass';
-    const logEvents = [
-      {
-        type: 'assistant',
-        message: {
-          role: 'assistant',
-          content: [{
-            type: 'tool_use',
-            id: 'toolu_halt_001',
-            name: 'Bash',
-            input: { command: 'check deps', description: 'Check prerequisites' },
-          }],
-        },
-      },
-      {
-        type: 'user',
-        message: {
-          role: 'user',
-          content: [{
-            type: 'tool_result',
-            tool_use_id: 'toolu_halt_001',
-            content: 'dep not ready',
-            is_error: false,
-          }],
-        },
-      },
-      {
-        type: 'result',
-        subtype: 'success',
-        result: 'HALT: prerequisite not met\nSCHEDULER_VERDICT: PASS',
-      },
-    ];
-    writeLog(tmp, slug, logEvents);
-    const prdPath = writePrd(tmp, slug, '# Halt with sentinel');
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running' },
-      allJobs: [],
-      committedDuringRun: true,
-    });
-    assert.equal(verdict.verdict, 'halt', `halt must survive even with PASS sentinel, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'pending');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// ─── -merge-main postcondition exemption tests (2026-07-18 feedback) ──────────
-
-const { isMergeMainSlug, extractMergeMainPrNumber } = require('../runVerify.cjs');
-
-test('isMergeMainSlug matches the NN-prXXX-merge-main and NN-fix-prXXX-merge-main conventions', () => {
-  assert.equal(isMergeMainSlug('565-pr188-merge-main'), true);
-  assert.equal(isMergeMainSlug('571-fix-pr141-merge-main'), true);
-  assert.equal(isMergeMainSlug('523-fix-bounded-fix-plan-retry'), false);
-  assert.equal(isMergeMainSlug('406-capture-panel-ui'), false);
-  assert.equal(isMergeMainSlug(''), false);
-  assert.equal(isMergeMainSlug(undefined), false);
-});
-
-test('extractMergeMainPrNumber prefers "PR #<n>" in text, falls back to slug', () => {
-  assert.equal(extractMergeMainPrNumber('565-pr188-merge-main', 'title: Merge current main into sigma PR #188 (round 2)'), 188);
-  assert.equal(extractMergeMainPrNumber('565-pr188-merge-main', 'no pr mention here'), 188);
-  assert.equal(extractMergeMainPrNumber('non-standard-slug', 'no pr mention here'), null);
-});
-
-// (1) -merge-main + PASS + no commit + gh reports MERGEABLE/non-CONFLICTING → verified, no issue
-test('merge-main exemption: gh reports MERGEABLE → pass_no_commit_target_verified, no issue pushed', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '565-pr188-merge-main';
-    writeLog(tmp, slug, noOpRunEvents('PR #188 was already merged into main by another actor.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Merge current main into sigma PR #188');
-    let ghCalls = 0;
-    const ghExecImpl = (cmd, args, opts) => {
-      ghCalls++;
-      assert.equal(cmd, 'gh');
-      assert.deepEqual(args, ['pr', 'view', '188', '--json', 'mergeable,mergeStateStatus']);
-      assert.equal(opts.cwd, '/tmp/sigma');
-      return JSON.stringify({ mergeable: 'MERGEABLE', mergeStateStatus: 'BEHIND' });
-    };
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: '/tmp/sigma' },
-      allJobs: [],
-      committedDuringRun: false,
-      ghExecImpl,
-    });
-    assert.equal(ghCalls, 1);
-    assert.equal(verdict.verdict, 'pass_no_commit_target_verified', `expected verified verdict, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, null);
-    assert.equal(verdict.verifiedPrNumber, 188);
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (2) -merge-main + PASS + no commit + gh reports CONFLICTING → falls through to pass_no_commit
-test('merge-main exemption: gh reports CONFLICTING → falls through to pass_no_commit', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '565-pr188-merge-main';
-    writeLog(tmp, slug, noOpRunEvents('PR #188 needs no further work.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Merge current main into sigma PR #188');
-    const ghExecImpl = () => JSON.stringify({ mergeable: 'MERGEABLE', mergeStateStatus: 'CONFLICTING' });
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: '/tmp/sigma' },
-      allJobs: [],
-      committedDuringRun: false,
-      ghExecImpl,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected fallback to pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (3) -merge-main + PASS + no commit + gh call fails/times out → falls through to pass_no_commit, never throws
-test('merge-main exemption: gh call throws → falls through to pass_no_commit, never throws', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '565-pr188-merge-main';
-    writeLog(tmp, slug, noOpRunEvents('PR #188 needs no further work.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Merge current main into sigma PR #188');
-    const ghExecImpl = () => { throw new Error('gh: command not found'); };
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: '/tmp/sigma' },
-      allJobs: [],
-      committedDuringRun: false,
-      ghExecImpl,
-    });
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected fallback to pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (4) non-merge-main slug, PASS + no commit → unchanged pass_no_commit behavior, zero gh invocations
-test('merge-main exemption: non-merge-main slug never attempts a gh call', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '406-capture-panel-ui';
-    writeLog(tmp, slug, noOpRunEvents('All acceptance criteria verified.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Browser capture panel UI');
-    let ghCalls = 0;
-    const ghExecImpl = () => { ghCalls++; return JSON.stringify({ mergeable: 'MERGEABLE', mergeStateStatus: 'CLEAN' }); };
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: '/tmp/proj' },
-      allJobs: [],
-      committedDuringRun: false,
-      ghExecImpl,
-    });
-    assert.equal(ghCalls, 0, 'gh must never be invoked for a non-merge-main slug');
-    assert.equal(verdict.verdict, 'pass_no_commit', `expected unchanged pass_no_commit, got ${verdict.verdict}: ${verdict.reason}`);
-    assert.equal(verdict.downgradeTo, 'needs_review');
-  } finally {
-    rmdir(tmp);
-  }
-});
-
-// (5) -merge-main slug but a real commit DID land → unaffected, path unreachable, zero gh invocations
-test('merge-main exemption: commit landed → path unreachable, zero gh invocations, verdict clean', async () => {
-  const tmp = makeTmpDir();
-  try {
-    const slug = '565-pr188-merge-main';
-    writeLog(tmp, slug, noOpRunEvents('Merged and pushed.\nSCHEDULER_VERDICT: PASS'));
-    const prdPath = writePrd(tmp, slug, '# Merge current main into sigma PR #188');
-    let ghCalls = 0;
-    const ghExecImpl = () => { ghCalls++; return JSON.stringify({ mergeable: 'MERGEABLE', mergeStateStatus: 'CLEAN' }); };
-    const verdict = await verifyRun({
-      runDir: tmp,
-      prdPath,
-      queueEntry: { slug, status: 'running', cwd: '/tmp/sigma' },
-      allJobs: [],
-      committedDuringRun: true,
-      ghExecImpl,
-    });
-    assert.equal(ghCalls, 0, 'gh must never be invoked when a commit already landed');
-    assert.equal(verdict.verdict, 'clean', `expected clean (commit landed, sentinel PASS), got ${verdict.verdict}: ${verdict.reason}`);
-  } finally {
-    rmdir(tmp);
-  }
-});