Files
gstack/test/autoplan-phase-observer.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

483 lines
28 KiB
TypeScript
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import { describe, expect, test } from 'bun:test';
import * as fs from 'node:fs';
import * as os from 'node:os';
import * as path from 'node:path';
import { pathToFileURL } from 'node:url';
import { autoplanPhaseCompletions } from './helpers/autoplan-phase-observer';
import type { PlanCountTranscript } from './helpers/plan-count-transcript';
import fixture_autoplan_phase_dash_ao from './fixtures/autoplan-phase-dash-ao.json';
import actual_autoplan_with_result_au from './fixtures/autoplan-with-result-au.json';
const START = Date.parse('2026-09-08T16:00:00.000Z');
const transcript = (...messages: Array<[number, string]>): PlanCountTranscript => ({
status: 'ready', calls: [],
assistantMessages: messages.map(([ms, text]) => ({
sessionId: 'autoplan-fixture', timestamp: new Date(START + ms).toISOString(), text,
})),
});
describe('native autoplan phase observation', () => {
test('AF public wrapped-up announcement retains its actual Phase 1 timestamp', () => {
// Exact reader-projected public parent narration from the AF run; no raw
// signature or private block is stored. The next-phase mention adds no hit.
const announcement = {
sessionId: '6ce27a04-7422-4687-9de9-138e68f308d8',
text: 'Phase 1 wrapped up: 11 findings from the Claude subagent, 30 of 34 spec issues fixed after 3 review rounds, and 21 obligations carried forward with 3 disagreements flagged as taste items. Moving on to Phase 2 (design review) now that UI scope was detected.\n\n',
timestamp: '2026-09-10T00:00:03.893Z',
};
const at = Date.parse(announcement.timestamp);
expect(autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: [announcement] }, at - 1)).toEqual([{ phase: 1, ts: at }]);
});
test('AF bare done announcement retains Phase 2 without crediting its DX transition', () => {
const announcement2 = {
sessionId: "6ce27a04-7422-4687-9de9-138e68f308d8",
text: "Phase 2 done: Claude subagent found 15 issues (1 critical, 8 high, 6 medium), with 14 fully accepted and 1 partially accepted; design score rose from 6/10 to 8.7/10, and 10 accepted items are now carried into the Implementation plan. Moving on to Phase 2.5 (DX Review) since developer-facing scope was detected.",
timestamp: "2026-09-10T00:08:31.892Z",
};
const at = Date.parse(announcement2.timestamp);
expect(autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: [announcement2] }, at - 1)).toEqual([{ phase: 2, ts: at }]);
});
test('AF named DX completion retains Phase 2.5 without crediting its Eng transition', () => {
const announcement3 = {
sessionId: "6ce27a04-7422-4687-9de9-138e68f308d8",
text: "Phase 2.5 (DX review) is done: DX score rose from 5.1 to 8.0/10, all 19 findings reviewed with 12 items accepted into the plan and one taste item flagged for gating. All pre-checks pass, so I'm moving on to Phase 3, the final Engineering Review of the amended plan.\n\n",
timestamp: "2026-09-10T00:16:29.888Z",
};
const at = Date.parse(announcement3.timestamp);
expect(autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: [announcement3] }, at - 1)).toEqual([{ phase: 2.5, ts: at }]);
});
test('completed-state declarations retain supported phases, punctuation and first native time', () => {
for (const state of ['wrapped up', 'done']) for (const phase of [1, 2, 2.5, 3]) for (const tail of ['', '.', ': Work retained.', '. Work retained.']) {
expect(autoplanPhaseCompletions(transcript([1, `Phase ${phase} ${state}${tail}`]), START))
.toEqual([{ phase, ts: START + 1 }]);
}
expect(autoplanPhaseCompletions(transcript([1, '**Phase 1 wrapped up.**']), START))
.toEqual([{ phase: 1, ts: START + 1 }]);
expect(autoplanPhaseCompletions(transcript([4, 'Phase 1 wrapped up.'],
[2, 'Phase 3 wrapped up.'], [3, 'Phase 1 wrapped up: Moving to Phase 2.']), START))
.toEqual([{ phase: 3, ts: START + 2 }, { phase: 1, ts: START + 3 }]);
});
test('affirmative completion words share punctuation and optional is without future tense', () => {
for (const state of ['complete', 'completed', 'done', 'finished', 'wrapped up']) {
for (const copula of ['', 'is ']) for (const tail of ['', '.', ': Work retained.']) {
expect(autoplanPhaseCompletions(transcript([1, `Phase 3 ${copula}${state}${tail}`]), START))
.toEqual([{ phase: 3, ts: START + 1 }]);
}
for (const text of [`Phase 3 ${state}?`, `Phase 3 will be ${state}.`,
`Phase 3 is not ${state}.`, `Phase 3 ${state} if the reviewer finishes.`,
`Phase 3 ${state} when the work ends.`, `**Phase 3 ${state}** if approved.`,
`Example:\nPhase 3 ${state}.`, `> Phase 3 ${state}.`,
`Phase 3 (Eng review) is not ${state}.`, `Phase 3 (Eng review) ${state} if approved.`]) {
expect(autoplanPhaseCompletions(transcript([1, text]), START), text).toEqual([]);
}
}
});
test('known phase names must agree with their number and never supply completion alone', () => {
const names = [[1, 'CEO'], [2, 'Design'], [2.5, 'DX'], [3, 'Eng'], [3, 'Engineering']] as const;
for (const [phase, name] of names) for (const suffix of ['', ' review']) {
const declaration = `Phase ${phase} (${name}${suffix}) is finished.`;
expect(autoplanPhaseCompletions(transcript([1, declaration]), START)).toEqual([{ phase, ts: START + 1 }]);
expect(autoplanPhaseCompletions(transcript([1, `**${declaration}**`]), START)).toEqual([{ phase, ts: START + 1 }]);
for (const other of [1, 2, 2.5, 3].filter(n => n !== phase)) {
expect(autoplanPhaseCompletions(transcript([1, declaration.replace(`Phase ${phase}`, `Phase ${other}`)]), START))
.toEqual([]);
}
}
for (const text of ['Phase 3 (Eng review).', 'Phase 2.5 (future DX review) is done.',
'Phase 2.5 (DX review if approved) is done.', 'Phase 1 (source) complete.',
'Phase 2.5 ((DX review)) is done.', 'Phase 2.5 (DX review) finished soon.',
'# Phase 2.5 (DX review) is done.', 'Example:\nPhase 2.5 (DX review) is done.']) {
expect(autoplanPhaseCompletions(transcript([1, text]), START), text).toEqual([]);
}
});
test('future, conditional, negative and quoted wrap-up claims do not complete a phase', () => {
for (const text of [
'Phase 1 will wrap up.', 'Phase 1 has not wrapped up.', 'Phase 1 is not wrapped up.',
'Phase 1 wrapped up if the reviewer finishes.', 'Phase 1 wrapped up when the review ends.',
'Phase 1 wrapped up but is not complete.', 'Phase 1 wrapped up?',
'Once Phase 1 wrapped up, we would start Phase 2.', 'I will announce Phase 1 wrapped up.',
'**Phase 1 wrapped up** if the tests pass.', 'Phase 4 wrapped up.', 'Phase 2.1 wrapped up.',
'# Phase 1 wrapped up.', '> Phase 1 wrapped up.', '"Phase 1 wrapped up."',
'- Phase 1 wrapped up.', '| Phase 1 wrapped up. |', ' Phase 1 wrapped up.',
'```text\nPhase 1 wrapped up.\n```', '~~~text\nPhase 1 wrapped up.\n~~~',
'Example:\nPhase 1 wrapped up.\nPhase 2 wrapped up.',
'The template says:\n\nPhase 1 wrapped up.',
'**Phase 1 wrapped up.** Emit phase-transition summary:',
]) for (const declaration of [text, text.replace(/wrapped up/g, 'done')]) {
expect(autoplanPhaseCompletions(transcript([1, declaration]), START), declaration).toEqual([]);
}
});
test('wrapped-up declarations retain ready transcript and native timestamp requirements', () => {
const current = transcript([1, 'Phase 1 wrapped up.']);
for (const status of ['missing', 'error'] as const) {
expect(autoplanPhaseCompletions({ ...current, status }, START)).toEqual([]);
}
expect(autoplanPhaseCompletions(current, START + 2)).toEqual([]);
expect(autoplanPhaseCompletions({ ...current, assistantMessages: current.assistantMessages.map(
message => ({ ...message, timestamp: 'invalid' })) }, START)).toEqual([]);
});
test('retains actual completion timestamps when several phases arrive between polls', () => {
expect(autoplanPhaseCompletions(transcript(
[1, '**Phase 1 complete.** Codex: 2 concerns. Native: 3 issues.'],
[2, 'Phase 2 complete. Design outputs are in the plan.'],
[3, '**Phase 2.5 complete.** DX overall: 8/10.'],
[4, 'Phase 3 complete. Both engineering reviews finished.'],
), START)).toEqual([1, 2, 2.5, 3].map((phase, index) => ({ phase, ts: START + index + 1 })));
});
test('duplicate announcements retain their first timestamp without reordering phases', () => {
expect(autoplanPhaseCompletions(transcript(
[3, 'Phase 1 complete.'], [2, 'Phase 2 complete.'],
[1, '**Phase 1 complete.**'], [4, '**Phase 3 complete.**'],
), START)).toEqual([{ phase: 1, ts: START + 1 }, { phase: 2, ts: START + 2 }, { phase: 3, ts: START + 4 }]);
});
test('headings, quoted skill text, examples, tables and planned checklists provide no completion', () => {
for (const text of [
'## Phase 3 complete.',
'**PHASE 3 COMPLETE.** Emit phase-transition summary:',
'> **Phase 3 complete.** Codex: [N concerns].',
'The section says "**Phase 3 complete.**".',
'Example: **Phase 3 complete.**',
'Example announcement:\n**Phase 3 complete.**',
'Example announcement:\nPhase 2 complete.\nPhase 3 complete.',
'The template requires this completion marker:\n\n**Phase 3 complete.**',
' **Phase 3 complete.**',
'\tPhase 3 complete.',
'```markdown\n**Phase 3 complete.**\n```',
'```markdown\n```still-code\nPhase 3 complete.\n```',
'````markdown\n```\nPhase 3 complete.\n````',
'~~~markdown\nPhase 3 complete.\n~~~',
'```markdown\nPhase 3 complete.',
'| **Phase 3 complete.** | pending |',
'- [ ] **Phase 3 complete.**',
'1. Phase 3 complete.',
'I will announce **Phase 3 complete.** after the review.',
'Phase 3 complete when the engineering review ends.',
'Phase 3 complete?',
'Phase 4 complete.',
]) expect(autoplanPhaseCompletions(transcript([1, text]), START), text).toEqual([]);
});
test('a real announcement after a closed example fence still establishes completion', () => {
expect(autoplanPhaseCompletions(transcript([1, 'Example:\n```markdown\nPhase 3 complete.\n```\n\n**Phase 1 complete.**']), START))
.toEqual([{ phase: 1, ts: START + 1 }]);
});
test('accepts a template-compliant quoted transition with actual consensus while rejecting quoted examples', () => {
const actual = '> **Phase 1 complete.** Codex: 2 concerns. Claude subagent: 3 issues.\n' +
'> Consensus: 4/6 confirmed, 2 disagreements → surfaced at gate.\n> Passing to Phase 2.';
expect(autoplanPhaseCompletions(transcript([1, actual]), START)).toEqual([{ phase: 1, ts: START + 1 }]);
expect(autoplanPhaseCompletions(transcript([1, actual.replace('Codex: 2 concerns.', 'Codex: unavailable.')]), START))
.toEqual([{ phase: 1, ts: START + 1 }]);
for (const quoted of [
actual.replace('4/6', 'X/6'),
actual.replace('2 concerns', '[N concerns]'),
'The template contains this example:\n' + actual,
'Example announcement:\n' + actual + '\n' + actual.replace('Phase 1', 'Phase 3'),
'> **Phase 3 complete.**',
]) expect(autoplanPhaseCompletions(transcript([1, quoted]), START), quoted).toEqual([]);
});
test('missing/error transcripts and pre-command declarations cannot establish coverage', () => {
const prior = transcript([-1, '**Phase 3 complete.**']);
expect(autoplanPhaseCompletions(prior, START)).toEqual([]);
for (const status of ['missing', 'error'] as const) {
expect(autoplanPhaseCompletions({ ...transcript([1, '**Phase 3 complete.**']), status }, START)).toEqual([]);
}
});
test('does not repair an out-of-order chain or synthesize omitted phases', () => {
expect(autoplanPhaseCompletions(transcript([1, 'Phase 3 complete.'], [2, 'Phase 1 complete.']), START))
.toEqual([{ phase: 3, ts: START + 1 }, { phase: 1, ts: START + 2 }]);
expect(autoplanPhaseCompletions(transcript([1, 'Phase 2 skipped — no UI scope.']), START)).toEqual([]);
});
test.skipIf(process.platform === 'win32')('ANSI-rendered completions use native evidence while displayed Read/source markers do not', async () => {
const dir = fs.mkdtempSync(path.join(os.tmpdir(), 'autoplan-phase-replay-'));
const fake = path.join(dir, 'fake-claude');
const worker = path.join(dir, 'worker.ts');
const recordFile = path.join(dir, 'events.jsonl');
const resultFile = path.join(dir, 'result.json');
fs.writeFileSync(fake, `#!${process.execPath}\n` + String.raw`
import * as fs from 'node:fs';
import * as path from 'node:path';
fs.writeFileSync(process.env.PHASE_RECORD, JSON.stringify({pid:process.pid}) + '\n');
const sessionId = 'fake-autoplan-session';
const dir = path.join(process.env.CLAUDE_CONFIG_DIR, 'projects', 'fixture');
fs.mkdirSync(dir, {recursive:true});
const base = Date.now();
const write = (role, content, offset, extra = {}) => fs.appendFileSync(path.join(dir, sessionId + '.jsonl'), JSON.stringify({
sessionId, cwd:process.cwd(), isSidechain:false, timestamp:new Date(base + offset).toISOString(), message:{role,content}, ...extra,
}) + '\n');
write('user', [{type:'tool_result', tool_use_id:'read', content:'**Phase 3 complete.**'}], 0);
write('assistant', [{type:'text', text:'> **Phase 3 complete.** is the quoted source marker.'}], 0);
write('assistant', [{type:'text', text:'**Phase 3 wrapped up.**'}], 0, {isSidechain:true});
write('assistant', [{type:'text', text:'**Phase 3 wrapped up.**'}], 0, {cwd:path.join(process.cwd(), 'foreign')});
process.stdin.setRawMode?.(true);
let sent = false;
process.stdin.on('data', data => {
if (sent || !data.toString().includes('\r')) return;
sent = true;
process.stdout.write('\x1b[2J\x1b[H');
[1, 2, 2.5, 3].forEach((phase, index) => {
const message = '**Phase ' + phase + (phase === 1 ? ' wrapped up.**' : phase === 2 ? ' done.**' : phase === 2.5 ? ' (DX review) is finished.**' : ' complete.**');
write('assistant', [{type:'text', text:message}], index + 1);
process.stdout.write('● \x1b[1mPhase ' + phase + ' complete.\x1b[22m\n');
});
process.stdout.write('NATIVE_PHASES_READY\n');
});
process.stdout.write('Read: autoplan/sections/eng-phase.md\n> **Phase 3 complete.**\nSOURCE_READY\n');
process.on('SIGINT', () => process.exit(0));
process.stdin.resume();
`);
fs.chmodSync(fake, 0o755);
const moduleUrl = (name: string) => pathToFileURL(path.resolve(import.meta.dir, 'helpers', name)).href;
fs.writeFileSync(worker, `
import { launchClaudePty } from ${JSON.stringify(moduleUrl('claude-pty-runner.ts'))};
import { readPlanCountTranscript } from ${JSON.stringify(moduleUrl('plan-count-transcript.ts'))};
import { autoplanPhaseCompletions } from ${JSON.stringify(moduleUrl('autoplan-phase-observer.ts'))};
const start = Date.now();
const session = await launchClaudePty({cwd:${JSON.stringify(dir)}, timeoutMs:8000, env:{PHASE_RECORD:${JSON.stringify(recordFile)}}});
try {
await session.waitFor('SOURCE_READY', {timeoutMs:4000, pollMs:20});
const read = () => readPlanCountTranscript(session.hermeticConfigDir, ${JSON.stringify(dir)});
const sourceOnly = autoplanPhaseCompletions(read(), start);
const oldPattern = /\\*\\*Phase\\s+(\\d+(?:\\.\\d+)?)\\s+complete\\.?\\*\\*/g;
const sourceFalsePositive = [...session.visibleText().matchAll(oldPattern)].length;
const since = session.mark();
session.send('\\r');
await session.waitFor('NATIVE_PHASES_READY', {timeoutMs:4000, pollMs:20});
const rendered = session.visibleSince(since);
const oldPatternMissed = [...rendered.matchAll(oldPattern)].length;
const hits = autoplanPhaseCompletions(read(), start);
await Bun.write(${JSON.stringify(resultFile)}, JSON.stringify({sourceOnly, sourceFalsePositive, oldPatternMissed, hits, rendered}));
} finally { await session.close(); }
`);
const child = Bun.spawn([process.execPath, worker], {
env: { ...process.env, BROWSE_TERMINAL_BINARY: fake, EVALS_HERMETIC: '1' }, stdout: 'pipe', stderr: 'pipe',
});
const timer = setTimeout(() => child.kill('SIGKILL'), 12_000);
try {
const [code, stdout, stderr] = await Promise.all([child.exited, new Response(child.stdout).text(), new Response(child.stderr).text()]);
expect(code, stdout + stderr).toBe(0);
const result = JSON.parse(fs.readFileSync(resultFile, 'utf8'));
expect(result.sourceOnly).toEqual([]);
expect(result.sourceFalsePositive).toBe(1);
expect(result.oldPatternMissed).toBe(0);
expect(result.rendered).toContain('Phase 3 complete.');
expect(result.hits.map((hit: { phase: number }) => hit.phase)).toEqual([1, 2, 2.5, 3]);
for (let index = 1; index < result.hits.length; index++) {
expect(result.hits[index].ts).toBeGreaterThan(result.hits[index - 1].ts);
}
const pid = JSON.parse(fs.readFileSync(recordFile, 'utf8').split('\n')[0]!).pid;
expect(() => process.kill(pid, 0)).toThrow();
} finally {
clearTimeout(timer); child.kill('SIGKILL');
if (fs.existsSync(recordFile)) {
const pid = JSON.parse(fs.readFileSync(recordFile, 'utf8').split('\n')[0]!).pid;
try { process.kill(pid, 'SIGKILL'); } catch { /* already reaped */ }
}
fs.rmSync(dir, {recursive:true, force:true});
}
}, 15_000);
});
describe('autoplan-phase-dash-ao', () => {
const fixture = fixture_autoplan_phase_dash_ao;
const at = Date.parse(fixture.message.timestamp);
const transcript = (text = fixture.message.text): PlanCountTranscript => ({
status: 'ready', calls: [], assistantMessages: [{ ...fixture.message, text }],
});
const hits = (text: string) => autoplanPhaseCompletions(transcript(text), at - 1);
test('exact owned DX dash declaration adds only DX at its native timestamp', () => {
expect(hits(fixture.message.text)).toEqual([{ phase: 2.5, ts: at }]);
const all = autoplanPhaseCompletions({ status: 'ready', calls: [],
assistantMessages: fixture.orderedMessages }, fixture.commandLowerBound);
expect(all).toEqual([...fixture.actualHits, { phase: 2.5, ts: at }]);
expect(all.map(hit => hit.phase)).toEqual([1, 2, 2.5]);
});
test('em and en dash spacing share the existing completed declaration forms', () => {
for (const dash of ['—', '–']) for (const before of ['', ' ']) for (const after of ['', ' ']) {
expect(hits(fixture.message.text.replace('complete—', `complete${before}${dash}${after}`)))
.toEqual([{ phase: 2.5, ts: at }]);
for (const phase of [1, 2, 2.5, 3]) for (const state of ['complete', 'completed', 'done', 'finished', 'wrapped up']) {
expect(hits(`Phase ${phase} is ${state}${before}${dash}${after}Work retained.`))
.toEqual([{ phase, ts: at }]);
}
}
expect(hits('**Phase 2.5 complete** — Work retained.')).toEqual([{ phase: 2.5, ts: at }]);
});
test('dash continuations cannot turn a conditional, quotation, question or denial into completion', () => {
for (const dash of ['—', '–']) for (const tail of [
'', 'if approved.', 'unless the checks fail.', 'when review finishes.',
'once the reviewer signs off.', 'pending final checks.', 'maybe tomorrow.',
'perhaps it is complete.', 'would be complete after review.',
'not complete yet.', 'the phase is not complete.', 'this completion is withdrawn.',
'actually never finished.', 'this completion is superseded.',
'provided the remaining checks pass.', 'this completion is rejected.',
'the completion announcement is retracted.', 'actually incomplete.',
'the review remains pending.', 'Work retained?', 'is this complete?',
'Source excerpt: Work retained.', 'Earlier review: Work retained.',
'the historical example says work is retained.', '"Work retained."',
]) expect(hits(`Phase 2.5 complete ${dash} ${tail}`), tail).toEqual([]);
for (const text of [
'If approved, Phase 2.5 complete—Work retained.',
'Phase 2.5 is not complete—Work retained.',
'Phase 2.5 complete?—Work retained.',
'> Phase 2.5 complete—Work retained.',
'"Phase 2.5 complete—Work retained."',
'Source excerpt:\nPhase 2.5 complete—Work retained.',
'Example:\nPhase 2.5 complete—Work retained.\nPhase 3 complete—Work retained.',
'```text\nPhase 2.5 complete—Work retained.\n```',
' Phase 2.5 complete—Work retained.',
'# Phase 2.5 complete—Work retained.',
'Phase 2.5 (Eng review) complete—Work retained.',
]) expect(hits(text), text).toEqual([]);
});
test('dash support keeps ready/current native evidence and first-hit ordering', () => {
for (const status of ['missing', 'error'] as const) {
expect(autoplanPhaseCompletions({ ...transcript(), status }, at - 1)).toEqual([]);
}
expect(autoplanPhaseCompletions(transcript(), at + 1)).toEqual([]);
expect(autoplanPhaseCompletions({ ...transcript(), assistantMessages: [
{ ...fixture.message, timestamp: 'invalid' },
] }, at - 1)).toEqual([]);
const later = { ...fixture.message, timestamp: new Date(at + 1).toISOString() };
expect(autoplanPhaseCompletions({ ...transcript(), assistantMessages: [later, fixture.message] }, at - 1))
.toEqual([{ phase: 2.5, ts: at }]);
});
});
describe('autoplan-with-result-au', () => {
const actual = actual_autoplan_with_result_au;
const at=Date.parse(actual.timestamp);
const transcript=(text=actual.text):PlanCountTranscript=>({status:'ready',calls:[],assistantMessages:[{...actual,text}]});
const observe=(text:string)=>autoplanPhaseCompletions(transcript(text),at-1);
test('the exact first AU DX completion retains its native timestamp without crediting the Eng transition',()=>{
expect(autoplanPhaseCompletions(transcript(),at-1)).toEqual([{phase:2.5,ts:at}]);
expect(actual.sessionId).toBe('78ce9c42-e5f7-4595-81ea-7d9bb8b4345c');
expect(actual.timestamp).toBe('2026-09-10T21:35:46.209Z');
});
test('affirmative result clauses share phase identity and the existing completion vocabulary',()=>{
for(const [phase,name] of [[1,'CEO'],[2,'Design review'],[2.5,'DX'],[3,'Engineering review']] as const)
for(const state of ['complete','completed','done','finished','wrapped up'])
for(const result of ['22 findings recorded in the plan.','the score at 8/10.','all adopted changes written; moving to the next phase.']) {
expect(observe(`Phase ${phase} (${name}) is ${state} with ${result}`)).toEqual([{phase,ts:at}]);
}
expect(observe('**Phase 2.5 wrapped up** with 22 findings retained.')).toEqual([{phase:2.5,ts:at}]);
});
const rejected=[
'Phase 2.5 wrapped up with ',
'Phase 2.5 wrapped up without the review.',
'Phase 2.5 will be complete with 22 findings.',
'Phase 2.5 is not complete with 22 findings.',
'Phase 2.5 complete with no completed review.',
'Phase 2.5 complete with findings still pending.',
'Phase 2.5 complete with 22 findings if the review finishes.',
'Phase 2.5 complete with 22 findings once approved.',
'Phase 2.5 complete with 22 findings when the review ends.',
'Phase 2.5 complete with 22 findings unless the review fails.',
'Phase 2.5 complete with 22 findings provided the reviewer agrees.',
'Phase 2.5 complete with 22 findings?','Phase 2.5 complete with results that will arrive tomorrow.',
'Phase 2.5 complete with maybe 22 findings.','Phase 2.5 complete with an unfinished review.',
'Phase 2.5 complete with 22 findings. This phase is withdrawn.',
'Phase 2.5 complete with 22 findings. This phase is "withdrawn".',
'Phase 2.5 complete with 22 findings. This phase is not complete.',
'Phase 2.5 complete with 22 findings. This phase is retracted.',
'Phase 2.5 complete with 22 findings. The declaration is superseded.',
'Phase 2.5 complete with a historical example.',
'Phase 2.5 complete with source instructions.',
'Phase 2.5 complete with "22 findings recorded".',
'Phase 2.5 complete with \'22 findings recorded\'.',
'Phase 2.5 (Design) complete with 22 findings.',
'Phase 2.5 (DX review if approved) complete with 22 findings.',
'Phase 4 complete with 22 findings.','Phase 2.1 complete with 22 findings.',
'# Phase 2.5 complete with 22 findings.',
'> Phase 2.5 complete with 22 findings.',
'"Phase 2.5 complete with 22 findings."',
'- Phase 2.5 complete with 22 findings.',
'| Phase 2.5 complete with 22 findings. |',
' Phase 2.5 complete with 22 findings.',
'\tPhase 2.5 complete with 22 findings.',
'```text\nPhase 2.5 complete with 22 findings.\n```',
'~~~text\nPhase 2.5 complete with 22 findings.\n~~~',
'Source:\nPhase 2.5 complete with 22 findings.',
'Historical example:\nPhase 2.5 complete with 22 findings.',
'Historical review:\nPhase 2.5 complete with 22 findings.',
'**Historical review:**\nPhase 2.5 complete with 22 findings.',
'**Source:**\nPhase 2.5 complete with 22 findings.',
'Hypothetical scenario:\nPhase 2.5 complete with 22 findings.',
'Phase 2.5 complete with 22 findings.\n```text\nexample text\n````\nThis phase is withdrawn.',
'Earlier review:\nPhase 2.5 complete with 22 findings.',
'Phase 2.5 complete with a hypothetical 8/10 score.',
'Phase 2.5 complete with 22 findings.\nThis phase is withdrawn.',
'Phase 2.5 complete with 22 findings.\nThis phase is \"withdrawn\".',
'Phase 2.5 complete with 22 findings.\n**Phase 2.5** is ‘withdrawn’.',
'Phase 2.5 complete with 22 findings.\nCurrent status: this phase is no longer current.',
'The template says:\n\nPhase 2.5 complete with 22 findings.',
'Example:\nPhase 1 complete with findings.\nPhase 2.5 complete with findings.',
];
test.each(rejected)('%s cannot supply completion',text=>expect(observe(text)).toEqual([]));
test('quoted summaries retain their existing concrete-consensus requirement',()=>{
const summary='> Phase 2.5 complete with 22 findings retained.\n> Consensus: 22/22 accepted.\n> Moving to Phase 3.';
expect(observe(summary)).toEqual([{phase:2.5,ts:at}]);
for(const text of [summary.replace('22/22','[N]/22'),'Example:\n'+summary,summary.replace('22/22','X/Y')])
expect(observe(text)).toEqual([]);
});
test('native readiness, timestamp, duplicate and observed-order rules remain intact',()=>{
for(const status of ['missing','error'] as const)
expect(autoplanPhaseCompletions({...transcript(),status},at-1)).toEqual([]);
expect(autoplanPhaseCompletions(transcript(),at+1)).toEqual([]);
expect(autoplanPhaseCompletions({...transcript(),assistantMessages:[{...actual,timestamp:'invalid'}]},at-1)).toEqual([]);
const data=transcript();data.assistantMessages.push({...actual,timestamp:new Date(at+1).toISOString()});
expect(autoplanPhaseCompletions(data,at-1)).toEqual([{phase:2.5,ts:at}]);
data.assistantMessages.unshift({...actual,text:'Phase 3 complete with 7 findings retained.',timestamp:new Date(at-10).toISOString()});
expect(autoplanPhaseCompletions(data,at-11)).toEqual([{phase:3,ts:at-10},{phase:2.5,ts:at}]);
});
test('quoted history and a foreign phase withdrawal do not cancel the current completed result',()=>{
for(const suffix of [
'> This phase is withdrawn.',
'Historical note: "This phase is withdrawn."',
'Example:\nThis phase is withdrawn.',
'```text\nThis phase is withdrawn.\n```',
'Phase 2 is withdrawn.',
'Phase 3 complete.\nThis phase is withdrawn.',
]) expect(observe('Phase 2.5 complete with 22 findings retained.\n'+suffix).some(hit=>hit.phase===2.5)).toBe(true);
expect(observe('Phase 2.5 complete with 22 findings.\nHistorical note:\nThis phase is withdrawn.\nCurrent status: Phase 2.5 is withdrawn.')).toEqual([]);
});
test('a current Markdown status heading resets historical context for an owned withdrawal',()=>{
const prefix='Phase 2.5 complete with 22 findings retained.\nHistorical note:\nThis phase is withdrawn.\n';
expect(observe(prefix+'## Current status\nPhase 2.5 is withdrawn.')).toEqual([]);
expect(observe(prefix+'`## Current status`\nThis phase is withdrawn.')).toEqual([{phase:2.5,ts:at}]);
});
test('inline code around an owned status is scalar formatting while a whole quoted statement stays literal',()=>{
const prefix='Phase 2.5 complete with 22 findings retained.\n';
expect(observe(prefix+'This phase is `withdrawn`.')).toEqual([]);
for(const literal of ['`This phase is withdrawn.`','"This phase is withdrawn."','```text\nThis phase is withdrawn.\n```'])
expect(observe(prefix+literal)).toEqual([{phase:2.5,ts:at}]);
});
});