Files
gstack/test/autoplan-phase-handoff.test.ts
T
Garry Tan 943105f109 v1.91.8.0 test: smaller suite, derived paid selection, retired never-green evals (#2994)
* test: delete test-infrastructure dead code (G)

- exit-propagation drives the runner's real strict verdict
  (BunTestOutputClassifier + strictTestExitCode); delete the unused
  shardRunLooksTruncated predicate.
- delete skill-coverage-matrix registry + its gate (nothing reads it; the
  floor already iterates skillCensus()).
- delete touchfiles-facade export-parity tests (Bun fails missing imports
  at link time) and the duplicated E2E_TIERS tier-value test.
- delete brain-cache-spec TRANSPORT_DEFAULT_POLICY, SKILL_RUN_RETENTION_DAYS
  and the now-unused BrainTrustPolicy type with their literal tests.
  AUTOPLAN_PREFLIGHT_BUDGET_BYTES stays: skill-preflight-budget enforces it
  against real resolver output.
- delete audit-compliance's JSDoc-comment grep.

* test: replace product tests that fake the product with real-boundary tests (F)

- design: serve.test.ts drove an inline mirror server; now two tests run the
  real serve() on an ephemeral port (reload confinement, submit exit 0).
- setup-gbrain: rollback + voyage tests execute the template-extracted init
  blocks (3 sites) instead of drifted local bash copies.
- terminal-agent: internalHandler source greps replaced by a behavioral
  /internal/grant + /internal/revoke auth matrix (no/wrong/valid token).
- /health: server-security-surface and the server-auth / security-audit-r2 /
  sidebar-tabs source greps fold into one liveness-only check on the real
  body; the L4 sidecar wiring gets a behavioral /pty-inject-scan test.
- delete tautologies (browser-manager onDisconnect, memory-command #12),
  ios swiftui tap fixture self-check, memory-ingest put_page grep, detach
  source greps, sidebar-agent absence pins, dead-CSS pins + the dead CSS,
  security-audit-r2 Task 1 + the test-only meta-commands re-export,
  duplicate generated-SKILL.md checks.
- make-pdf coverage-gaps cases move into their owner test files.

* test: delete tests of dead eval code (A)

- A1: the retired Eng lexical oracle (evaluateEngSeedCoverage,
  isEngSeedDecisionAUQ), the completion-handoff detector and the retained
  corpus had no paid caller since v1.87.6; delete their 26 replay files,
  ~2.6k helper LOC and fixtures, and the dead blocks in 8 mixed files
  (live hasNativePlanTerminal / batching assertions stay).
- A2: dead viewport approvers in autoplan-artifact-permission and their 11
  replay files + fixtures; recorder/launcher cases stay.
- A3: never-wired oracles and seeders (autoplan-phase-order,
  eng-finding-fixture, ceo-paired-fixture, design-ui-scope,
  plan-skill-completion, pty-current-screen, required-reads,
  transcript-section-logger); plan-seed-submission now decodes through the
  production createPtyScreen; section manifests name their actual guard.
- A4: zero-reference helper exports, plus execGit and invokeAndObserve
  found by the reachability pass.
- 52 fixtures orphaned by the deletions; touchfile and selection-table
  entries for every deleted path.

* test: clean up the paid eval lane (B1-B4, B6, B7)

- B1: delete paid files that assert nothing or cannot pass meaningfully:
  skill-llm-eval-spec and skill-e2e-spec-execute (test.todo), gemini-e2e
  (+ gemini-session-runner; no gemini CLI in CI), ship-idempotency (red
  since v1.63), the two opus-4-7 *-sonnet overlay wrappers, conductor-prose
  (+ its source-evaluation replay), codex-e2e-plan-format; drop their keys,
  scripts and census rows.
- B2: skill-llm-eval grades browse/sections/command-list.md with one union
  judge that also carries the baseline score pin; regression-vs-baseline
  deleted (paid run: pass, c4/c4/a4).
- B3: memory-pipeline, ios-qa, ios-qa-swift-build and plan-tune-cathedral
  make no model calls; renamed out of the paid glob so they run on every
  PR. Swift builds need GSTACK_TEST_SWIFT=1; device stub deleted.
- B4: codex-e2e*, outside-voice, aside and ios-device cannot run in the CI
  image; excluded from the weekly lane with a tracked re-entry condition.
- B6: fold opus-47's negative routing controls into skill-routing-e2e
  journey-negatives (paid run: 3/3 unrouted) and delete the file.
- B7: delete the never-green brain-privacy-gate eval; a free
  gstack-skill-start test now proves consent precedes artifacts egress.

* test: retire the finding-count cluster and trim its helpers (C)

- C0/C1: the five never-green evals (skill-e2e-autoplan-chain and
  skill-e2e-plan-{ceo,eng,design,devex}-finding-count) failed on harness and
  budget, never on skill behavior; delete them, their touchfile/tier ids,
  AUTOPLAN_CHAIN_BUDGET and the dedicated eighth periodic slice (--slices 7).
- C2: delete the helper groups whose only paid consumers were those files
  (11 modules), trim claude-pty-runner and eng-seeded-coverage to the paid
  closure, and delete the free replay tests whose assertions exercised only
  that dead code (89 files, 135 orphaned fixtures). Blocks that used dead code
  only as input for a live subject keep their assertions: the multiSelect
  default moved to plan-review-decisions, runner PTY tests use inline caller
  policies, and the timer-safe budget checks moved to eng-finding-retry-budget.
- The eight production-touching files stay except ceo-current-decision-record
  (its template read only feeds the retired counter).
- CARVE_GUARDS.autoplan is behavioral 'none'; TODOS records the lost chain
  and per-finding cadence coverage with their re-entry tests.

* test: fold per-incident replay series into their detector owners (D)

Twelve detector families move into one owner test each: 73 incident files
become describe blocks in ceo-section-loading-fixture (stale-fill race),
model-overlays, coverage-audit-evidence, autoplan-phase-observer,
native-auto-decide, outside-voice-evidence, eng-first-review,
plan-count-completion, plan-count-file-permission, ceo-mode-option,
plan-scope-selection and plan-count-prerequisite. Each block keeps its original
code and fixture, so every case still runs; only tests asserting the incident
file's own touchfile registration are dropped (41). Touchfile lists that named
an incident now name its owner.

* test: start the plan-count history PTY on its readiness marker (H)

The fake CLI prints a startup marker and the runner waits for it instead of the
fixed 8 s startup sleep (8.6 s -> 0.9 s locally). eng-semantic-terminal's
sleeping registration cases went with C; plan-count-timeout keeps the fixed wait
because it asserts deadline behavior.

* test: derive paid touchfiles from each eval's static closure (E)

touchfiles.test.ts now checks, per key, that the paid file's static
test/helpers and test/fixtures closure (plus fixture paths it names in string
literals) is covered, and names the file, path, chain and key to fix when it is
not. Free *.test.ts files are no longer touchfiles, so editing a free replay
test stops selecting paid evals: 950 entries removed, 653 real closure paths
added. The hand-copied inventories go: periodic-fixture-selection,
fake-impeccable-touchfiles and 45 per-file selection examples. Selection for
the sample edits (plan-eng-review template, claude-pty-runner,
plan-count-fixture, gstack-config) loses no case under either profile.
CONTRIBUTING documents the rule and its lower bound.

* test: skip hollow tier shards and census judges in the paid planner (B5)

A paid file is now skipped for a tier lane only when every E2E id it registers
is known statically and none has that tier; ids come from the touchfile
registrations and literal testName/*IfSelected arguments, so a comment or
skill path that quotes another id cannot unschedule it, and computed names
keep today's scheduling. --list and the manifest show each skip as
"skipped: no E2E_TIERS id has tier <tier>". The weekly gate census drops the
LLM judges (--skip-judges); they still run in the periodic census and PR gate
lanes. Gate lane 52 -> 42 files, census 41; periodic 77 -> 69.

* test: run seven paid evals on the current default capture model (B8)

skill-e2e-{auq-matrix,plan-format,qa-bugs,retro,workflow} pinned
claude-opus-4-7 and skill-e2e-office-hours plus -brain-writeback pinned
claude-sonnet-4-6; none tests a historical model, so they now capture with
resolveEvalModel('capture'), and the free harness tests that execute these
registrations receive the same resolver. The paid re-pin run passed all of
them. skill-e2e-{design,office-hours-phase4,plan-prosons,plan} keep
claude-opus-4-7: six of their cases failed on the default model (three
timeouts, a missing report file, a format miss and a posture score of 3), so
per the plan's fallback they keep their pins with a TODOS entry. The pre-spend
estimate and drop threshold are in docs/test-audit-2026-09.md.

* test: guard the reduced suite against new test-of-test files

- test/test-of-test-ratchet.test.ts records the 228 free tests that import only
  test/ code and fails on a new one, naming the owner test to extend instead;
  a stale baseline entry fails with the remove instruction.
- test/helpers/resolve-repo-path.ts is the one specifier/literal resolver for
  the ratchet and the touchfile closure invariant, with its own unit tests.
- CONTRIBUTING "Test tiers" describes the paid-failure workflow (fix, then one
  row in the detector's owner test) and the ratchet; TEST_PORTFOLIO gains the
  detector -> owner-test table and no longer claims an Autoplan chain eval.
- TODOS: automatic exclusion policy for chronically red periodic files (P3),
  the deferred native-completion table collapse, the unused CEO payment
  seeder; the PTY readiness item is narrowed to the paid runner.
- docs/test-audit-2026-09.md collects the triage, security mapping, inventories,
  selection proof, behavior-commit decisions and retained false positives.

* v1.91.8.0 test: smaller suite, derived paid selection, retired never-green evals

Release metadata for the test-reduction branch: VERSION 1.91.8.0 (1.91.7.0 is
claimed by #2983), CHANGELOG with the measured before/after table and a
contributor section, durations re-recorded on Ubicloud standard-16 (857 files,
0 failures), the agents digest, CONTRIBUTING's after-measurement row, the B8
fallback TODOS entry, and the after metrics, kept-vs-plan notes, B8 run and
census estimate in docs/test-audit-2026-09.md.

* fix(ubicloud): skip retrieval globs that match nothing instead of reporting a failed pull

* test: count issue-numbered Design findings in the UI-scope gate eval
2026-09-29 08:58:49 -07:00

305 lines
21 KiB
TypeScript

import { afterEach, expect, test } from 'bun:test';
import { mkdtempSync, mkdirSync, readFileSync, rmSync, writeFileSync } from 'node:fs';
import { tmpdir } from 'node:os';
import { dirname, join, resolve } from 'node:path';
import { createHash } from 'node:crypto';
import { initializePlan, prepareMethodology, createSnapshot, amendImplementation,
extractImplementationPlan, checkPhaseImplementation } from '../bin/gstack-autoplan-snapshot';
import { autoplanPhaseCompletions } from './helpers/autoplan-phase-observer';
import { auditAutoplanMethodReads, loadAutoplanMethodologyBinding } from './helpers/autoplan-method-read-audit';
import { readPlanCountTranscript, type NativePublicToolEvent } from './helpers/plan-count-transcript';
import captured from './fixtures/autoplan-phase-handoff-6714.json';
const ROOT = resolve(import.meta.dir, '..');
const source = (file: string) => readFileSync(join(ROOT, file), 'utf8');
const owned: string[] = [];
afterEach(() => { for (const dir of owned.splice(0)) rmSync(dir, { recursive: true, force: true }); });
function fixture() {
const root = mkdtempSync(join(tmpdir(), 'gstack-autoplan-handoff-')); owned.push(root);
const input = join(root, 'input.md'), active = join(root, 'active.md'), restore = join(root, 'restore.md');
writeFileSync(input, captured.originalImplementation);
initializePlan(input, active, restore);
const methodology = prepareMethodology('ceo', join(ROOT, 'plan-ceo-review/SKILL.md'), restore);
const create = () => createSnapshot('ceo', active, restore, methodology.methodologyPath);
const record = (block: string) => {
const plan = readFileSync(active, 'utf8');
writeFileSync(active, plan.slice(0, plan.indexOf('## Review record\n')) + '## Review record\n' + block + '\n');
};
return { root, active, create, record };
}
test('captured accepted CEO decisions remained outside the operative input', () => {
const f = fixture(); f.record(captured.acceptedCeoBlock);
const premature = f.create();
expect(premature.sha256).toBe(captured.originalImplementationSha256);
expect(premature.sourceBytes).toBe(4607);
expect(readFileSync(premature.snapshotPath, 'utf8')).not.toContain('per-source deadline');
expect(captured.acceptedCeoBlock).toContain('per-source deadline');
});
test('a Step 0 checkpoint applies full accepted bytes before a fresh voice snapshot', () => {
const f = fixture(), step0 = f.create();
f.record(captured.acceptedCeoBlock);
amendImplementation('ceo', f.active, step0.snapshotPath);
expect(checkPhaseImplementation('ceo', f.active, step0.snapshotPath, 'changed').recordedObligations.none).toBe(false);
expect(extractImplementationPlan(readFileSync(f.active, 'utf8'))).toContain(captured.acceptedCeoBlock);
const revised = captured.acceptedCeoBlock.replace('<!-- /autoplan-accepted:ceo -->',
'- Synthetic accepted spec follow-up: verify the completed response.\n<!-- /autoplan-accepted:ceo -->');
f.record(revised);
amendImplementation('ceo', f.active, step0.snapshotPath);
const voices = f.create();
expect(voices.snapshotPath).not.toBe(step0.snapshotPath);
expect(voices.sha256).not.toBe(captured.originalImplementationSha256);
expect(readFileSync(voices.snapshotPath, 'utf8')).toContain(revised.split('\n').slice(1, -1).join('\n'));
expect(readFileSync(step0.snapshotPath, 'utf8')).toBe(captured.originalImplementation);
const final = revised.replace('<!-- /autoplan-accepted:ceo -->',
'- Synthetic accepted full-review follow-up: retain the earlier spec fix.\n<!-- /autoplan-accepted:ceo -->');
f.record(final);
amendImplementation('ceo', f.active, voices.snapshotPath);
expect(checkPhaseImplementation('ceo', f.active, voices.snapshotPath, 'changed').recordedObligations.none).toBe(false);
expect(extractImplementationPlan(readFileSync(f.active, 'utf8'))).toContain(final);
});
test('no-change or a draft record does not invent an accepted amendment', () => {
const f = fixture(), step0 = f.create();
f.record('<!-- autoplan-accepted:ceo -->\nNone: existing requirements cover this phase.\n<!-- /autoplan-accepted:ceo -->');
amendImplementation('ceo', f.active, step0.snapshotPath);
expect(checkPhaseImplementation('ceo', f.active, step0.snapshotPath, 'unchanged').recordedObligations.none).toBe(true);
expect(f.create().sha256).toBe(captured.originalImplementationSha256);
f.record(captured.acceptedCeoBlock);
expect(() => checkPhaseImplementation('ceo', f.active, step0.snapshotPath, 'changed')).toThrow();
expect(f.create().sha256).toBe(captured.originalImplementationSha256);
});
test('captured file summaries and real tool progress cannot replace missing parent messages', () => {
expect(captured.observedOutcome).toBe('timeout');
const transcript = { status: 'ready' as const, calls: [], assistantMessages: captured.assistantMessages };
expect(autoplanPhaseCompletions(transcript, Date.parse('2026-09-15T12:29:00Z'))).toEqual([]);
// Explicitly synthetic message; it proves only CEO, never the other three phases.
const message = { sessionId: captured.assistantMessages[0]!.sessionId,
timestamp: '2026-09-15T13:24:30Z', text: 'Phase 1 complete.' };
expect(autoplanPhaseCompletions({ ...transcript, assistantMessages: [...transcript.assistantMessages, message] }, 0))
.toEqual([{ phase: 1, ts: Date.parse(message.timestamp) }]);
for (const text of ['# Phase 1 complete.', '> Phase 1 complete.', '```\nPhase 1 complete.\n```', 'I will send Phase 1 complete. later.']) {
expect(autoplanPhaseCompletions({ ...transcript, assistantMessages: [{ ...message, text }] }, 0)).toEqual([]);
}
});
test('compaction restores the exact dispatch field from the existing immutable manifest', () => {
const f = fixture(), snapshot = f.create();
const restored = JSON.parse(readFileSync(join(dirname(snapshot.snapshotPath), 'snapshot.json'), 'utf8'));
expect(restored.nativeDispatchPrompt).toBe(snapshot.nativeDispatchPrompt);
expect(loadAutoplanMethodologyBinding(restored.nativeDispatchPrompt, [f.root]).phase).toBe('ceo');
expect(restored.nativeDispatchPrompt).not.toContain(captured.originalImplementation);
expect(createHash('sha256').update(readFileSync(snapshot.nativePromptPath)).digest('hex')).toBe(snapshot.nativePromptSha256);
// The exact captured DX prefix is insufficient by design; the inline body is not a dispatch.
expect(auditAutoplanMethodReads([{ sessionId: 'captured-parent', timestamp: captured.dxDispatch.timestamp,
toolUseId: captured.dxDispatch.toolUseId, kind: 'use', name: 'Agent', input: { prompt: captured.dxDispatch.promptPrefix } }],
() => { throw new Error('An inline body must not select a methodology binding'); })).toEqual([]);
expect(() => loadAutoplanMethodologyBinding(snapshot.nativePrompt, [f.root])).toThrow();
expect(() => loadAutoplanMethodologyBinding(restored.nativeDispatchPrompt.replace('CEO', 'DESIGN'), [f.root])).toThrow();
});
test('the working-plan destination rule leaves conversation messages in the conversation', () => {
const intake = source('autoplan/SKILL.md.tmpl').split('### Step 1: Capture restore point')[1]!.split('### Step 2:')[0]!;
expect(intake).not.toContain('Write all amendments/outputs to ACTIVE_PLAN');
expect(intake).toContain('Save plan amendments and review artifacts to ACTIVE_PLAN');
expect(intake).toContain('Send phase announcements and the final approval request in the conversation');
});
test('CEO applies Step 0 decisions before each spec dispatch and refreshes the native input afterward', () => {
const section = source('autoplan/sections/ceo-phase.md.tmpl');
const preliminary = section.slice(section.indexOf('**At 0H'), section.indexOf('Step 0.5 (Dual Voices)'));
expect(preliminary).toContain('`snapshotPath` as `<CEO_STEP0_CHECKPOINT>`');
expect(preliminary).toContain('Before every spec dispatch, including after each accepted spec fix');
expect(preliminary).toContain('amend-input ceo "<ACTIVE_PLAN>" "<CEO_STEP0_CHECKPOINT>" "<RESTORE_PATH>" "<methodologyPath>"');
expect(preliminary).toContain('use returned `reviewInputPath` as `<CEO_SPEC_INPUT>`');
expect(preliminary).toContain('returned `readRanges` offset/limit through EOF');
expect(preliminary).toContain('Supply the complete CEO scope summary and `<CEO_SPEC_INPUT>`');
expect(preliminary).toContain('three-launch cap');
expect(preliminary).toContain('User Challenges retain the original requirements');
expect(section).toContain('create a fresh snapshot below for both voices');
});
test('compaction recovery separates saved artifacts, sent messages and pending reviewer state', () => {
const contract = source('autoplan/SKILL.md.tmpl').split('## Sequential Execution')[1]!.split('---')[0]!;
expect(contract).toContain('reconcile saved artifacts and sent conversation messages separately');
expect(contract).toContain("verified phase lacks its announcement, resume the close procedure at step 6 (Publish) before advancing");
expect(contract).toContain('regenerate and reread the full packet if the implementation or accepted decisions changed');
expect(contract).toContain('If its reviewer is pending, wait for that same reviewer');
expect(contract).toContain('Read `snapshot.json` beside that final `<PHASE_INPUT>`');
expect(contract).toContain('use its `nativeDispatchPrompt` unchanged');
});
test('a same-phase draft checkpoint cannot substitute for the post-Step-0 voice input', () => {
const f = fixture(), draft = f.create();
f.record(captured.acceptedCeoBlock);
amendImplementation('ceo', f.active, draft.snapshotPath);
const final = f.create();
// Both are valid immutable CEO manifests. Phase identity alone cannot choose the final input.
expect(loadAutoplanMethodologyBinding(draft.nativeDispatchPrompt, [f.root]).phase).toBe('ceo');
expect(loadAutoplanMethodologyBinding(final.nativeDispatchPrompt, [f.root]).phase).toBe('ceo');
expect(draft.sha256).toBe(captured.originalImplementationSha256);
expect(final.sha256).not.toBe(draft.sha256);
const contract = source('autoplan/SKILL.md.tmpl').split('## Sequential Execution')[1]!.split('---')[0]!;
expect(contract).toContain('finish any incomplete preliminary work before recovering a voice input');
expect(contract).toContain('Never dispatch `<CEO_STEP0_CHECKPOINT>`');
expect(contract).toContain('If the final voice input does not exist, create it after the preliminary gates');
});
test('taste overrides follow the existing affected-phase and final Eng rerun rule', () => {
const skill = source('autoplan/SKILL.md.tmpl');
const override = skill.split('- B:')[1]!.split('- B2:')[0]!;
expect(override).toContain("follow D's affected-phase rerun rule (including Eng last)");
expect(override).toContain('before re-presenting the gate');
expect(override).toContain('same 3-cycle cap as D');
expect(skill).toContain('a re-run of any earlier phase re-runs Eng after it');
expect(skill).toContain('scope→1, design→2, dx→2.5, test plan→3, arch→3');
expect(skill).not.toContain('scope→1B');
});
test('captured parent text and a following tool can share a response without ending the turn', () => {
const root = mkdtempSync(join(tmpdir(), 'gstack-autoplan-continuation-')); owned.push(root);
const project = join(root, 'projects', 'fixture'); mkdirSync(project, { recursive: true });
const [textEnvelope, toolEnvelope] = captured.nativeContinuation.records;
expect(textEnvelope!.message.id).toBe(toolEnvelope!.message.id);
expect(textEnvelope!.message.stop_reason).toBe('tool_use');
expect(toolEnvelope!.message.stop_reason).toBe('tool_use');
const file = join(project, `${textEnvelope!.sessionId}.jsonl`);
const read = (rows: unknown[]) => {
writeFileSync(file, rows.map(row => JSON.stringify(row)).join('\n') + '\n');
const tools: NativePublicToolEvent[] = [];
const transcript = readPlanCountTranscript(root, root, event => tools.push(event));
return { transcript, tools, hits: autoplanPhaseCompletions(transcript, 0) };
};
const start = Date.parse(textEnvelope!.timestamp);
// Explicit synthetic phase text/input inside the captured public response envelope.
// The actual original text made no completion claim and retains zero credit.
const originals = captured.nativeContinuation.records.map(row => ({ ...row, cwd: root,
message: { ...row.message, content: row.message.content.map(block => block.type === 'tool_use'
? { ...block, input: { command: 'true' } } : block) } }));
expect(read(originals).hits).toEqual([]);
const phases = [1, 2, 2.5, 3];
const rows = phases.flatMap((phase, index) => {
const id = `msg_synthetic_phase_${String(phase).replace('.', '_')}`;
return [
{ ...textEnvelope, cwd: root, timestamp: new Date(start + index * 10).toISOString(),
message: { ...textEnvelope!.message, id, content: [{ type: 'text', text: `Phase ${phase} complete.` }] } },
{ ...toolEnvelope, cwd: root, timestamp: new Date(start + index * 10 + 1).toISOString(),
message: { ...toolEnvelope!.message, id, content: [{ type: 'tool_use', id: `next_${phase}`,
name: 'Read', input: { file_path: join(root, 'next-phase.md') } }] } },
];
});
const result = read(rows);
expect(result.hits.map(hit => hit.phase)).toEqual(phases);
expect(result.tools).toHaveLength(4);
expect(result.hits.every((hit, index) => hit.ts < Date.parse(result.tools[index]!.timestamp))).toBe(true);
// Tool arguments, tool results and sidechain text are not parent announcements.
expect(read([{ ...rows[1], message: { ...rows[1]!.message, content: [{ type: 'tool_use', id: 'source',
name: 'Bash', input: { command: 'echo "Phase 1 complete."' } }] } }]).hits).toEqual([]);
expect(read([{ ...rows[0], type: 'user', message: { role: 'user', content: [{ type: 'tool_result',
tool_use_id: 'source', content: 'Phase 1 complete.' }] } }]).hits).toEqual([]);
expect(read([{ ...rows[0], isSidechain: true }]).hits).toEqual([]);
expect(read([{ ...rows[0], message: { ...rows[0]!.message,
content: [{ type: 'text', text: 'Phase 2 skipped — no UI scope.' }] } }]).hits).toEqual([]);
});
test('phase progress text permits immediate tool continuation in the same turn', () => {
const contract = source('autoplan/SKILL.md.tmpl').split('## Sequential Execution')[1]!.split('---')[0]!;
expect(contract.replace(/\s+/g, ' ')).toContain("load its `phase-close` section afresh");
expect(contract).toContain('in the same turn');
expect(contract).not.toContain('This parent response contains no tool calls');
const shared = source('autoplan/sections/phase-close.md.tmpl').replace(/\s+/g, ' ');
const operations = ["3. **Prepare this phase's close packet.**", '4. **Read the complete current packet.**',
'5. **Verify the current implementation.**', '6. **Publish the parent report.**', '7. **Return to the driver.**'];
const positions = operations.map(operation => shared.indexOf(operation));
expect(positions.every(position => position >= 0)).toBe(true);
expect(positions).toEqual([...positions].sort((a, b) => a - b));
expect(shared).toContain('SEND the filled report below now as visible parent assistant text');
expect(shared).toContain('This message is the next operation before any next-phase tool call');
expect(shared).toContain('After sending the actual parent report, continue to the driver in the same turn');
expect(shared).toContain("The sent conversation message is step 6's output");
expect(shared).not.toContain('The packet owns the close continuation');
expect(shared).not.toContain('This message contains no tool calls');
for (const phase of ['ceo', 'design', 'dx', 'eng']) {
const close = source(`autoplan/sections/${phase}-phase.md.tmpl`).split('**Close this phase:**')[1]!;
expect(close).toContain('{{SECTION:phase-close}}');
expect(close.trim().endsWith('{{SECTION:phase-close}}')).toBe(true);
expect(close).not.toContain('**Phase ');
}
});
test('the captured cf74 full readback did not publish a CEO report before the Design Read', () => {
// Public projection from chain-ceo-close/receipt.json (0b1ede1e…0e9de0),
// source cf74db538a2f4c4361f2573316abb91e01663564. The complete retained
// parent window contains no assistant text between this Read ACK and Design.
const lastParentMessage = {
sessionId: '647c542b-5f7a-45d7-b1d9-8e2d984eab17',
timestamp: '2026-09-16T01:27:36.112Z',
text: 'Tasks JSONL written (13 rows). Now the phase-close procedure for CEO.',
};
const readbackAck = { toolUseId: 'toolu_01SpARevKkyf9P9MqwN8w9Ms',
timestamp: '2026-09-16T01:27:53.896Z', startLine: 1, numLines: 109, totalLines: 109 };
const nextPhaseUse = { toolUseId: 'toolu_01XvX1QbuKqv1xWjpdHsFLnj',
timestamp: '2026-09-16T01:28:04.501Z', name: 'Read',
input: { file_path: '/home/vercel-sandbox/gstack/autoplan/sections/design-phase.md' } };
const boundary = Date.parse(nextPhaseUse.timestamp);
expect(Date.parse(readbackAck.timestamp)).toBeLessThan(boundary);
expect(readbackAck.numLines).toBe(readbackAck.totalLines);
const observe = (messages: typeof lastParentMessage[], through = boundary) =>
autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: messages.filter(message => Date.parse(message.timestamp) <= through) },
Date.parse(lastParentMessage.timestamp));
expect(observe([lastParentMessage])).toEqual([]);
// Synthetic controls exercise the unchanged public observer's timestamps.
// A later announcement remains later; it cannot populate the earlier boundary.
const report = { ...lastParentMessage, timestamp: new Date(boundary - 1).toISOString(),
text: '**Phase 1 complete.**\nCodex: disabled. Claude subagent: completed: 9 issues.\n'
+ 'Consensus: N/A (voice coverage missing).\nPassing to Phase 2 (Design Review).' };
expect(observe([lastParentMessage, report])).toEqual([{ phase: 1, ts: boundary - 1 }]);
const late = { ...report, timestamp: new Date(boundary + 1).toISOString() };
expect(observe([lastParentMessage, late])).toEqual([]);
expect(observe([lastParentMessage, late], boundary + 1)).toEqual([{ phase: 1, ts: boundary + 1 }]);
for (const text of [`\`\`\`text\n${report.text}\n\`\`\``, 'I will send the CEO completion report after Design.']) {
expect(observe([lastParentMessage, { ...report, text }])).toEqual([]);
}
});
// Minimal exact public projection of the two f359 CEO close failures. Packet
// hashes/complete ranges were authenticated against each original Read result;
// only this parent-message boundary is replayed here, not semantic compliance.
test.each([
{ attempt: 'uREF54', sessionId: '94599121-1626-4188-a553-e68579eeb329',
lastAt: '2026-09-16T07:37:49.858Z',
lastText: 'One stale phrase in R1: "production p95 ≤ 300ms over each 48h cohort hold" contradicts row 40 (hold = max(48h, power-based minimum)). Fixing both copies, then regenerating the packet.',
readId: 'toolu_01AEEg5eZxFJdurZgdDQsUs3', readAt: '2026-09-16T07:38:07.954Z', lines: 137,
packetSha256: '6e1349a485c2d21bb62588df2730439f5dfffbcc5faaf6f9dc500178dbb3fe1d',
designId: 'toolu_01VCEDJEdVHpqX3d6VuLBMrx', designAt: '2026-09-16T07:38:18.984Z' },
{ attempt: 'RhTXQ5', sessionId: '45abf2fa-0d62-471f-9efa-9a0d5b2ec1b5',
lastAt: '2026-09-16T08:26:23.970Z',
lastText: 'Task JSONL written (11 lines). Now reading `phase-close.md` afresh to close Phase 1.',
readId: 'toolu_01AnPNRCY2c9Lg6nvSMpdYsp', readAt: '2026-09-16T08:26:42.105Z', lines: 138,
packetSha256: '0de1420a316b222df62b5454bcea6d812b4e5040d873a225bfe89840c5a1652f',
designId: 'toolu_0116k1GsR8JxowqSYBJtJbpi', designAt: '2026-09-16T08:26:52.230Z' },
])('captured f359 $attempt complete packet leaves publication pending', evidence => {
const message = { sessionId: evidence.sessionId, timestamp: evidence.lastAt, text: evidence.lastText };
const boundary = Date.parse(evidence.designAt);
const observe = (messages: typeof message[]) => autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: messages.filter(row => Date.parse(row.timestamp) <= boundary) }, Date.parse(evidence.lastAt));
expect(Date.parse(evidence.readAt)).toBeLessThan(boundary);
expect(observe([message])).toEqual([]);
// Synthetic publication tests only the missing native operation and its order.
const report = { ...message, timestamp: new Date(boundary - 1).toISOString(),
text: '**Phase 1 complete.**\nOutside review: disabled. Native subagent: completed: 9 issues.\n'
+ 'Consensus: N/A (voice coverage missing).\nPassing to Phase 2 (Design Review).' };
expect(observe([message, report])).toEqual([{phase: 1, ts: boundary - 1}]);
expect(observe([message, {...report, timestamp: new Date(boundary + 1).toISOString()}])).toEqual([]);
for (const text of ['```text\n' + report.text + '\n```', '> Phase 1 complete.',
'Expected output:\n' + report.text, 'I will publish Phase 1 complete after Design.']) {
expect(observe([message, {...report, text}])).toEqual([]);
}
});