Merge remote-tracking branch 'origin/main' into capy/gsta-95-bun-config-windows

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com> # Conflicts: # CHANGELOG.md # bin/gstack-question-log
2026-11-06 08:54:04 +01:00 · 2026-10-25 23:58:12 +00:00
commit d77948c6cb
@@ -18,6 +18,42 @@
 ### Gate integrity
 - `test/bun-entry-isolation.test.ts` is strengthened. On POSIX the required shell spelling still resolves to `/dev/null`. The literal `--config=/dev/null` in a shipped shell file is now an offender, as is an undefined or redefined `GSTACK_BUN_CONFIG`. The shell-caller detector also skips a `--config` value before deciding whether a call starts a program, so `bun --no-env-file --config=<anything> -e` no longer passes unchecked; the tree has no such call. Negative controls: the old spelling, a variable without the default, an empty `--config=`, a missing, late or altered definition, and a reverted bin each fail the check. No threshold, budget, timeout, trial count or case kind changed.

+## [1.91.223.0] - 2026-10-25
+
+**/plan-tune's inferred profile finally moves: a clicked answer is logged as the question's registered option key, not its label.** #2623 showed every inferred dimension stuck at 0.5 on a 1,286-event log: the AskUserQuestion hook recorded labels like "A) Fix it now (recommended)", while the signal map that turns answers into dimension nudges is keyed by registered option keys like `fix-now`, so 0 of 1,286 answers matched. Now five hook-captured label answers to the registered architecture-finding question move `architecture_care` from 0.5 to 0.632, and `--trace` lists all five.
+
+### Fixed
+- **`bin/gstack-question-log`**: on a hook-captured event (`source: hook`) for a registered `question_id`, the clicked label is mapped to that question's registered option key before it is written, and the label is kept in a new `choice_label` field. A label in `recommended` is mapped the same way, so `followed_recommendation` compares keys. Ad-hoc and `hook-<hash>` ids and unmatched or ambiguous labels are written as given. Rows from other writers (the agent's own log call, auto-decide, Codex import, typed free text) are unchanged, because readers such as the plan-ceo-review auto-decide check match the full mode names those rows carry. Existing logs are not rewritten.
+- **`scripts/psychographic-signals.ts`**: new `resolveOptionKey(question_id, label)`. It strips the `A)` selector and `(recommended)` suffix, then matches the label's leading words against the question's own options only: each key's own words (`fix-now` → "fix now") plus a short cue list ("Add to TODOs" → `defer`, "Ship as-is" → `skip`, "HOLD SCOPE" → `hold`). The longest cue wins and a tie stays unmapped.
+
+### Tests
+- `test/plan-tune.test.ts`: every registered option resolves from its own words, bare and as an `A) … (recommended)` label; 20 common phrasings, including the four plan-ceo-review mode names, map to their keys; negative controls (unregistered ids, a question with no options, a label naming no option, a mid-sentence cue, another question's key, a number that looks like a selector) stay unmapped. Two capture tests run the real bin: hook labels become keys with `choice_label` kept, while agent rows (including a `HOLD SCOPE` auto-decide row), ad-hoc ids, unmatched labels and free text are untouched, and a picked label follows a key recommendation. Both capture tests fail against the old bin.
+
+### Gate integrity
+- No test expectation, threshold, budget, timeout, trial count or case kind changed. Existing tests that pin label rows from the agent, auto-decide and Codex import paths pass unchanged.
+
+## [1.91.222.0] - 2026-10-25
+
+**Five flaky tests now fail only for real defects, and a judge reply that isn't valid JSON says so in its failure line (daily flaky-fix pass, GSTA-91).** `context-restore-fragment-match` failed 0/3 on main cb99df4, although every run restored the right checkpoint: to find the checkpoint whose title matches, the agent often prints all three, and the case failed whenever another checkpoint's body appeared anywhere in the transcript. `test/gstack-ff-update.test.ts` counted every `sleep 300` on the machine, so heartbeats from parallel test files changed its count (remote pregate at #3412's head: expected "1", received "3"). The `GSTACK_CAPY=1` install-admission case in `test/gstack-skill-start.test.ts` makes ten `gstack-skill-start` spawns under bun's 5 s default timeout and took 3.6–5.7 s on a 4-core machine. The QA-only fixture's late scenarios mocked a 1000 ms capture budget but ran each attempt's real setup inside it. The setup-browser-cookies workflow judge in train-4 gate 40173759210 returned JSON with an unescaped quote in its reasoning, and the flake record read only "JSON Parse error: Expected '}'".
+
+### Fixed
+- **`test/helpers/context-restore-fragment.ts`** (new): `fragmentRestoreVerdict` reads only the agent's report (top-level assistant text and the final result), never tool output. The report must carry the payments checkpoint's body and no other checkpoint's body. `test/skill-e2e-context-skills.test.ts` uses it for the fragment-match case.
+- **`test/helpers/llm-judge.ts`**: when a judge response is not valid JSON, `callJudge` throws a `SyntaxError` that names the judge output, its stop reason and length, and the retained `llm-judge-response-parse-error` event. It still fails closed and parses no score; the request is unchanged.
+
+### Tests
+- `test/context-restore-fragment-verdict.test.ts` replays the three captured false reds from main (`test/fixtures/context-restore-fragment-match-captures.json`) as passes. Controls that must fail: captures rewritten to resume alpha or omega, synthetic runs that read and resume only alpha or omega, resume summaries that mix two checkpoints, and a payments body that appears only in tool output. Restoring the old any-byte surface fails all four positive controls. Focused paid runs of the case: 0/3 on main, 3/3 after (about $0.11 a trial).
+- `test/llm-judge-frontier.test.ts` replays the exact captured cookie-judge reply: it fails closed with no score, one request and the new diagnosis. It fails on the previous helper.
+- `test/gstack-ff-update.test.ts`: the reap case runs `gstack-ff-update` with a heartbeat interval unique to the test process and counts only sleepers of that interval, starting from zero. The old count failed with a loop starting unrelated `sleep 300` processes beside it (expected "5", received "14"); the new one passes, and still fails when the bin stops reaping its sleeper.
+- `test/qa-only-fixture.test.ts`: in the late scenarios the fixture clock stands still from the start of each attempt until its runner starts, so the real setup work (directory copies and five `git` calls) costs no fixture time. The runner and cleanup stay on the real clock, as the deadline-evidence check requires. Each late capture must now get exactly the 1000 ms budget. A new `late-slow-retry` scenario stalls each attempt's setup for 1100 ms, longer than the whole budget, and its attempt 2 still has to succeed. Before, a loaded machine either used up the work budget in setup (`QA-only work budget exhausted`) or Bun's 1580 ms deadline killed a setup `git` child (exit `null` at `skill-e2e-qa-workflow.test.ts:133`); this is what failed late-retry in the 2026-10-24 and 2026-10-25 remote pregates. Under 24 busy loops on 4 cores: late-retry 1/20 → 20/20, late-failure-retry 2/20 → 20/20, late-timeout 6/20 → 20/20, late-slow-retry 20/20. Reverting the clock fails all four late scenarios unloaded; reverting the deadline fails `late-slow-retry`.
+- `test/gstack-skill-start.test.ts`: the `GSTACK_CAPY=1` install-admission case gets an explicit 20 s timeout. Measured on a 4-core Capy machine with a temporary 60 s cap: 3.6–5.7 s through Capy's git shim (9 of 10 over 5 s), 2.6–4.9 s with the real git binary. Four of its ten spawns are full admissions at about 800 ms each.
+
+### Gate integrity
+- One test budget changed, approved by GStack Minder on 2026-10-25: the install-admission case's timeout goes from bun's 5 s default to 20 s, about 3.5 times the 5.7 s worst case measured above. No assertion changed.
+- The QA-only late scenarios register each attempt with its real outer budget (1000 + 80 + 5000 = 6080 ms, the value `outerMs` already asserts) instead of a shortened 1580 ms Bun deadline, which left cleanup 500 ms of real time under load. Measured child totals with the fix under 24 busy loops, 20 runs each: late-timeout 2.4–2.7 s, late-retry 4.1–4.3 s, late-failure-retry 4.7–5.0 s, late-slow-retry 6.2–6.6 s, against the parent's unchanged 15 s spawn timeout. The mocked 1000 ms capture budget, the 80 ms drain and every existing assertion are unchanged. Verdict: justified, because the deadline was a fixture artifact shorter than the budget the case declares, and the clock freeze, not the deadline, is what removes the load sensitivity.
+- `context-restore-fragment-match`: the failure surface moves from the whole transcript to the agent's report. Alpha or omega bytes in tool output no longer fail the case; the payments body in tool output no longer counts as loaded, so the report must show it. A run that resumes another checkpoint, or mixes two, still fails. `test/test-of-test-ratchet.test.ts` gains one BASELINE entry, the replay test, with its reason, and the new helper joins the touchfiles of the nine context-skills cases, as the derived-closure rule requires.
+- `scripts/harness-version.json` records `test/helpers/llm-judge.ts` as non-behavioral at HARNESS_VERSION 6: only a parse failure's message changes, before any score is parsed, so no case's verdict or pass-rate series changes.
+- No threshold, trial count, case kind, rubric, judge request or prompt changed.
+
 ## [1.91.221.0] - 2026-10-25

 **The session-start updater no longer skips because of its own session's workflow lease, and the autostash tests stop failing their first pass in Free Tests.** Since #3385, `gstack-session-update` skips the fast-forward while another session holds a workflow lease, and it means to ignore the session that started it (`--exclude-ancestry-of`). The check runs in the forked background updater, though, and passed the hook's own `$$`, which has exited by then, so the ancestry walk excluded nothing. A Claude Code process that had already run a skill (and so held a lease) and then started a session again in the same process (`/clear`, resume or compaction) logged `SKIP leased pid=<its own pid>` and never updated. `test/session-update-autostash.test.ts` hit the same gap plus a fixture leak: its first two fixtures spawned the updater with the whole parent environment, so it read the test preload's `GSTACK_LEASE_DIR`, which every test file in the same Bun process shares. A lease another file left there gave "Received: SKIP leased pid=3196/1m" where UPDATING or PULL_FAILED was expected, in 13 Free Tests runs since #3385 merged (each passed on the serial retry).
@@ -1 +1 @@
-1.91.221.0
+1.91.223.0
@@ -1,4 +1,4 @@
-# gstack digest v1.91.221.0 — regenerate/re-copy after upgrading gstack
+# gstack digest v1.91.223.0 — regenerate/re-copy after upgrading gstack

 Behavioral rules from gstack (https://github.com/garrytan/gstack), compressed
 for agent hosts without a full skill install. The full skills add workflows,
@@ -18,8 +18,13 @@
 #   door_type          — one-way | two-way
 #                        (optional — looked up from registry if omitted)
 #   options_count      — number of options presented (positive integer)
-#   user_choice        — key user selected (free string; registry-options preferred)
-#   recommended        — option key the agent recommended (optional)
+#   user_choice        — key user selected (free string; registry-options preferred).
+#                        On hook-captured events (source=hook) for a registered
+#                        question_id, the clicked UI label is mapped to its
+#                        registered option key (#2623) and kept in choice_label.
+#                        Unmatched labels and other sources stay as given.
+#   recommended        — option key the agent recommended (optional; mapped
+#                        the same way on hook-captured events)
 #   followed_recommendation — bool (optional — computed if both present)
 #   session_id         — stable session identifier
 #   ts                 — ISO 8601 timestamp (auto-injected if missing)
@@ -47,8 +52,9 @@ INPUT="$1"
 TMPERR=$(mktemp "${TMPDIR:-/tmp}/gstack-question-log.XXXXXX") || { echo "gstack-question-log: could not create a temp file in ${TMPDIR:-/tmp}, so the question was not logged. Fix: point TMPDIR at a writable directory." >&2; exit 1; }
 trap 'rm -f "$TMPERR"' EXIT
 set +e
-VALIDATED=$(printf '%s' "$INPUT" | GSTACK_LIB_DIR="$SCRIPT_DIR/../lib" bun --no-env-file --config="${GSTACK_BUN_CONFIG:-/dev/null}" -e "
+VALIDATED=$(printf '%s' "$INPUT" | GSTACK_LIB_DIR="$SCRIPT_DIR/../lib" GSTACK_SCRIPTS_DIR="$SCRIPT_DIR/../scripts" bun --no-env-file --config="${GSTACK_BUN_CONFIG:-/dev/null}" -e "
 const { hasInjection } = await import(process.env.GSTACK_LIB_DIR + '/jsonl-store.ts');
+const { resolveOptionKey } = await import(process.env.GSTACK_SCRIPTS_DIR + '/psychographic-signals.ts');
 const path = require('path');
 const raw = await Bun.stdin.text();
 let j;
@@ -174,6 +180,23 @@ if (j.recommended !== undefined) {
  if (j.recommended.length > 64) j.recommended = j.recommended.slice(0, 64);
 }

+// Map the clicked UI label to its registered option key (#2623). SIGNAL_MAP
+// and question preferences are keyed by the registry's option keys, but the
+// PostToolUse hook only sees the label. Hook events and registered ids only;
+// an unmatched or ambiguous label stays as given. Agent-written rows keep
+// what the agent logged: skills such as plan-ceo-review log full mode names
+// that their own readers match on.
+if (j.source === 'hook') {
+  const key = resolveOptionKey(j.question_id, j.user_choice);
+  if (key !== undefined && key !== j.user_choice) {
+    j.choice_label = j.user_choice;
+    j.user_choice = key;
+  }
+  if (j.recommended !== undefined) {
+    j.recommended = resolveOptionKey(j.question_id, j.recommended) ?? j.recommended;
+  }
+}
+
 // followed_recommendation — compute if both sides present. An __unknown__
 // choice means extraction failed, not that the user rejected the
 // recommendation — leave the field absent so metrics can't be poisoned.
@@ -1,6 +1,6 @@
 {
  "name": "gstack",
-  "version": "1.91.221",
+  "version": "1.91.223",
  "description": "Garry's Stack — Claude Code skills + fast headless browser. One repo, one install, entire AI engineering workflow.",
  "license": "MIT",
  "type": "module",
@@ -29,7 +29,7 @@
    "test/helpers/eval-budgets.ts": {"blob":"667d0e09b20fe5d7366f57b77b5fb9b04baa37ef","since":6},
    "test/helpers/eval-store.ts": {"blob":"a50ea2c28f508ad9f35ea9353dd2f502ee0dcfc2","since":6,"nonBehavioral":"credential rejections are classed infra instead of assertion: no case's pass or fail changes; only failed trials whose provider rejected the run's key move class, so no rerun spends on a rejected key"},
    "test/helpers/hermetic-env.ts": {"blob":"cc072895461dc52ea317d1beff72121342420d30","since":6},
-    "test/helpers/llm-judge.ts": {"blob":"d0bc6bbbe9f478603d11a1ece95e66c812ad34e6","since":6},
+    "test/helpers/llm-judge.ts": {"blob":"a3177665b662932ce485ccc976c2b36a1531a0bb","since":6,"nonBehavioral":"callJudge rewrites only the message of a JSON parse failure (names the judge output, stop reason, length and the retained llm-judge-response-parse-error event); it still throws a SyntaxError before any score is parsed, the request is unchanged, and no case's pass or fail changes"},
    "test/helpers/paid-test-set.ts": {"blob":"ff9bfe43c43ec2bdd11114f7b8ec7c5736116f75","since":6},
    "test/helpers/pty-screen.ts": {"blob":"4f77569f9ce11f9cee9f347216e468cd7c956820","since":6},
    "test/helpers/pty/auq.ts": {"blob":"16e5ea2bb5b9f28f44808a1a09024dbd3e6f28c4","since":6},
@@ -223,6 +223,89 @@ export const SIGNAL_MAP: Record<string, Record<string, DimensionDelta[]>> = {
  },
 };

+/**
+ * Label cues: option key → leading words of UI labels that mean that key.
+ *
+ * The AskUserQuestion hook only sees the label the user clicked ("A) Fix it
+ * now (recommended)"), never a stable key, while SIGNAL_MAP is keyed by the
+ * registry's option keys (#2623). resolveOptionKey() maps a label back to a
+ * key when bin/gstack-question-log records a hook event. Every key also
+ * matches its own words ("fix-now" → "fix now"), so only phrasings beyond
+ * that need an entry here. Cues are
+ * lowercase words separated by single spaces and match at the start of the
+ * normalized label; candidates are limited to the question's own registered
+ * options, and the longest cue wins.
+ */
+const OPTION_LABEL_CUES: Record<string, readonly string[]> = {
+  accept: ['yes', 'approve', 'agree', 'add', 'include', 'apply', 'adopt', 'enable', 'go ahead', 'do it', 'proceed', 'run', 'create', 'merge', 'roll back', 'migrate', 'open', 'promote', 'reorganize', 'polish'],
+  reject: ['no', 'disagree', 'decline', 'cancel', 'abort', 'dont', 'do not', 'keep current', 'keep as is', 'keep it as is'],
+  skip: ['no', 'no thanks', 'not now', 'dont', 'do not', 'pass', 'drop', 'ignore', 'leave it', 'ship as is', 'ship it as is'],
+  defer: ['later', 'fix later', 'add to todos', 'add to todo', 'add a todo', 'todo', 'postpone', 'backlog', 'park', 'track it'],
+  'fix-now': ['fix', 'fix it', 'fix them', 'fix before shipping'],
+  'fix-anyway': ['fix', 'fix it'],
+  investigate: ['dig in', 'look into', 'debug', 'diagnose', 'root cause'],
+  'ack-and-ship': ['ack', 'acknowledge', 'ship', 'ship anyway', 'ship it'],
+  'accept-risk': ['accept', 'accept the risk', 'live with it'],
+  'false-positive': ['not a bug', 'mark false positive', 'mark as false positive', 'mark it false positive'],
+  'add-test': ['add', 'add tests', 'add the test', 'write', 'write the test', 'test it'],
+  expand: ['scope expansion'],
+  selective: ['cherry pick', 'selective expansion'],
+  hold: ['keep', 'keep scope', 'hold scope'],
+  reduce: ['cut', 'trim', 'shrink', 'scope down', 'scope reduction'],
+  revise: ['edit', 'change', 'rework'],
+  restart: ['start over', 'redo'],
+  'build-now': ['build', 'build it now', 'include in this pr'],
+  'keep-local': ['keep'],
+  'keep-vendored': ['keep'],
+  'yes-upgrade': ['yes', 'upgrade'],
+  'always-auto': ['always'],
+  'not-now': ['later', 'snooze'],
+  'never-ask': ['never', 'disable'],
+};
+
+function normalizeOptionLabel(label: string): string {
+  return label
+    .toLowerCase()
+    .replace(/\s*\(recommended\)\s*$/, '')
+    .replace(/['\u2019]/g, '')
+    .replace(/^\s*(?:option\s+)?[a-z0-9]\s*(?:[).:]|-(?=\s))\s+/, '')
+    .replace(/[^a-z0-9]+/g, ' ')
+    .trim();
+}
+
+/**
+ * Map a chosen UI label to the registered option key of its question.
+ *
+ * Returns the key when exactly one of the question's registered options
+ * matches with the longest cue; undefined for unregistered ids, questions
+ * without options, unmatched labels and ties, so an unclear answer stays
+ * unmapped instead of nudging the wrong dimension.
+ */
+export function resolveOptionKey(question_id: string, choice: string): string | undefined {
+  const options = (QUESTIONS as Record<string, QuestionDef>)[question_id]?.options as string[] | undefined;
+  if (!options?.length) return undefined;
+  if (options.includes(choice)) return choice;
+  const label = normalizeOptionLabel(choice);
+  if (!label) return undefined;
+  let best: string | undefined;
+  let bestWords = 0;
+  let tie = false;
+  for (const key of options) {
+    for (const cue of [key.replace(/-/g, ' '), ...(OPTION_LABEL_CUES[key] ?? [])]) {
+      if (label !== cue && !label.startsWith(cue + ' ')) continue;
+      const words = cue.split(' ').length;
+      if (words > bestWords) {
+        best = key;
+        bestWords = words;
+        tie = false;
+      } else if (words === bestWords && best !== key) {
+        tie = true;
+      }
+    }
+  }
+  return tie ? undefined : best;
+}
+
 /**
 * Apply a user choice for a question to the running dimension totals.
 *
@@ -0,0 +1,121 @@
+import { describe, test, expect } from 'bun:test';
+import * as fs from 'fs';
+import * as path from 'path';
+import { parseNDJSON } from './helpers/session-runner';
+import { fragmentRestoreVerdict } from './helpers/context-restore-fragment';
+
+const MARKERS = { restored: 'FRAGMATCH_PAYMENTS_BUILD', others: ['FRAGMATCH_ALPHA_BUILD', 'FRAGMATCH_OMEGA_BUILD'] };
+const BODIES: Record<string, string> = {
+  '20260101-120000-alpha-feature.md': 'branch: feat/alpha\n## Working on: alpha feature\n### Summary\nAlpha content FRAGMATCH_ALPHA_BUILD\n',
+  '20260202-120000-middle-payments.md': 'branch: feat/payments\n## Working on: middle payments\n### Summary\nPayments content FRAGMATCH_PAYMENTS_BUILD\n',
+  '20260303-120000-omega-release.md': 'branch: feat/omega\n## Working on: omega release\n### Summary\nOmega content FRAGMATCH_OMEGA_BUILD\n',
+};
+const ALL_FILES = Object.keys(BODIES);
+
+const captures: { sessions: Array<{ session_id: string; ndjson: string[] }> } = JSON.parse(
+  fs.readFileSync(path.join(import.meta.dir, 'fixtures', 'context-restore-fragment-match-captures.json'), 'utf8'),
+);
+
+function replay(lines: string[]) {
+  const parsed = parseNDJSON(lines);
+  return { ...parsed, output: parsed.resultLine?.result || '' };
+}
+
+function toolOutput(run: ReturnType<typeof replay>): string {
+  return run.toolCalls.map((call) => call.output).join('\n');
+}
+
+function isReportLine(line: string): boolean {
+  const event = JSON.parse(line);
+  return event.type === 'result'
+    || (event.type === 'assistant' && event.message.content.some((block: any) => block.type === 'text'));
+}
+
+function rewriteReport(lines: string[], rewrite: (line: string) => string): string[] {
+  return lines.map((line) => (isReportLine(line) ? rewrite(line) : line));
+}
+
+function synthetic(files: string[], report: string): string[] {
+  const lines: string[] = [
+    JSON.stringify({ type: 'assistant', parent_tool_use_id: null, message: { content: [{ type: 'tool_use', id: 'skill', name: 'Skill', input: { skill: 'context-restore', args: 'payments' } }] } }),
+    JSON.stringify({ type: 'user', parent_tool_use_id: null, message: { content: [{ type: 'tool_result', tool_use_id: 'skill', content: 'Launching skill: context-restore' }] } }),
+  ];
+  files.forEach((file, i) => {
+    lines.push(JSON.stringify({ type: 'assistant', parent_tool_use_id: null, message: { content: [{ type: 'tool_use', id: `read-${i}`, name: 'Read', input: { file_path: `.gstack-home/projects/p/checkpoints/${file}` } }] } }));
+    lines.push(JSON.stringify({ type: 'user', parent_tool_use_id: null, message: { content: [{ type: 'tool_result', tool_use_id: `read-${i}`, content: BODIES[file] }] } }));
+  });
+  if (report) {
+    lines.push(JSON.stringify({ type: 'assistant', parent_tool_use_id: null, message: { content: [{ type: 'text', text: report }] } }));
+  }
+  lines.push(JSON.stringify({ type: 'result', subtype: report ? 'success' : 'error_max_turns', result: report }));
+  return lines;
+}
+
+function resumeReport(file: string, ...summaries: string[]): string {
+  return `I loaded \`${file}\`.\n\nRESUMING CONTEXT\n### Summary\n${summaries.join('\n')}\n`;
+}
+
+describe('context-restore-fragment-match verdict', () => {
+  test('captured runs hold three sessions that each printed another checkpoint body', () => {
+    expect(captures.sessions).toHaveLength(3);
+    for (const session of captures.sessions) {
+      const run = replay(session.ndjson);
+      expect(run.toolCalls.map((call) => call.tool)).toContain('Skill');
+      expect(toolOutput(run)).toContain('FRAGMATCH_PAYMENTS_BUILD');
+      expect(MARKERS.others.some((marker) => toolOutput(run).includes(marker))).toBe(true);
+    }
+  });
+
+  test.each(captures.sessions.map((s) => [s.session_id, s.ndjson] as const))(
+    'captured run %s: inspected every checkpoint, restored payments -> pass',
+    (_id, ndjson) => {
+      expect(fragmentRestoreVerdict(replay(ndjson), MARKERS)).toEqual({ loadedTarget: true, didNotLoadOthers: true });
+    },
+  );
+
+  test('synthetic: prints every checkpoint, reports payments -> pass', () => {
+    const run = replay(synthetic(ALL_FILES, resumeReport(ALL_FILES[1], 'Payments content FRAGMATCH_PAYMENTS_BUILD')));
+    expect(fragmentRestoreVerdict(run, MARKERS)).toEqual({ loadedTarget: true, didNotLoadOthers: true });
+  });
+
+  test.each([
+    ['alpha', 'FRAGMATCH_ALPHA_BUILD'],
+    ['omega', 'FRAGMATCH_OMEGA_BUILD'],
+  ])('captured run rewritten to resume %s -> fail', (_name, marker) => {
+    for (const session of captures.sessions) {
+      const run = replay(rewriteReport(session.ndjson, (line) => line.replaceAll('FRAGMATCH_PAYMENTS_BUILD', marker)));
+      expect(toolOutput(run)).toContain('FRAGMATCH_PAYMENTS_BUILD');
+      expect(fragmentRestoreVerdict(run, MARKERS)).toEqual({ loadedTarget: false, didNotLoadOthers: false });
+    }
+  });
+
+  test.each([
+    ['alpha', ALL_FILES[0], 'Alpha content FRAGMATCH_ALPHA_BUILD'],
+    ['omega', ALL_FILES[2], 'Omega content FRAGMATCH_OMEGA_BUILD'],
+  ])('synthetic: reads only %s and resumes it -> fail', (_name, file, summary) => {
+    const run = replay(synthetic([file], resumeReport(file, summary)));
+    expect(fragmentRestoreVerdict(run, MARKERS)).toEqual({ loadedTarget: false, didNotLoadOthers: false });
+  });
+
+  test('captured run with omega content mixed into the resume summary -> fail', () => {
+    for (const session of captures.sessions) {
+      const mixed = rewriteReport(session.ndjson, (line) => line.replaceAll(
+        'Payments content FRAGMATCH_PAYMENTS_BUILD', 'Payments content FRAGMATCH_PAYMENTS_BUILD\\nOmega content FRAGMATCH_OMEGA_BUILD'));
+      expect(fragmentRestoreVerdict(replay(mixed), MARKERS)).toEqual({ loadedTarget: true, didNotLoadOthers: false });
+    }
+  });
+
+  test('synthetic: resume summary mixes alpha and payments -> fail', () => {
+    const run = replay(synthetic(ALL_FILES, resumeReport(ALL_FILES[1],
+      'Payments content FRAGMATCH_PAYMENTS_BUILD', 'Alpha content FRAGMATCH_ALPHA_BUILD')));
+    expect(fragmentRestoreVerdict(run, MARKERS)).toEqual({ loadedTarget: true, didNotLoadOthers: false });
+  });
+
+  test('payments body only in tool output, no report of it -> fail', () => {
+    const noReport = replay(synthetic([ALL_FILES[1]], ''));
+    expect(toolOutput(noReport)).toContain('FRAGMATCH_PAYMENTS_BUILD');
+    expect(fragmentRestoreVerdict(noReport, MARKERS).loadedTarget).toBe(false);
+    const nameOnly = replay(synthetic(ALL_FILES, `I loaded \`${ALL_FILES[1]}\`.`));
+    expect(fragmentRestoreVerdict(nameOnly, MARKERS).loadedTarget).toBe(false);
+  });
+});
@@ -0,0 +1,51 @@
+{
+ "provenance": {
+  "case": "context-restore-fragment-match",
+  "run": "local paid baseline: scripts/test-paid-shards.ts --tier periodic --case context-restore-fragment-match --trials 3 (0/3, all on didNotLoadOthers)",
+  "head": "cb99df4a2fc45bc983e0a64f862f79fab0ce899a",
+  "claude_code": "2.1.292",
+  "source_sha256": "2d05a7257f7d559d60ec2e644178a3d65cd0799121260633c931d5b54755b495",
+  "projection": "Raw NDJSON lines of each session's assistant, user and result events, unchanged; system events, thinking-only assistant events and the synthetic user message carrying the extracted SKILL.md body dropped."
+ },
+ "sessions": [
+  {
+   "session_id": "d1147327-49e3-4371-ab54-784d135573ac",
+   "ndjson": [
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2Sp1UcGZ3JFpbUbxQg\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01HHxMjn7G6dQ36WFwvafm5W\",\"name\":\"Skill\",\"input\":{\"skill\":\"context-restore\",\"args\":\"payments\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6256,\"cache_read_input_tokens\":12042,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6256,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"691912ce-6742-4e21-8f4c-a3cfa15b4112\",\"timestamp\":\"2026-10-25T13:21:55.808Z\",\"request_id\":\"req_011CgP2Som6KWQBvN4W6W3B8\",\"wire_tool_inputs\":{\"toolu_01HHxMjn7G6dQ36WFwvafm5W\":{\"skill\":\"context-restore\",\"args\":\"payments\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"type\":\"tool_result\",\"tool_use_id\":\"toolu_01HHxMjn7G6dQ36WFwvafm5W\",\"content\":\"Launching skill: context-restore\"}]},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"9e05cd18-bf0c-4d75-93d8-db0f817d8c87\",\"timestamp\":\"2026-10-25T13:21:55.840Z\",\"tool_use_result\":{\"success\":true,\"commandName\":\"context-restore\",\"allowedTools\":[\"Bash\",\"Read\",\"Glob\",\"Grep\",\"AskUserQuestion\"]}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2SwJ2idXv6jZBjmMRS\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01EyWkH3v7Gu7cBDAgVzCiSu\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa; ls bin; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT); echo ROOT=$GSTACK_STATE_ROOT\\nSLUG=$($B/gstack-slug --get SLUG 2>/dev/null); echo SLUG=$SLUG\\nCHECKPOINT_DIR=\\\"$GSTACK_STATE_ROOT/projects/$SLUG/checkpoints\\\"; ls -la \\\"$CHECKPOINT_DIR\\\"; find .gstack-home -type f\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6339,\"cache_read_input_tokens\":18298,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6339,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":8,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"c6cad35f-a6b6-4eb2-8741-39112ddc1cae\",\"timestamp\":\"2026-10-25T13:21:58.959Z\",\"request_id\":\"req_011CgP2Sw39kvpLEXQaPZoUk\",\"wire_tool_inputs\":{\"toolu_01EyWkH3v7Gu7cBDAgVzCiSu\":{\"command\":\"cd /tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa; ls bin; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT); echo ROOT=$GSTACK_STATE_ROOT\\nSLUG=$($B/gstack-slug --get SLUG 2>/dev/null); echo SLUG=$SLUG\\nCHECKPOINT_DIR=\\\"$GSTACK_STATE_ROOT/projects/$SLUG/checkpoints\\\"; ls -la \\\"$CHECKPOINT_DIR\\\"; find .gstack-home -type f\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_01EyWkH3v7Gu7cBDAgVzCiSu\",\"type\":\"tool_result\",\"content\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-svqfPa\\ntotal 16\\ndrwxr-xr-x 2 user user 4096 Oct 25 13:21 .\\ndrwxr-xr-x 3 user user   60 Oct 25 13:21 ..\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260101-120000-alpha-feature.md\\n-rw-r--r-- 1 user user  168 Oct 25 13:21 20260202-120000-middle-payments.md\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260303-120000-omega-release.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260101-120000-alpha-feature.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260202-120000-middle-payments.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260303-120000-omega-release.md\\n.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa_.git\\n.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"1beb6923-1013-4359-a964-2b06e7594c17\",\"timestamp\":\"2026-10-25T13:21:59.621Z\",\"tool_use_result\":{\"stdout\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-svqfPa\\ntotal 16\\ndrwxr-xr-x 2 user user 4096 Oct 25 13:21 .\\ndrwxr-xr-x 3 user user   60 Oct 25 13:21 ..\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260101-120000-alpha-feature.md\\n-rw-r--r-- 1 user user  168 Oct 25 13:21 20260202-120000-middle-payments.md\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260303-120000-omega-release.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260101-120000-alpha-feature.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260202-120000-middle-payments.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260303-120000-omega-release.md\\n.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa_.git\\n.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false,\"bashEditDiff\":{\"files\":[{\"filePath\":\"/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-svqfPa\"]}],\"created\":true},{\"filePath\":\"/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa_.git\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-svqfPa\"]}],\"created\":true}],\"moreFiles\":0,\"changedFiles\":[\"/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa\",\"/tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/slug-cache/_tmp_gstack-paid-shard-O15pOD_tmp_skill-e2e-ctx-fragment-svqfPa_.git\"]}}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TDR8mJC7cqYJgSDrA\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01E9t3wnBeL5brVygUoRfuoh\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/projects/*/checkpoints; for f in *; do echo \\\"== $f\\\"; cat \\\"$f\\\"; done; cd ../../../..; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; ls .gstack-home/projects/*/checkpoints/*.md | ./bin/gstack-slug --classify-checkpoints; echo rc=$?\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":1039,\"cache_read_input_tokens\":24637,\"cache_creation\":{\"ephemeral_5m_input_tokens\":1039,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"9fa3b490-1be5-4221-b712-c671e2422c46\",\"timestamp\":\"2026-10-25T13:22:01.775Z\",\"request_id\":\"req_011CgP2TD6nbxFnYubbkBVGv\",\"wire_tool_inputs\":{\"toolu_01E9t3wnBeL5brVygUoRfuoh\":{\"command\":\"cd /tmp/gstack-paid-shard-O15pOD/tmp/skill-e2e-ctx-fragment-svqfPa/.gstack-home/projects/*/checkpoints; for f in *; do echo \\\"== $f\\\"; cat \\\"$f\\\"; done; cd ../../../..; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; ls .gstack-home/projects/*/checkpoints/*.md | ./bin/gstack-slug --classify-checkpoints; echo rc=$?\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_01E9t3wnBeL5brVygUoRfuoh\",\"type\":\"tool_result\",\"content\":\"== 20260101-120000-alpha-feature.md\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\\n== 20260202-120000-middle-payments.md\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n== 20260303-120000-omega-release.md\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260101-120000-alpha-feature.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260202-120000-middle-payments.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260303-120000-omega-release.md\\nrc=0\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"96853ca6-5db9-445b-ab54-ac9103ca336a\",\"timestamp\":\"2026-10-25T13:22:02.045Z\",\"tool_use_result\":{\"stdout\":\"== 20260101-120000-alpha-feature.md\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\\n== 20260202-120000-middle-payments.md\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n== 20260303-120000-omega-release.md\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260101-120000-alpha-feature.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260202-120000-middle-payments.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-svqfPa/checkpoints/20260303-120000-omega-release.md\\nrc=0\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TPmyymz1BeaE3qngr\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"text\",\"text\":\"I loaded the checkpoint matching \\\"payments\\\": `20260202-120000-middle-payments.md`. It's the only one of the three saved checkpoints whose title contains that word. All three are unstamped and none come from another project, so there's no project mismatch.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       Feb 2, 2026, 12:00 UTC\\nDuration:    (not recorded)\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\n(none recorded in the checkpoint)\\n\\n### Notes\\n(none recorded)\\n```\\n\\n**Branch:** this context was saved on `feat/payments`, but you're on `main`. You may want to switch to `feat/payments` before continuing.\\n\\nThe other two checkpoints, which I did not load:\\n- `alpha feature` (`feat/alpha`, 2026-01-01)\\n- `omega release` (`feat/omega`, 2026-03-03), the newest\\n\\nYou asked me not to use AskUserQuestion, so I skipped the follow-up prompt. The checkpoint lists no remaining work. You can continue on the payments work, see the full saved file (it's only the short content above), or stop here.\"}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":718,\"cache_read_input_tokens\":25676,\"cache_creation\":{\"ephemeral_5m_input_tokens\":718,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":7,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"uuid\":\"833cfc7c-5f23-4673-b45c-d7d70c0e3f35\",\"timestamp\":\"2026-10-25T13:22:06.611Z\",\"request_id\":\"req_011CgP2TPV8JNDtZyf4D3mYB\"}",
+    "{\"duration_api_ms\":11552,\"stop_reason\":\"end_turn\",\"session_id\":\"d1147327-49e3-4371-ab54-784d135573ac\",\"total_cost_usd\":0.1106026,\"usage\":{\"input_tokens\":8,\"cache_creation_input_tokens\":14352,\"cache_read_input_tokens\":80653,\"output_tokens\":1134,\"output_tokens_details\":{\"thinking_tokens\":86},\"server_tool_use\":{\"web_search_requests\":0,\"web_fetch_requests\":0},\"service_tier\":\"standard\",\"cache_creation\":{\"ephemeral_1h_input_tokens\":0,\"ephemeral_5m_input_tokens\":14352},\"inference_geo\":\"global\",\"iterations\":[{\"input_tokens\":2,\"output_tokens\":484,\"cache_read_input_tokens\":25676,\"cache_creation_input_tokens\":718,\"cache_creation\":{\"ephemeral_5m_input_tokens\":718,\"ephemeral_1h_input_tokens\":0},\"type\":\"message\"}],\"speed\":\"standard\",\"fallback_credit\":null},\"modelUsage\":{\"claude-opus-5-5\":{\"inputTokens\":8,\"outputTokens\":1134,\"cacheReadInputTokens\":80653,\"cacheCreationInputTokens\":14352,\"webSearchRequests\":0,\"costUSD\":0.1106026,\"contextWindow\":1000000,\"maxOutputTokens\":128000,\"thinkingTokens\":86,\"canonicalModel\":\"claude-opus-5-5\",\"provider\":\"firstParty\",\"costBasis\":\"list\"}},\"permission_denials\":[],\"terminal_reason\":\"completed\",\"fast_mode_state\":\"off\",\"fast_mode_disabled_reason\":\"sdk_opt_in_required\",\"subagent_stats\":{\"spawned\":0,\"requested\":{\"background\":0,\"foreground\":0,\"unset\":0},\"started_in_background\":0,\"max_depth\":0,\"spawned_by_subagents\":0,\"completed\":0,\"failed\":0,\"killed\":{\"parent\":0,\"user\":0,\"system\":0},\"refused\":{\"depth_limit\":0,\"concurrency_limit\":0,\"budget\":0},\"by_type\":{}},\"safety_stops\":0,\"is_error\":false,\"num_turns\":5,\"subtype\":\"success\",\"api_error_status\":null,\"result\":\"I loaded the checkpoint matching \\\"payments\\\": `20260202-120000-middle-payments.md`. It's the only one of the three saved checkpoints whose title contains that word. All three are unstamped and none come from another project, so there's no project mismatch.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       Feb 2, 2026, 12:00 UTC\\nDuration:    (not recorded)\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\n(none recorded in the checkpoint)\\n\\n### Notes\\n(none recorded)\\n```\\n\\n**Branch:** this context was saved on `feat/payments`, but you're on `main`. You may want to switch to `feat/payments` before continuing.\\n\\nThe other two checkpoints, which I did not load:\\n- `alpha feature` (`feat/alpha`, 2026-01-01)\\n- `omega release` (`feat/omega`, 2026-03-03), the newest\\n\\nYou asked me not to use AskUserQuestion, so I skipped the follow-up prompt. The checkpoint lists no remaining work. You can continue on the payments work, see the full saved file (it's only the short content above), or stop here.\",\"ttft_ms\":1839,\"type\":\"result\",\"duration_ms\":12657,\"uuid\":\"170f9374-7a93-44aa-9ce6-c54054faf5cc\",\"ttft_stream_ms\":902,\"time_to_request_ms\":199,\"first_content_frame_ms\":902,\"queued_turn_count\":0,\"result_index\":0}"
+   ]
+  },
+  {
+   "session_id": "8889e81b-d4bf-4eba-9ade-cd8805219972",
+   "ndjson": [
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2SpsKBS3fLGqXc3Bwm\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01DewVZTi4xMmGQYH14JLR52\",\"name\":\"Skill\",\"input\":{\"skill\":\"context-restore\",\"args\":\"payments\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6260,\"cache_read_input_tokens\":12042,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6260,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"73df9278-656a-4350-9f2b-11f00fc83e23\",\"timestamp\":\"2026-10-25T13:21:55.982Z\",\"request_id\":\"req_011CgP2SpdBaxUoNcNgUqkdU\",\"wire_tool_inputs\":{\"toolu_01DewVZTi4xMmGQYH14JLR52\":{\"skill\":\"context-restore\",\"args\":\"payments\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"type\":\"tool_result\",\"tool_use_id\":\"toolu_01DewVZTi4xMmGQYH14JLR52\",\"content\":\"Launching skill: context-restore\"}]},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"d884aeed-d430-4d45-93b4-29855244b2ab\",\"timestamp\":\"2026-10-25T13:21:56.013Z\",\"tool_use_result\":{\"success\":true,\"commandName\":\"context-restore\",\"allowedTools\":[\"Bash\",\"Read\",\"Glob\",\"Grep\",\"AskUserQuestion\"]}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2Swzw8ptggHdz4yMNm\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01YASRbf3r6DvzQdp2WBoi5v\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; ls bin; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT); echo ROOT=$GSTACK_STATE_ROOT\\nSLUG=$($B/gstack-slug --get SLUG 2>/dev/null); echo SLUG=$SLUG\\nCHECKPOINT_DIR=\\\"$GSTACK_STATE_ROOT/projects/$SLUG/checkpoints\\\"; ls -la \\\"$CHECKPOINT_DIR\\\"; find \\\"$GSTACK_HOME\\\" -name \\\"*.md\\\"\",\"description\":\"Locate saved checkpoints\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6340,\"cache_read_input_tokens\":18302,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6340,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":8,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"88f53b0b-92b0-433b-be2e-30a6e9c4f180\",\"timestamp\":\"2026-10-25T13:21:59.368Z\",\"request_id\":\"req_011CgP2SwkZ9Z8jUt6x2i38U\",\"wire_tool_inputs\":{\"toolu_01YASRbf3r6DvzQdp2WBoi5v\":{\"command\":\"cd /tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; ls bin; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT); echo ROOT=$GSTACK_STATE_ROOT\\nSLUG=$($B/gstack-slug --get SLUG 2>/dev/null); echo SLUG=$SLUG\\nCHECKPOINT_DIR=\\\"$GSTACK_STATE_ROOT/projects/$SLUG/checkpoints\\\"; ls -la \\\"$CHECKPOINT_DIR\\\"; find \\\"$GSTACK_HOME\\\" -name \\\"*.md\\\"\",\"description\":\"Locate saved checkpoints\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_01YASRbf3r6DvzQdp2WBoi5v\",\"type\":\"tool_result\",\"content\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-mF7D6T\\ntotal 16\\ndrwxr-xr-x 2 user user 4096 Oct 25 13:21 .\\ndrwxr-xr-x 3 user user   60 Oct 25 13:21 ..\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260101-120000-alpha-feature.md\\n-rw-r--r-- 1 user user  168 Oct 25 13:21 20260202-120000-middle-payments.md\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260303-120000-omega-release.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260101-120000-alpha-feature.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260202-120000-middle-payments.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260303-120000-omega-release.md\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"d8dab3b1-3cb6-4750-a8b1-30e12536377f\",\"timestamp\":\"2026-10-25T13:21:59.932Z\",\"tool_use_result\":{\"stdout\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-mF7D6T\\ntotal 16\\ndrwxr-xr-x 2 user user 4096 Oct 25 13:21 .\\ndrwxr-xr-x 3 user user   60 Oct 25 13:21 ..\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260101-120000-alpha-feature.md\\n-rw-r--r-- 1 user user  168 Oct 25 13:21 20260202-120000-middle-payments.md\\n-rw-r--r-- 1 user user  157 Oct 25 13:21 20260303-120000-omega-release.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260101-120000-alpha-feature.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260202-120000-middle-payments.md\\n/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/skill-e2e-ctx-fragment-mF7D6T/checkpoints/20260303-120000-omega-release.md\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false,\"bashEditDiff\":{\"files\":[{\"filePath\":\"/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/slug-cache/_tmp_gstack-paid-shard-XShYTB_tmp_skill-e2e-ctx-fragment-mF7D6T\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-mF7D6T\"]}],\"created\":true},{\"filePath\":\"/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/slug-cache/_tmp_gstack-paid-shard-XShYTB_tmp_skill-e2e-ctx-fragment-mF7D6T_.git\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-mF7D6T\"]}],\"created\":true}],\"moreFiles\":0,\"changedFiles\":[\"/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/slug-cache/_tmp_gstack-paid-shard-XShYTB_tmp_skill-e2e-ctx-fragment-mF7D6T\",\"/tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/slug-cache/_tmp_gstack-paid-shard-XShYTB_tmp_skill-e2e-ctx-fragment-mF7D6T_.git\"]}}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TEimtnXFfvDRRUFde\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_0148N1vosxsxAic6XBPYrXGQ\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/*/checkpoints; for f in *; do echo \\\"== $f\\\"; cat \\\"$f\\\"; done\",\"description\":\"Read checkpoint files\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":1098,\"cache_read_input_tokens\":24642,\"cache_creation\":{\"ephemeral_5m_input_tokens\":1098,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"83ba03a6-8c2c-4fb6-8a99-168b4da4b9ab\",\"timestamp\":\"2026-10-25T13:22:01.555Z\",\"request_id\":\"req_011CgP2TEStrRVjCJmLjGxUQ\",\"wire_tool_inputs\":{\"toolu_0148N1vosxsxAic6XBPYrXGQ\":{\"command\":\"cd /tmp/gstack-paid-shard-XShYTB/tmp/skill-e2e-ctx-fragment-mF7D6T/.gstack-home/projects/*/checkpoints; for f in *; do echo \\\"== $f\\\"; cat \\\"$f\\\"; done\",\"description\":\"Read checkpoint files\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_0148N1vosxsxAic6XBPYrXGQ\",\"type\":\"tool_result\",\"content\":\"== 20260101-120000-alpha-feature.md\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\\n== 20260202-120000-middle-payments.md\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n== 20260303-120000-omega-release.md\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"686ae24d-a3eb-4707-ae3b-330574244652\",\"timestamp\":\"2026-10-25T13:22:01.697Z\",\"tool_use_result\":{\"stdout\":\"== 20260101-120000-alpha-feature.md\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\\n== 20260202-120000-middle-payments.md\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n== 20260303-120000-omega-release.md\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TNGCLmG6DD1SMJXr4\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"text\",\"text\":\"I loaded the checkpoint matching \\\"payments\\\": `20260202-120000-middle-payments.md`. Three checkpoints were saved in this project (alpha-feature, middle-payments, omega-release), and only this one has \\\"payments\\\" in its title. Omega-release is newer, but you asked for this one by name.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       2026-02-02 12:00:00 UTC\\nDuration:    (not recorded)\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\n(none recorded)\\n\\n### Notes\\n(none recorded)\\n```\\n\\n- **Branch:** this context was saved on `feat/payments`, but you're on `main`. You may want to switch branches before continuing.\\n- **Project check:** the checkpoint has no project identity stamp. That's fine here, because no other project has saved checkpoints in this folder.\\n\\nI didn't change any code, and I skipped the usual follow-up question as you asked. Your options are to keep working on this context, see the full file (it's short and shown above in full), or stop here.\"}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":635,\"cache_read_input_tokens\":25740,\"cache_creation\":{\"ephemeral_5m_input_tokens\":635,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":2,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"uuid\":\"70979650-0199-44e3-956e-0824d1de507a\",\"timestamp\":\"2026-10-25T13:22:06.359Z\",\"request_id\":\"req_011CgP2TN24xUJz1E2rmAPb1\"}",
+    "{\"duration_api_ms\":11304,\"stop_reason\":\"end_turn\",\"session_id\":\"8889e81b-d4bf-4eba-9ade-cd8805219972\",\"total_cost_usd\":0.10918220000000001,\"usage\":{\"input_tokens\":8,\"cache_creation_input_tokens\":14333,\"cache_read_input_tokens\":80726,\"output_tokens\":1067,\"output_tokens_details\":{\"thinking_tokens\":71},\"server_tool_use\":{\"web_search_requests\":0,\"web_fetch_requests\":0},\"service_tier\":\"standard\",\"cache_creation\":{\"ephemeral_1h_input_tokens\":0,\"ephemeral_5m_input_tokens\":14333},\"inference_geo\":\"global\",\"iterations\":[{\"input_tokens\":2,\"output_tokens\":439,\"cache_read_input_tokens\":25740,\"cache_creation_input_tokens\":635,\"cache_creation\":{\"ephemeral_5m_input_tokens\":635,\"ephemeral_1h_input_tokens\":0},\"type\":\"message\"}],\"speed\":\"standard\",\"fallback_credit\":null},\"modelUsage\":{\"claude-opus-5-5\":{\"inputTokens\":8,\"outputTokens\":1067,\"cacheReadInputTokens\":80726,\"cacheCreationInputTokens\":14333,\"webSearchRequests\":0,\"costUSD\":0.10918220000000001,\"contextWindow\":1000000,\"maxOutputTokens\":128000,\"thinkingTokens\":71,\"canonicalModel\":\"claude-opus-5-5\",\"provider\":\"firstParty\",\"costBasis\":\"list\"}},\"permission_denials\":[],\"terminal_reason\":\"completed\",\"fast_mode_state\":\"off\",\"fast_mode_disabled_reason\":\"sdk_opt_in_required\",\"subagent_stats\":{\"spawned\":0,\"requested\":{\"background\":0,\"foreground\":0,\"unset\":0},\"started_in_background\":0,\"max_depth\":0,\"spawned_by_subagents\":0,\"completed\":0,\"failed\":0,\"killed\":{\"parent\":0,\"user\":0,\"system\":0},\"refused\":{\"depth_limit\":0,\"concurrency_limit\":0,\"budget\":0},\"by_type\":{}},\"safety_stops\":0,\"is_error\":false,\"num_turns\":5,\"subtype\":\"success\",\"api_error_status\":null,\"result\":\"I loaded the checkpoint matching \\\"payments\\\": `20260202-120000-middle-payments.md`. Three checkpoints were saved in this project (alpha-feature, middle-payments, omega-release), and only this one has \\\"payments\\\" in its title. Omega-release is newer, but you asked for this one by name.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       2026-02-02 12:00:00 UTC\\nDuration:    (not recorded)\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\n(none recorded)\\n\\n### Notes\\n(none recorded)\\n```\\n\\n- **Branch:** this context was saved on `feat/payments`, but you're on `main`. You may want to switch branches before continuing.\\n- **Project check:** the checkpoint has no project identity stamp. That's fine here, because no other project has saved checkpoints in this folder.\\n\\nI didn't change any code, and I skipped the usual follow-up question as you asked. Your options are to keep working on this context, see the full file (it's short and shown above in full), or stop here.\",\"ttft_ms\":1797,\"type\":\"result\",\"duration_ms\":12189,\"uuid\":\"3003a3de-20c9-4a22-8387-f4c23597d9c3\",\"ttft_stream_ms\":873,\"time_to_request_ms\":179,\"first_content_frame_ms\":874,\"queued_turn_count\":0,\"result_index\":0}"
+   ]
+  },
+  {
+   "session_id": "6d684089-4cb5-4016-bc86-203a9a8cdab3",
+   "ndjson": [
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2SoyF1qZULHrLjmM3K\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01X9nJt9XVFgAKfbeHTY7jhD\",\"name\":\"Skill\",\"input\":{\"skill\":\"context-restore\",\"args\":\"payments\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6260,\"cache_read_input_tokens\":12042,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6260,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"7756244d-e82e-4aeb-9437-732412ccb96f\",\"timestamp\":\"2026-10-25T13:21:56.218Z\",\"request_id\":\"req_011CgP2Soi7ME2b1kXN8NCRG\",\"wire_tool_inputs\":{\"toolu_01X9nJt9XVFgAKfbeHTY7jhD\":{\"skill\":\"context-restore\",\"args\":\"payments\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"type\":\"tool_result\",\"tool_use_id\":\"toolu_01X9nJt9XVFgAKfbeHTY7jhD\",\"content\":\"Launching skill: context-restore\"}]},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"376841c6-0fe2-417f-8da9-5d79c52f0694\",\"timestamp\":\"2026-10-25T13:21:56.247Z\",\"tool_use_result\":{\"success\":true,\"commandName\":\"context-restore\",\"allowedTools\":[\"Bash\",\"Read\",\"Glob\",\"Grep\",\"AskUserQuestion\"]}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2Sy7Apg39RghKe7nhs\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_016oZY36mo3ibs1cL3rNt82T\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5; ls bin; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT 2>&1); echo \\\"ROOT=$GSTACK_STATE_ROOT\\\"\\nSLUG=$($B/gstack-slug --get SLUG 2>&1); echo \\\"SLUG=$SLUG\\\"\\nfind .gstack-home -name \\\"*.md\\\" | sort -r\",\"description\":\"Locate gstack state and checkpoints\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":6340,\"cache_read_input_tokens\":18302,\"cache_creation\":{\"ephemeral_5m_input_tokens\":6340,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":8,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"46203839-ca20-4fe1-a98b-d965e965f665\",\"timestamp\":\"2026-10-25T13:22:00.005Z\",\"request_id\":\"req_011CgP2SxnaBEsoXQB2CY6kN\",\"wire_tool_inputs\":{\"toolu_016oZY36mo3ibs1cL3rNt82T\":{\"command\":\"cd /tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5; ls bin; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; B=./bin\\nGSTACK_STATE_ROOT=$($B/gstack-paths --get GSTACK_STATE_ROOT 2>&1); echo \\\"ROOT=$GSTACK_STATE_ROOT\\\"\\nSLUG=$($B/gstack-slug --get SLUG 2>&1); echo \\\"SLUG=$SLUG\\\"\\nfind .gstack-home -name \\\"*.md\\\" | sort -r\",\"description\":\"Locate gstack state and checkpoints\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_016oZY36mo3ibs1cL3rNt82T\",\"type\":\"tool_result\",\"content\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-GZoJH5\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260202-120000-middle-payments.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"d7748b04-b48b-498b-89e5-ce4df4e6d58c\",\"timestamp\":\"2026-10-25T13:22:00.921Z\",\"tool_use_result\":{\"stdout\":\"gstack-config\\ngstack-learnings-log\\ngstack-learnings-search\\ngstack-paths\\ngstack-remote-identity.sh\\ngstack-repo-mode\\ngstack-slug\\ngstack-state-root.sh\\ngstack-timeline-log\\ngstack-timeline-read\\ngstack-update-check\\nROOT=/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home\\nSLUG=skill-e2e-ctx-fragment-GZoJH5\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260202-120000-middle-payments.md\\n.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false,\"bashEditDiff\":{\"files\":[{\"filePath\":\"/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home/slug-cache/_tmp_gstack-paid-shard-V2G2le_tmp_skill-e2e-ctx-fragment-GZoJH5\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-GZoJH5\"]}],\"created\":true},{\"filePath\":\"/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home/slug-cache/_tmp_gstack-paid-shard-V2G2le_tmp_skill-e2e-ctx-fragment-GZoJH5_.git\",\"hunks\":[{\"oldStart\":0,\"oldLines\":0,\"newStart\":1,\"newLines\":1,\"lines\":[\"+v2:skill-e2e-ctx-fragment-GZoJH5\"]}],\"created\":true}],\"moreFiles\":0,\"changedFiles\":[\"/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home/slug-cache/_tmp_gstack-paid-shard-V2G2le_tmp_skill-e2e-ctx-fragment-GZoJH5\",\"/tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5/.gstack-home/slug-cache/_tmp_gstack-paid-shard-V2G2le_tmp_skill-e2e-ctx-fragment-GZoJH5_.git\"]}}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TJxWctw7y2Vz5qBJH\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"tool_use\",\"id\":\"toolu_01FMDVzkCc6y3vknp52LsMBF\",\"name\":\"Bash\",\"input\":{\"command\":\"cd /tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; D=.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints\\nls $D/*.md | ./bin/gstack-slug --classify-checkpoints 2>&1; echo ---; cat $D/20260202-120000-middle-payments.md; echo ---; head -12 $D/20260303-120000-omega-release.md $D/20260101-120000-alpha-feature.md\",\"description\":\"Classify checkpoints and read payments file\"},\"caller\":{\"type\":\"direct\"}}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":718,\"cache_read_input_tokens\":24642,\"cache_creation\":{\"ephemeral_5m_input_tokens\":718,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":16,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"bf17bfed-5ea2-42c9-8b0d-df4adf7e20ff\",\"timestamp\":\"2026-10-25T13:22:03.980Z\",\"request_id\":\"req_011CgP2TJg9vKd8HjaMC2abX\",\"wire_tool_inputs\":{\"toolu_01FMDVzkCc6y3vknp52LsMBF\":{\"command\":\"cd /tmp/gstack-paid-shard-V2G2le/tmp/skill-e2e-ctx-fragment-GZoJH5; export GSTACK_HOME=\\\"$PWD/.gstack-home\\\"; D=.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints\\nls $D/*.md | ./bin/gstack-slug --classify-checkpoints 2>&1; echo ---; cat $D/20260202-120000-middle-payments.md; echo ---; head -12 $D/20260303-120000-omega-release.md $D/20260101-120000-alpha-feature.md\",\"description\":\"Classify checkpoints and read payments file\"}}}",
+    "{\"type\":\"user\",\"message\":{\"role\":\"user\",\"content\":[{\"tool_use_id\":\"toolu_01FMDVzkCc6y3vknp52LsMBF\",\"type\":\"tool_result\",\"content\":\"unstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260202-120000-middle-payments.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md\\n---\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n---\\n==> .gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md <==\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\\n\\n==> .gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md <==\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\",\"is_error\":false}]},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"21ec78ab-abbc-4ea3-8ba6-544d6b05fc91\",\"timestamp\":\"2026-10-25T13:22:04.223Z\",\"tool_use_result\":{\"stdout\":\"unstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260202-120000-middle-payments.md\\nunstamped\\t.gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md\\n---\\n---\\nstatus: in-progress\\nbranch: feat/payments\\ntimestamp: 2026-02-02T12:00:00Z\\n---\\n## Working on: middle payments\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n---\\n==> .gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260303-120000-omega-release.md <==\\n---\\nstatus: in-progress\\nbranch: feat/omega\\ntimestamp: 2026-03-03T12:00:00Z\\n---\\n## Working on: omega release\\n\\n### Summary\\nOmega content FRAGMATCH_OMEGA_BUILD\\n\\n==> .gstack-home/projects/skill-e2e-ctx-fragment-GZoJH5/checkpoints/20260101-120000-alpha-feature.md <==\\n---\\nstatus: in-progress\\nbranch: feat/alpha\\ntimestamp: 2026-01-01T12:00:00Z\\n---\\n## Working on: alpha feature\\n\\n### Summary\\nAlpha content FRAGMATCH_ALPHA_BUILD\",\"stderr\":\"\",\"interrupted\":false,\"isImage\":false,\"noOutputExpected\":false}}",
+    "{\"type\":\"assistant\",\"message\":{\"model\":\"claude-opus-5-5\",\"id\":\"msg_011CgP2TZ93aTmrLHgHkg2Je\",\"type\":\"message\",\"role\":\"assistant\",\"content\":[{\"type\":\"text\",\"text\":\"I loaded the checkpoint `20260202-120000-middle-payments.md`. It's the only one of the three saved checkpoints whose title contains \\\"payments\\\". It isn't the newest: `omega-release` (2026-03-03) is, so I chose this one because its title matched, not because it was the latest.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       Feb 2, 2026, 12:00 UTC\\nDuration:    not recorded\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\nNone recorded in the checkpoint.\\n\\n### Notes\\nNone recorded.\\n```\\n\\n- **Branch:** this was saved on `feat/payments`, but you're on `main`. You may want to switch branches before continuing.\\n- **Project check:** none of the checkpoints have a project identity stamp, but none belong to another project, so this one counts as this project's.\\n- **Other checkpoints:** `alpha feature` on `feat/alpha` (2026-01-01) and `omega release` on `feat/omega` (2026-03-03).\\n\\nYou asked me not to use AskUserQuestion, so I skipped the skill's usual \\\"what next?\\\" prompt. I made no code changes; restoring only reads the saved files.\"}],\"container\":null,\"stop_reason\":null,\"stop_sequence\":null,\"stop_details\":null,\"usage\":{\"input_tokens\":2,\"cache_creation_input_tokens\":859,\"cache_read_input_tokens\":25360,\"cache_creation\":{\"ephemeral_5m_input_tokens\":859,\"ephemeral_1h_input_tokens\":0},\"output_tokens\":8,\"service_tier\":\"standard\",\"inference_geo\":\"global\",\"speed\":\"standard\"},\"input_transformations\":[],\"diagnostics\":null,\"context_management\":null},\"parent_tool_use_id\":null,\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"uuid\":\"e8864f23-50e9-4a59-ab4a-b984590d2559\",\"timestamp\":\"2026-10-25T13:22:09.545Z\",\"request_id\":\"req_011CgP2TYqh2XjMw3Nc9cuQu\"}",
+    "{\"duration_api_ms\":14272,\"stop_reason\":\"end_turn\",\"session_id\":\"6d684089-4cb5-4016-bc86-203a9a8cdab3\",\"total_cost_usd\":0.1112262,\"usage\":{\"input_tokens\":8,\"cache_creation_input_tokens\":14177,\"cache_read_input_tokens\":80346,\"output_tokens\":1212,\"output_tokens_details\":{\"thinking_tokens\":88},\"server_tool_use\":{\"web_search_requests\":0,\"web_fetch_requests\":0},\"service_tier\":\"standard\",\"cache_creation\":{\"ephemeral_1h_input_tokens\":0,\"ephemeral_5m_input_tokens\":14177},\"inference_geo\":\"global\",\"iterations\":[{\"input_tokens\":2,\"output_tokens\":501,\"cache_read_input_tokens\":25360,\"cache_creation_input_tokens\":859,\"cache_creation\":{\"ephemeral_5m_input_tokens\":859,\"ephemeral_1h_input_tokens\":0},\"type\":\"message\"}],\"speed\":\"standard\",\"fallback_credit\":null},\"modelUsage\":{\"claude-opus-5-5\":{\"inputTokens\":8,\"outputTokens\":1212,\"cacheReadInputTokens\":80346,\"cacheCreationInputTokens\":14177,\"webSearchRequests\":0,\"costUSD\":0.1112262,\"contextWindow\":1000000,\"maxOutputTokens\":128000,\"thinkingTokens\":88,\"canonicalModel\":\"claude-opus-5-5\",\"provider\":\"firstParty\",\"costBasis\":\"list\"}},\"permission_denials\":[],\"terminal_reason\":\"completed\",\"fast_mode_state\":\"off\",\"fast_mode_disabled_reason\":\"sdk_opt_in_required\",\"subagent_stats\":{\"spawned\":0,\"requested\":{\"background\":0,\"foreground\":0,\"unset\":0},\"started_in_background\":0,\"max_depth\":0,\"spawned_by_subagents\":0,\"completed\":0,\"failed\":0,\"killed\":{\"parent\":0,\"user\":0,\"system\":0},\"refused\":{\"depth_limit\":0,\"concurrency_limit\":0,\"budget\":0},\"by_type\":{}},\"safety_stops\":0,\"is_error\":false,\"num_turns\":5,\"subtype\":\"success\",\"api_error_status\":null,\"result\":\"I loaded the checkpoint `20260202-120000-middle-payments.md`. It's the only one of the three saved checkpoints whose title contains \\\"payments\\\". It isn't the newest: `omega-release` (2026-03-03) is, so I chose this one because its title matched, not because it was the latest.\\n\\n```\\nRESUMING CONTEXT\\n════════════════════════════════════════\\nTitle:       middle payments\\nBranch:      feat/payments\\nSaved:       Feb 2, 2026, 12:00 UTC\\nDuration:    not recorded\\nStatus:      in-progress\\n════════════════════════════════════════\\n\\n### Summary\\nPayments content FRAGMATCH_PAYMENTS_BUILD\\n\\n### Remaining Work\\nNone recorded in the checkpoint.\\n\\n### Notes\\nNone recorded.\\n```\\n\\n- **Branch:** this was saved on `feat/payments`, but you're on `main`. You may want to switch branches before continuing.\\n- **Project check:** none of the checkpoints have a project identity stamp, but none belong to another project, so this one counts as this project's.\\n- **Other checkpoints:** `alpha feature` on `feat/alpha` (2026-01-01) and `omega release` on `feat/omega` (2026-03-03).\\n\\nYou asked me not to use AskUserQuestion, so I skipped the skill's usual \\\"what next?\\\" prompt. I made no code changes; restoring only reads the saved files.\",\"ttft_ms\":2268,\"type\":\"result\",\"duration_ms\":15609,\"uuid\":\"39ae4b99-23c4-4444-b94e-4cf48961a993\",\"ttft_stream_ms\":1914,\"time_to_request_ms\":192,\"first_content_frame_ms\":1915,\"queued_turn_count\":0,\"result_index\":0}"
+   ]
+  }
+ ]
+}
@@ -336,10 +336,12 @@ describe.skipIf(process.platform === 'win32')('gstack-ff-update: the install-sco
  test('the heartbeat and its sleeper are reaped when the bin exits', () => {
    const fx = makeFixture();
    release(fx, { 'notes.txt': 'top UPSTREAM\nmiddle\nbottom\n' });
-    const before = execFileSync('bash', ['-c', 'pgrep -f "^sleep 300$" | wc -l'], { encoding: 'utf8', timeout: 30_000 }).trim();
-    expect(ff(fx).status).toBe(0);
-    const after = execFileSync('bash', ['-c', 'sleep 0.3; pgrep -f "^sleep 300$" | wc -l'], { encoding: 'utf8', timeout: 30_000 }).trim();
-    expect(after).toBe(before);
+    const beat = String(3000 + (process.pid % 997));
+    const sleepers = (delay = '') => execFileSync('bash', ['-c', `${delay}pgrep -f "^sleep ${beat}$" | wc -l`], { encoding: 'utf8', timeout: 30_000 }).trim();
+    const before = sleepers();
+    expect(before).toBe('0');
+    expect(ff(fx, 'origin/main', { GSTACK_FF_HEARTBEAT_SECONDS: beat }).status).toBe(0);
+    expect(sleepers('sleep 0.3; ')).toBe(before);
  });
 });

@@ -932,7 +932,7 @@ describe('gstack-skill-start install admission', () => {
    const elsewhere = verdict(gh, {}, path.join(gh, 'named.json'));
    fs.rmSync(path.join(gh, 'install-verdicts'), { recursive: true, force: true });
    expect(start(gh, 'autoplan', { ...capy, GSTACK_INSTALL_VERDICT: elsewhere }).status).toBe(0);
-  });
+  }, 20_000);

  test('an installed runtime root follows its .source-path to the source checkout for the verdict and HEAD', () => {
    const gh = scratch();
@@ -0,0 +1,34 @@
+/**
+ * Verdict for the context-restore-fragment-match E2E: which checkpoint did the
+ * agent restore? The evidence is the agent's own report (top-level assistant
+ * text plus the final result), never tool output. Restore has to inspect the
+ * candidates to find the one whose title matches, so every checkpoint body can
+ * appear in Bash or Read output without being the restored context.
+ */
+export interface FragmentRestoreRun {
+  output?: string;
+  transcript?: any[];
+}
+
+export function agentReport(result: FragmentRestoreRun): string {
+  const parts: string[] = [];
+  for (const event of result.transcript || []) {
+    if (event?.type !== 'assistant' || event.parent_tool_use_id) continue;
+    for (const block of event.message?.content || []) {
+      if (block?.type === 'text' && typeof block.text === 'string') parts.push(block.text);
+    }
+  }
+  if (result.output) parts.push(result.output);
+  return parts.join('\n');
+}
+
+export function fragmentRestoreVerdict(
+  result: FragmentRestoreRun,
+  markers: { restored: string; others: string[] },
+): { loadedTarget: boolean; didNotLoadOthers: boolean } {
+  const report = agentReport(result);
+  return {
+    loadedTarget: report.includes(markers.restored),
+    didNotLoadOthers: markers.others.every((marker) => !report.includes(marker)),
+  };
+}
@@ -60,6 +60,7 @@ export const FREE_FIXTURES: Readonly<Record<string, readonly string[]>> = {
  'test/fixtures/codex-sandbox/**': ['test/codex-eval-recording.test.ts', 'test/codex-model-probe.test.ts', 'test/outside-review-result.test.ts', 'test/outside-voice-invocation.test.ts'],
  'test/fixtures/context-bill/**': ['test/context-bill.test.ts', 'test/context-budget-ratchet.test.ts'],
  'test/fixtures/context-budget.json': ['test/helpers/capture-context-budget.ts', 'test/test-value-bar.test.ts'],
+  'test/fixtures/context-restore-fragment-match-captures.json': ['test/context-restore-fragment-verdict.test.ts'],
  'test/fixtures/coverage-audit-cd-chain.json': ['test/coverage-audit-evidence.test.ts'],
  'test/fixtures/coverage-audit-parallel-column.json': ['test/coverage-audit-evidence.test.ts'],
  'test/fixtures/coverage-audit-separator-collision.json': ['test/coverage-audit-evidence.test.ts'],
@@ -187,6 +187,13 @@ export async function callJudge<T>(
    .filter(block => block.type === 'text')
    .map(block => block.text)
    .join('\n');
+  const parseJudgeText = (json: string): T => {
+    try {
+      return JSON.parse(json) as T;
+    } catch (error) {
+      throw new SyntaxError(`Judge response is not valid JSON (stop_reason=${response.stop_reason}, ${json.length} chars; full text in llm-judge-response-parse-error): ${error instanceof Error ? error.message : String(error)}`);
+    }
+  };
  try {
    if (response.stop_reason === 'max_tokens') {
      throw new Error(`Judge response truncated at max_tokens=${maxTokens} (model=${resolvedModel})`);
@@ -194,11 +201,11 @@ export async function callJudge<T>(
    if (response.stop_reason === 'refusal') throw new JudgeRefusalError(response);
    if (opts?.jsonSchema !== undefined) {
      if (response.stop_reason !== 'end_turn') throw new Error(`Structured judge did not complete: stop_reason=${response.stop_reason}`);
-      return JSON.parse(text) as T;
+      return parseJudgeText(text);
    }
    const jsonMatch = text.match(/\{[\s\S]*\}/);
    if (!jsonMatch) throw new Error(`Judge returned non-JSON: ${text.slice(0, 200)}`);
-    return JSON.parse(jsonMatch[0]) as T;
+    return parseJudgeText(jsonMatch[0]);
  } catch (error) {
    // The canonical full stderr spool retains this public response even when
    // parsing fails before a caller can record a judgment. Never copy content
@@ -682,15 +682,15 @@ export const E2E_TOUCHFILES: Record<string, string[]> = {
  // test/skill-e2e-context-skills.test.ts. These are periodic-tier because
  // each one spawns claude -p and costs ~$0.20-$0.40. Collectively they
  // verify the thing the /checkpoint → /context-save rename was for.
-  'context-save-routing':                  [ 'context-save/**', 'scripts/resolvers/preamble.ts', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-save-then-restore-roundtrip':   ['bin/gstack-state-root.sh', 'lib/state-root.ts', 'context-save/**', 'context-restore/**', 'bin/gstack-slug', 'bin/gstack-remote-identity.sh', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-restore-fragment-match':        [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-restore-provenance-order':      ['context-save/**', 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-restore-empty-state':           [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-restore-list-delegates':        [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-restore-legacy-compat':         [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-save-list-current-branch':      [ 'context-save/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
-  'context-save-list-all-branches':        [ 'context-save/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts'],
+  'context-save-routing':                  [ 'context-save/**', 'scripts/resolvers/preamble.ts', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-save-then-restore-roundtrip':   ['bin/gstack-state-root.sh', 'lib/state-root.ts', 'context-save/**', 'context-restore/**', 'bin/gstack-slug', 'bin/gstack-remote-identity.sh', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-restore-fragment-match':        [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-restore-provenance-order':      ['context-save/**', 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-restore-empty-state':           [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-restore-list-delegates':        [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-restore-legacy-compat':         [ 'context-restore/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-save-list-current-branch':      [ 'context-save/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],
+  'context-save-list-all-branches':        [ 'context-save/**', 'test/skill-e2e-context-skills.test.ts', 'bin/gstack-skill-start', 'test/helpers/shipped-skill-routing.ts', 'test/helpers/context-restore-fragment.ts'],

  // Document-release
  'document-release': [ 'test/helpers/outside-voice-evidence.ts', 'bin/gstack-version-bump', 'document-release/**', 'test/skill-e2e-workflow.test.ts', 'test/fixtures/coverage-audit-fixture.ts', 'test/helpers/coverage-audit.ts', 'test/helpers/coverage-audit-evidence.ts', 'test/helpers/office-hours-attempt.ts'],
@@ -218,6 +218,22 @@ describe('frontier Claude judge compatibility', () => {
    expect(create).toHaveBeenCalledTimes(1);
  });

+  // Train-4 gate 40173759210 slice 4 attempt 1: Sonnet's free-form cookie workflow judgment left the quotes around "one `--domain` per run" unescaped.
+  const COOKIE_GATE_UNESCAPED_REPLY = "{\"clarity\": 4, \"completeness\": 4, \"actionability\": 4, \"reasoning\": \"The workflow is well ordered: Aside check, then consent questions before any `$B` command, then daemon prep, import, and an honest three-line report. It states the boundaries precisely: profile directory vs display name, `--all` exclusions, Chromium-only reset with its non-rollback caveat, and the rule that copied cookies, counts, and HTTP 200 never prove sign-in. The error and failure-reason tables cover recovery and platform limits. Gaps and ambiguities remain. (1) Step 3.1 sets env vars and runs `$B stop` before `$B status` and `goto`, which is workable but slightly awkward. (2) Step 2 says to ask everything before any `$B` command, yet it also needs the Aside probe, which comes earlier in step 1 and so is an implicit exception. (3) It is unclear how the picker or direct import maps the receipt's states to the three report lines, such as `Imported: <count>` for partial or failed results, or the storage-reset line when the picker is used. (4) The `--all` flow is not given a command example or a step. (5) For the multiple-domain case the instruction is \"one `--domain` per run\", but it does not say how to handle the target tab or the receipt across runs. (6) `decryption_failed` says to close the source browser, which sits in tension with the Chrome 136+ note, though it is not contradictory. (7) The `handoff` path requires a display, and the fallback when none is available is not defined beyond 'offer only with a display'. (8) The Windows example's `$B reload` without verification is fine, but the unverified case is not tied to the report. Overall it is strong with a few minor missing decisions.\"}";
+
+  test('the captured cookie judge reply fails closed and names the judge output, stop reason and retained text', async () => {
+    create.mockResolvedValue({ id: 'msg_011CgNbBNCNcZ2C6WJDCmjgU', model: 'claude-sonnet-5-5', stop_reason: 'end_turn',
+      content: [{ type: 'text', text: COOKIE_GATE_UNESCAPED_REPLY }], usage: { input_tokens: 4521, output_tokens: 572 } } as never);
+    let failure: unknown;
+    try { await callJudge('score the workflow', 'claude-sonnet-5-5'); } catch (error) { failure = error; }
+    expect(failure).toBeInstanceOf(SyntaxError);
+    expect((failure as Error).message).toStartWith(`Judge response is not valid JSON (stop_reason=end_turn, ${COOKIE_GATE_UNESCAPED_REPLY.length} chars; full text in llm-judge-response-parse-error): `);
+    expect(create).toHaveBeenCalledTimes(1);
+    expect(diagnostics).toHaveBeenCalledTimes(1);
+    expect(JSON.parse(diagnostics.mock.calls[0][0])).toMatchObject({ type: 'llm-judge-response-parse-error', stopReason: 'end_turn',
+      textBlocks: [COOKIE_GATE_UNESCAPED_REPLY], error: { name: 'SyntaxError', message: (failure as Error).message } });
+  });
+
  test('a weak structured score still fails the substance bar', async () => {
    create.mockResolvedValue({ stop_reason: 'end_turn', content: [{ type: 'text',
      text: JSON.stringify({ reason_substance: 1, reasoning: 'Boilerplate.' }) }] } as never);
@@ -33,6 +33,7 @@ import {
  validateRegistrySignalKeys,
  newDimensionTotals,
  normalizeToDimensionValue,
+  resolveOptionKey,
  ALL_DIMENSIONS,
 } from '../scripts/psychographic-signals';
 import {
@@ -290,6 +291,125 @@ describe('psychographic signal map', () => {
  });
 });

+// -----------------------------------------------------------------------
+// #2623 — chosen UI label → registered option key at capture time
+// -----------------------------------------------------------------------
+
+describe('option label → registered key (#2623)', () => {
+  const words = (key: string) => key.replace(/-/g, ' ');
+  const title = (s: string) => s.charAt(0).toUpperCase() + s.slice(1);
+
+  test('every registered option resolves from its own words, as a bare label and as an AUQ label', () => {
+    const misses: string[] = [];
+    for (const q of Object.values(QUESTIONS as Record<string, QuestionDef>)) {
+      for (const key of (q.options ?? []) as string[]) {
+        for (const label of [key, words(key), `A) ${title(words(key))} (recommended)`, `B. ${title(words(key))}`]) {
+          if (resolveOptionKey(q.id, label) !== key) misses.push(`${q.id}: ${JSON.stringify(label)} → ${resolveOptionKey(q.id, label)}`);
+        }
+      }
+    }
+    expect(misses).toEqual([]);
+  });
+
+  test('common phrasings map to the question\'s own key', () => {
+    const cases: Array<[string, string, string]> = [
+      ['plan-eng-review-arch-finding', 'A) Fix it now (recommended)', 'fix-now'],
+      ['plan-eng-review-arch-finding', 'Fix later', 'defer'],
+      ['plan-eng-review-arch-finding', 'B) Defer to TODOS.md', 'defer'],
+      ['plan-eng-review-arch-finding', 'C) Accept the risk and ship', 'accept-risk'],
+      ['plan-ceo-review-expansion-proposal', 'A) Add to plan (recommended)', 'accept'],
+      ['plan-ceo-review-expansion-proposal', 'B) Add to TODOs', 'defer'],
+      ['plan-ceo-review-mode', 'A) Expand scope (Recommended)', 'expand'],
+      ['plan-ceo-review-mode', 'A) SCOPE EXPANSION', 'expand'],
+      ['plan-ceo-review-mode', 'B) SELECTIVE EXPANSION', 'selective'],
+      ['plan-ceo-review-mode', 'C) HOLD SCOPE (recommended)', 'hold'],
+      ['plan-ceo-review-mode', 'D) SCOPE REDUCTION', 'reduce'],
+      ['plan-ceo-review-mode', 'Cherry-pick expansions', 'selective'],
+      ['plan-ceo-review-mode', 'Cut to the minimum', 'reduce'],
+      ['ship-pre-landing-review-fix', 'Ship as-is', 'skip'],
+      ['ship-test-failure-triage', 'Ship anyway, the failures are known', 'ack-and-ship'],
+      ['review-finding-fix', 'Not a bug, the input is trusted', 'false-positive'],
+      ['land-and-deploy-merge-confirm', 'Yes, merge it', 'accept'],
+      ['land-and-deploy-merge-confirm', 'Don’t merge yet', 'reject'],
+      ['plan-eng-review-test-gap', 'Add the test now', 'add-test'],
+      ['office-hours-mode-goal', 'Hackathon / demo', 'hackathon'],
+    ];
+    for (const [id, label, key] of cases) {
+      expect({ id, label, key: resolveOptionKey(id, label) }).toEqual({ id, label, key });
+    }
+  });
+
+  test('negative controls stay unmapped', () => {
+    // Unregistered (ad-hoc or hook-hash) ids: no registered options to map to.
+    expect(resolveOptionKey('plan-eng-review-ci-typecheck', 'Fix now')).toBeUndefined();
+    expect(resolveOptionKey('hook-0123456789', 'Accept')).toBeUndefined();
+    // Registered question with no options list.
+    expect(resolveOptionKey('plan-devex-review-persona', 'Accept')).toBeUndefined();
+    // A label that names none of the question's options.
+    expect(resolveOptionKey('plan-eng-review-arch-finding', 'Restore the parallel typecheck job')).toBeUndefined();
+    // Cues anchor at the start of the label, never mid-sentence.
+    expect(resolveOptionKey('review-sql-safety', 'Ship it, then fix')).toBeUndefined();
+    expect(resolveOptionKey('review-sql-safety', "Don't fix")).toBeUndefined();
+    // Another question's key is not a candidate.
+    expect(resolveOptionKey('qa-tier', 'Fix now')).toBeUndefined();
+    // A selector-looking number is not stripped into a match.
+    expect(resolveOptionKey('qa-tier', '1.5x deeper')).toBeUndefined();
+    expect(resolveOptionKey('plan-eng-review-arch-finding', '')).toBeUndefined();
+  });
+
+  function logAndRead(entries: Array<Record<string, unknown>>): Array<Record<string, unknown>> {
+    const tmpHome = fs.mkdtempSync(path.join(require('os').tmpdir(), 'gstack-2623-'));
+    try {
+      const { spawnSync } = require('child_process');
+      const env = { ...process.env, GSTACK_HOME: tmpHome, GSTACK_QUESTION_LOG_NO_DERIVE: '1' };
+      const logBin = path.join(ROOT, 'bin', 'gstack-question-log');
+      for (const e of entries) {
+        const r = spawnSync(logBin, [JSON.stringify({ skill: 'plan-eng-review', question_summary: 'q?', ...e })], { env, cwd: ROOT, encoding: 'utf-8', timeout: 30_000 });
+        expect(r.status).toBe(0);
+      }
+      const [slug] = fs.readdirSync(path.join(tmpHome, 'projects'));
+      return fs
+        .readFileSync(path.join(tmpHome, 'projects', slug, 'question-log.jsonl'), 'utf-8')
+        .trim()
+        .split('\n')
+        .map((l) => JSON.parse(l));
+    } finally {
+      fs.rmSync(tmpHome, { recursive: true, force: true });
+    }
+  }
+
+  test('gstack-question-log records the key and keeps the label for hook events on registered ids only', () => {
+    const [hook, agentKey, agentMode, adHoc, unmatched, freeText] = logAndRead([
+      { question_id: 'plan-eng-review-arch-finding', user_choice: 'A) Fix it now (recommended)', recommended: 'Fix it now', source: 'hook' },
+      { question_id: 'plan-eng-review-arch-finding', user_choice: 'defer', recommended: 'fix-now' },
+      { skill: 'plan-ceo-review', question_id: 'plan-ceo-review-mode', user_choice: 'HOLD SCOPE', recommended: 'HOLD SCOPE', auto_decided: true },
+      { question_id: 'plan-eng-review-ci-typecheck', user_choice: 'Fix now', source: 'hook' },
+      { question_id: 'plan-eng-review-arch-finding', user_choice: 'Restore the parallel typecheck job', source: 'hook' },
+      { question_id: 'plan-eng-review-arch-finding', user_choice: 'Fix it but rename the module first', free_text: 'Fix it but rename the module first', source: 'auq-other' },
+    ]);
+    expect(hook).toMatchObject({ user_choice: 'fix-now', choice_label: 'A) Fix it now (recommended)', recommended: 'fix-now', followed_recommendation: true });
+    expect(agentKey).toMatchObject({ user_choice: 'defer', recommended: 'fix-now', followed_recommendation: false });
+    expect(agentKey.choice_label).toBeUndefined();
+    // Agent-written rows keep what the agent logged; the auto-decide harness
+    // matches plan-ceo-review's full mode names.
+    expect(agentMode).toMatchObject({ source: 'agent', user_choice: 'HOLD SCOPE', recommended: 'HOLD SCOPE', followed_recommendation: true });
+    expect(agentMode.choice_label).toBeUndefined();
+    expect(adHoc).toMatchObject({ user_choice: 'Fix now' });
+    expect(adHoc.choice_label).toBeUndefined();
+    expect(unmatched).toMatchObject({ user_choice: 'Restore the parallel typecheck job' });
+    expect(unmatched.choice_label).toBeUndefined();
+    expect(freeText).toMatchObject({ user_choice: 'Fix it but rename the module first', source: 'auq-other' });
+    expect(freeText.choice_label).toBeUndefined();
+  });
+
+  test('a picked label scores as following a key recommendation', () => {
+    const [row] = logAndRead([
+      { question_id: 'plan-ceo-review-mode', user_choice: 'A) Expand scope (Recommended)', recommended: 'expand', source: 'hook' },
+    ]);
+    expect(row).toMatchObject({ user_choice: 'expand', followed_recommendation: true });
+  });
+});
+
 // -----------------------------------------------------------------------
 // Archetypes
 // -----------------------------------------------------------------------
@@ -19,7 +19,7 @@ test('QA-only fixture helpers and owned sections select the no-fix consumer', ()
 });

 test.each(['success', 'max-turns', 'dependencies', 'metadata', 'edit', 'source-write', 'source-delete', 'source-add', 'setup', 'runner', 'recording', 'runner-recording', 'artifact-error',
-  'process-error', 'slow-setup', 'exhausted-setup', 'late-timeout', 'late-retry', 'late-failure-retry', 'unknown-timeout',
+  'process-error', 'slow-setup', 'exhausted-setup', 'late-timeout', 'late-retry', 'late-slow-retry', 'late-failure-retry', 'unknown-timeout',
  'cleanup', 'cleanup-retry', 'runner-cleanup', 'cleanup-recording', 'remove-error', 'server-stop',
  'missing-deadline', 'unguarded-probe', 'forged-receipt', 'deadline-write', 'reset-deadline',
  'preparation-missing', 'preparation-failed', 'preparation-unacknowledged', 'preparation-wrong-path', 'preparation-late',
@@ -47,11 +47,12 @@ if (scenario === 'remove-error') spyOn(fs, 'rmSync').mockImplementation((target,
  return originalRm(target, options);
 });
 mock.module(path.join(root, 'test/helpers/eval-store.ts'), () => ({ getProjectEvalDir: () => ${JSON.stringify(path.join(dir, 'evals'))} }));
-const late = scenario.startsWith('late-'), workMs = 1000, drainMs = 80;
+const late = scenario.startsWith('late-'), lateRetry = ['late-retry', 'late-slow-retry'].includes(scenario), workMs = 1000, drainMs = 80;
 const realNow = Date.now, controlledClock = ['slow-setup', 'exhausted-setup'].includes(scenario);
 const clockStart = realNow();
-let clock = clockStart;
+let clock = clockStart, frozenAt;
 if (controlledClock) Date.now = () => clock;
+else if (late) Date.now = () => frozenAt ?? realNow();
 let finalized = false, outerMs;
 const collector = { addTest: row => {
  const cwd = directories.at(-1);
@@ -77,8 +78,7 @@ mock.module(path.join(root, 'test/helpers/e2e-helpers.ts'), () => ({
  testConcurrentIfSelected: (id, body, budget) => {
    if (id !== 'qa-only-no-fix') return;
    outerMs = budget;
-    const deadline = late ? (budget >= workMs + 5000 ? budget - 5000 + 500 : budget) : 5000;
-    test(id, body, deadline);
+    test(id, late ? () => { frozenAt = realNow(); return body(); } : body, late ? budget : 5000);
  },
  copyDirSync: (source, target) => {
    if (scenario === 'setup') throw new Error('QA-only setup failed');
@@ -87,6 +87,7 @@ mock.module(path.join(root, 'test/helpers/e2e-helpers.ts'), () => ({
  setupBrowseShims: cwd => {
    directories.push(cwd); fs.mkdirSync(path.join(cwd, 'browse/bin'), { recursive: true });
    if (controlledClock) clock += scenario === 'slow-setup' ? 295000 : 300000;
+    if (scenario === 'late-slow-retry') Bun.sleepSync(workMs + 100);
  },
  logCost() {}, createEvalCollector: () => collector, finalizeEvalCollector: async () => { finalized = true; },
  recordE2E: (_collector, name, suite, result, extra) => collector.addTest({ name, suite,
@@ -96,6 +97,7 @@ mock.module(path.join(root, 'test/helpers/e2e-helpers.ts'), () => ({
 mock.module(path.join(root, 'test/helpers/session-runner.ts'), () => ({
  SESSION_DRAIN_GRACE_MS: late ? drainMs : 5000,
  runSkillTest: async options => {
+    frozenAt = undefined;
    const cwd = options.workingDirectory;
    const fact = { cwd, exists: fs.existsSync(cwd), recordedBeforeCapture: records.length,
      reportBefore: fs.existsSync(path.join(cwd, 'qa-reports/qa-only-report.md')), captureMs: options.timeout,
@@ -145,7 +147,7 @@ mock.module(path.join(root, 'test/helpers/session-runner.ts'), () => ({
      fact.existsAfterDrain = fs.existsSync(cwd);
      return { ...result, exitReason: 'timeout' };
    }
-    if (late && (scenario !== 'late-retry' || attempts.length === 1)) {
+    if (late && (!lateRetry || attempts.length === 1)) {
      await new Promise(resolve => setTimeout(resolve, options.timeout + drainMs));
      fact.existsAfterDrain = fs.existsSync(cwd);
      return { ...result, exitReason: 'timeout' };
@@ -237,7 +239,7 @@ await import(path.join(root, 'test/skill-e2e-qa-workflow.test.ts'));
      expect(records[index].cleanupDone).toBe(true);
      expect(cleanups[index].budget).toBeGreaterThan(0);
      expect(cleanups[index].budget).toBeLessThanOrEqual(4000);
-      expect(records[index].passed).toBe(scenario === 'recording' || !shouldFail && !(['late-retry', 'cleanup-retry'].includes(scenario) && index === 0));
+      expect(records[index].passed).toBe(scenario === 'recording' || !shouldFail && !(['late-retry', 'late-slow-retry', 'cleanup-retry'].includes(scenario) && index === 0));
      if (scenario.includes('cleanup') && !(scenario === 'cleanup-retry' && index === 1)) expect(records[index].error).toContain('QA-only browser cleanup: invalid daemon identity');
    }
    for (const [index, attempt] of attempts.entries()) {
@@ -257,6 +259,7 @@ await import(path.join(root, 'test/skill-e2e-qa-workflow.test.ts'));
    }
    if (scenario.startsWith('late-')) {
      expect(outerMs).toBe(1000 + 80 + 5000);
+      for (const attempt of attempts) expect(attempt.captureMs).toBe(1000);
      expect(records[0]).toMatchObject({ passed: false, exit_reason: 'timeout', cost_usd: 0.25 });
      expect(records[0].error).toContain('timeout');
      if (count === 2) expect(directories[0]).not.toBe(directories[1]);
@@ -23,6 +23,7 @@ import {
 } from './helpers/e2e-helpers';
 import { extractSkillBody } from './helpers/skill-fixture';
 import { readShippedSkillRouting } from './helpers/shipped-skill-routing';
+import { fragmentRestoreVerdict } from './helpers/context-restore-fragment';
 import { spawnSync } from 'child_process';
 import * as fs from 'fs';
 import * as path from 'path';
@@ -260,11 +261,14 @@ Do NOT use AskUserQuestion.`,

    logCost('context-restore-fragment-match', result);

-    // Broader surface — agent may stop at Skill call without echoing the
-    // body marker. The payments file's body is in tool outputs (Read/Bash).
-    const out = fullOutputSurface(result);
-    const loadedPayments = out.includes('FRAGMATCH_PAYMENTS_BUILD');
-    const didNotLoadOthers = !out.includes('FRAGMATCH_ALPHA_BUILD') && !out.includes('FRAGMATCH_OMEGA_BUILD');
+    // What was restored is what the agent reports, not what its tools printed:
+    // finding the matching title can cat every checkpoint, so all three bodies
+    // may appear in Bash/Read output. The report must carry the payments body
+    // and no other checkpoint's body.
+    const { loadedTarget: loadedPayments, didNotLoadOthers } = fragmentRestoreVerdict(result, {
+      restored: 'FRAGMATCH_PAYMENTS_BUILD',
+      others: ['FRAGMATCH_ALPHA_BUILD', 'FRAGMATCH_OMEGA_BUILD'],
+    });
    const routedToRestore = skillCalls(result).includes('context-restore');
    const exitOk = ['success', 'error_max_turns'].includes(result.exitReason);

@@ -60,6 +60,7 @@ const BASELINE = [
  'test/codex-carve-fixture.test.ts',
  'test/codex-eval-recording.test.ts',
  'test/codex-resume-flag-semantics.test.ts',
+  'test/context-restore-fragment-verdict.test.ts', // GSTA-91: replays three captured false reds (0/3 on main cb99df4) through the paid context-restore-fragment-match case's fragmentRestoreVerdict
  'test/cookie-workflow-judge-input.test.ts', // slices test/skill-llm-eval.test.ts source (recorded 2026-10 when the ratchet learned paid-source slicing)
  'test/cso-ntfs-fixture.test.ts',
  'test/cso-windows-build-contract.test.ts',