Files
gstack/test/evals-workflow-matrix.test.ts
T
Garry TanandClaude Fable 5 a3749bfa4b v1.70.1.0 fix: ship names the /document-release subagent at every decision point (tripwire + gate E2E) (#2700)
* fix(ship): name the /document-release subagent at every Step 18 decision point

The v1.54.0.0 carve moved Step 18 (documentation sync) into
ship/sections/pr-body.md and the Claude-host skeleton stopped saying
"document-release" anywhere in the workflow body — the dispatch became
invisible at exactly the moments an agent decides whether to open the
section. Restore visibility at three touchpoints, all subagent-framed
(never bare-slash-framed, which would invite an inline Skill invocation
that bypasses the fresh-context subagent + JSON contract):

- manifest trigger (renders into the section-index row AND the STOP
  pointer): "dispatching the /document-release subagent to sync docs
  (Step 18) and then creating or updating the PR/MR (Step 19)"
- Step 17 handoff line names Step 18's dispatch explicitly
- new hoisted doc-sync invariant beside the PR-title invariant: the
  dispatch itself is never skipped; only a failed subagent is
  non-blocking

Pin it in carve-guards: 'the /document-release subagent' (all three
touchpoints) + 'dispatches the /document-release subagent' (invariant)
must stay in the skeleton; the carved imperative 'Dispatch
/document-release as a subagent' must stay carved. Skeleton cap
91,600 → 92,300 (measured 91,764; trigger renders twice). Goldens
regenerated for all three hosts.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: pin the ship→document-release Step 18 wiring with a free tripwire

Five substring/structure asserts across the carved section, the Claude
skeleton's three touchpoints, the manifest trigger, and the codex/factory
goldens (inlined Step 18 ordered before Step 19). Claude-golden asserts
deliberately omitted: host-config.test.ts already enforces golden ==
generated byte-for-byte.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* test: gate-tier E2E proving /ship dispatches the document-release subagent

New skill-e2e-ship-docsync: a live agent gets the sliced Step 17→19 tail
of the generated ship skeleton in a bare-remote git fixture (Steps 0-16
"done"), under a fake HOME so the STOP pointer and the Step 18 subagent
prompt resolve to planted copies, with a stub document-release skill that
returns the empty-result JSON contract. Hard assert: an Agent/Task
tool-call matching /document-release/i exists in result.toolCalls and
precedes any `gh pr create`. Neutral prompt (no STOP-Read priming, no
document-release mention — the prompt echoes into the transcript, so
asserts read toolCalls only).

Hardening from review: throw-on-marker-drift fixture slice; per-test
GSTACK_HOME + .redact-prepush-prompted marker (routes Step 17's
credential guard to its silent branch — the hermetic GSTACK_HOME pin
defeats a HOME-only override); 480s/540s timeouts (nested subagent adds
wall clock the 300s sibling never carried); 'timeout' accepted in
exitReason only because the dispatch assert is independently hard;
whole-file describeE2ETier('gate') composed with diff selection (keeps
the file out of the periodic shard census, which sits at its ceiling,
and under the hard tier-alignment invariant).

Registered as 'ship-docsync' in E2E_TOUCHFILES + E2E_TIERS (gate) in the
same commit — touchfiles.test.ts rejects either half landing first.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: fix stale document-release TODOS entry + three review-deferred items

The SHIPPED entry still described the deleted Step 8.5 post-PR cat-delegation
design from v0.8.4; replace with the current Step 18 subagent design and its
test pins. Add the three P3 items deferred from the v1.69 plan review:
dispatch receipt enforcement, land-and-deploy→canary dispatch-pin pattern,
and the periodic shard-census boundary.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: pre-landing review fixes

Testing-specialist findings, all mechanical: (1) pin the E2E fixture's git
branch (-b main / init.defaultBranch=main) and assert every setup command's
exit status so operator git config can't silently corrupt a paid run;
(2) tighten the dispatch matcher to Step 18-prompt-specific markers
(document-release/SKILL.md | executing the /document-release workflow) so a
subagent merely quoting section text can't false-pass the regression assert
(verified against recorded burn-in transcripts); (3) replace the subsumed
carve-guards anchor with three non-overlapping per-touchpoint anchors
(gerund/imperative/3rd-person) so each touchpoint is independently enforced.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: red-team review fixes

Five informational findings: TODOS shard-census arithmetic corrected (census
is 67 with one free ungated slot; the SECOND ungated file trips the floor)
and version pointer fixed (v0.18.2.0, not v0.18.1.0); the free tripwire now
pins the two dispatch-matcher marker strings so a pr-body prompt reword
fails the free suite instead of surfacing as a paid-tier mystery; the E2E
matcher gains a section-paste exclusion (scaffold strings disqualify) —
verified against all recorded runs; the E2E header documents the tierless
test:evals invisibility tradeoff.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: adversarial review fixes

Pin the E2E matcher's two EXCLUSION markers in the free tripwire (an
unpinned 'Parent processing:' reword would silently deaden the
section-paste guard while every test stayed green); add an ordering pin
(the hoisted doc-sync invariant must sit above the pr-body STOP pointer —
presence-only anchors can't catch drift below it); plant a third
cwd-relative pr-body copy inside the fixture repo, gitignored so the agent
never tries to commit test scaffolding.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* chore: bump version and changelog (v1.70.1.0)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* docs: CHANGELOG accuracy fixes from the doc-release review

Three factual corrections the Step 18 doc subagent caught in the fresh
v1.70.1.0 entry: 5 tripwire tests (not 6), cost floor $0.63 per the cited
eval store (not $0.59), and the visibility claim scoped to decision points
(the re-run checklist mention survived the carve). Plus the E2E header's
stale pending-burn-in note replaced with the observed numbers.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix: raise bun-polyfill subprocess budget to 60s for degraded Windows runners

The 50ms-sleep test blew the 20s budget on BOTH bun retry attempts on PR
#2700's windows-latest runner (run 32989821401) — sustained AV/runner
pressure, not just the documented cold-start. Same flake passed-on-rerun on
the prompt-token-load-reduction branch yesterday. Budget only; every
assertion still checks exact output.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(ci): run the ship-docsync gate E2E in the evals matrix + silent-skip tripwire

The evals.yml matrix is hand-enumerated and the Run step never exported
EVALS_TIER, so the new whole-file-gated ship-docsync E2E would have
self-skipped even with a row — a hollow green one layer deeper than the
documented rehomed-monolith incident. Add the e2e-ship-docsync row with a
row-level `tier: gate` property, exported as EVALS_TIER by the Run step
(empty = unset for every existing row: all readers are `=== '<tier>'` or
truthiness).

New free tripwire test/evals-workflow-matrix.test.ts ratchets the class:
matrix files must exist; gate-hosting files must have a row; whole-file-gated
matrix files must carry a matching row tier; and the burn-down lists enforce
their own cleanup. It enumerates the PRE-EXISTING holes found while wiring
this (8 gate-hosting files with no row; codex/gemini rows running zero tests;
the pty-plan-smoke row hollow since its files adopted describeE2ETier) —
tracked in TODOS as the CI gate-lane hollow-coverage burn-down.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-27 08:46:41 -07:00

197 lines
8.3 KiB
TypeScript

/**
* CI eval-matrix completeness tripwire — kills the silent-skip class where a
* gate-tier test exists in the repo but the hand-enumerated matrix in
* .github/workflows/evals.yml never runs it, so "gate tier blocks merge" is
* quietly false in CI. This has happened before (see the "rehomed from the
* deleted pre-split monolith" comment in evals.yml) and was found again on
* PR #2700: nine gate-hosting files absent from the matrix, plus matrix rows
* whose whole-file tier guards can never fire because the Run step exported
* no EVALS_TIER.
*
* Ratchet, not amnesty: the KNOWN_* lists below enumerate the PRE-EXISTING
* gaps with reasons, so no NEW gap can land while the backlog burns down
* (same pattern as SCANNER_EXEMPT in egress-receipt-wiring). If you fix a
* listed gap (add its matrix row / tier property), this test FAILS until you
* remove the entry — stale exemptions are enforced, not decorative.
*
* Wiring pinned:
* - every matrix `file:` path exists on disk (no stale rows),
* - every gate-hosting paid file (whole-file gate self-gate, or named in the
* dep list of a gate-tier E2E_TOUCHFILES key) appears in the matrix or in
* KNOWN_MATRIX_GAPS,
* - every matrix file with a whole-file tier guard has a matching row-level
* `tier:` property (else the suite self-skips and the job is hollow-green)
* or sits in KNOWN_TIER_UNSET.
*/
import { describe, test, expect } from 'bun:test';
import * as fs from 'fs';
import * as path from 'path';
import { E2E_TOUCHFILES, E2E_TIERS } from './helpers/touchfiles-data';
import { isPaidTestFile } from './helpers/paid-test-set';
const ROOT = path.join(import.meta.dir, '..');
const WORKFLOW = path.join(ROOT, '.github', 'workflows', 'evals.yml');
/**
* Pre-existing gate-hosting files with no matrix row (found 2026-08-26,
* PR #2700). Adding a row activates real paid runs on every PR — a cost and
* flake-surface decision per file, tracked in TODOS.md ("CI gate-lane
* hollow-coverage burn-down"). Fix = add a matrix row (plus `tier: gate` when
* the file is whole-file gated), then DELETE the entry here.
*/
const KNOWN_MATRIX_GAPS = new Set([
'test/skill-e2e-ask-user-question-format-compliance.test.ts',
'test/skill-e2e-hermetic-canary.test.ts',
'test/skill-e2e-ios.test.ts',
'test/skill-e2e-plan-ceo-finding-floor.test.ts',
'test/skill-e2e-plan-ceo-plan-mode.test.ts',
'test/skill-e2e-plan-design-with-ui.test.ts',
'test/skill-e2e-plan-devex-finding-floor.test.ts',
'test/skill-e2e-plan-devex-plan-mode.test.ts',
]);
/**
* Matrix files whose whole-file tier guard has no matching row `tier:`
* property (pre-existing, found 2026-08-26). Consequences today:
* - codex-e2e / gemini-e2e declare 'periodic' → both jobs run ZERO tests and
* report green on every PR (vestigial rows; the periodic cron lane owns
* these suites).
* - the two PTY plan-mode smokes declare 'gate' → the e2e-pty-plan-smoke job
* spends ~7 min on container setup and skill registration, then bun test
* skips every describe — hollow-green since the files adopted
* describeE2ETier.
* Fixing either means deliberately (re)activating paid suites on every PR —
* tracked in the same TODOS burn-down. Fix = add `tier:` to the row (or
* delete the vestigial row), then DELETE the entry here.
*/
const KNOWN_TIER_UNSET = new Map([
['test/codex-e2e.test.ts', 'periodic'],
['test/gemini-e2e.test.ts', 'periodic'],
['test/skill-e2e-office-hours-auto-mode.test.ts', 'gate'],
['test/skill-e2e-plan-mode-no-op.test.ts', 'gate'],
]);
interface MatrixRow {
name: string;
files: string[];
tier?: string;
}
/** Parse the `matrix: suite:` rows (name / file / optional tier) from evals.yml. */
function parseMatrixRows(source: string): MatrixRow[] {
const rows: MatrixRow[] = [];
let current: MatrixRow | null = null;
for (const line of source.split('\n')) {
const name = line.match(/^\s+- name: (\S+)\s*$/);
if (name) {
if (current) rows.push(current);
current = { name: name[1], files: [] };
continue;
}
if (!current) continue;
const file = line.match(/^\s+file: (.+?)\s*$/);
if (file) current.files.push(...file[1].trim().split(/\s+/));
const tier = line.match(/^\s+tier: (\S+)\s*$/);
if (tier) current.tier = tier[1];
// `steps:` ends the strategy block — stop before step-level keys leak in.
if (/^\s{4}steps:\s*$/.test(line)) break;
}
if (current) rows.push(current);
return rows.filter((r) => r.files.length > 0);
}
const wholeFileTier = (source: string): string | null => {
const m =
/\b(?:describeE2ETier|e2eTierEnabled)\(\s*['"`](gate|periodic)['"`]/.exec(source) ||
/EVALS_TIER\s*===\s*['"`](gate|periodic)['"`]/.exec(source);
return m ? m[1] : null;
};
const workflowSource = fs.readFileSync(WORKFLOW, 'utf-8');
const rows = parseMatrixRows(workflowSource);
const matrixFiles = new Map<string, MatrixRow>();
for (const row of rows) for (const f of row.files) matrixFiles.set(f, row);
const paidFiles = fs
.readdirSync(path.join(ROOT, 'test'))
.filter((f) => f.endsWith('.test.ts'))
.map((f) => `test/${f}`)
.filter(isPaidTestFile);
describe('evals.yml matrix completeness (gate-lane silent-skip tripwire)', () => {
test('matrix parse sanity: rows and known suites present', () => {
expect(rows.length).toBeGreaterThanOrEqual(15);
expect(matrixFiles.has('test/skill-e2e-workflow.test.ts')).toBe(true);
expect(matrixFiles.has('test/skill-e2e-ship-docsync.test.ts')).toBe(true);
});
test('every matrix file exists on disk', () => {
const missing = [...matrixFiles.keys()].filter(
(f) => !fs.existsSync(path.join(ROOT, f))
);
expect(missing).toEqual([]);
});
test('every gate-hosting paid file is in the matrix (or the documented backlog)', () => {
const gaps: string[] = [];
for (const file of paidFiles) {
const source = fs.readFileSync(path.join(ROOT, file), 'utf-8');
const declaresGate = wholeFileTier(source) === 'gate';
const inGateDeps = Object.entries(E2E_TOUCHFILES).some(
([key, deps]) =>
(E2E_TIERS as Record<string, string>)[key] === 'gate' &&
(deps as string[]).includes(file)
);
if (!declaresGate && !inGateDeps) continue;
if (matrixFiles.has(file) || KNOWN_MATRIX_GAPS.has(file)) continue;
gaps.push(file);
}
expect(
gaps,
`Gate-hosting test file(s) missing from the evals.yml matrix — CI will ` +
`never run them and "gate tier blocks merge" becomes silently false. ` +
`Add a matrix row (with tier: gate when the file is whole-file gated). ` +
`Do NOT extend KNOWN_MATRIX_GAPS for new files.`
).toEqual([]);
});
test('matrix rows for whole-file-gated files carry a matching tier property', () => {
const mismatches: string[] = [];
for (const [file, row] of matrixFiles) {
if (!fs.existsSync(path.join(ROOT, file))) continue;
const declared = wholeFileTier(fs.readFileSync(path.join(ROOT, file), 'utf-8'));
if (!declared) continue;
if (row.tier === declared) continue;
if (KNOWN_TIER_UNSET.get(file) === declared && row.tier === undefined) continue;
mismatches.push(`${file} declares '${declared}' but row '${row.name}' has tier: ${row.tier ?? 'unset'}`);
}
expect(
mismatches,
`A whole-file tier guard with no matching row tier means the suite ` +
`self-skips and the CI job reports a hollow green. Set tier: <declared> ` +
`on the row (the Run step exports it as EVALS_TIER).`
).toEqual([]);
});
test('burn-down lists hold only live gaps (ratchet cleanup enforcement)', () => {
const staleGaps = [...KNOWN_MATRIX_GAPS].filter(
(f) => matrixFiles.has(f) || !fs.existsSync(path.join(ROOT, f))
);
expect(
staleGaps,
'Entry fixed or file removed — delete it from KNOWN_MATRIX_GAPS.'
).toEqual([]);
const staleTiers = [...KNOWN_TIER_UNSET.entries()].filter(([f, declared]) => {
const row = matrixFiles.get(f);
if (!row) return true; // row deleted — entry no longer applies
if (row.tier === declared) return true; // fixed — entry must go
if (!fs.existsSync(path.join(ROOT, f))) return true;
return wholeFileTier(fs.readFileSync(path.join(ROOT, f), 'utf-8')) !== declared;
});
expect(
staleTiers.map(([f]) => f),
'Entry fixed, row removed, or guard changed — delete it from KNOWN_TIER_UNSET.'
).toEqual([]);
});
});