mirror of
https://github.com/garrytan/gstack.git
synced 2026-11-02 15:04:37 +01:00
v1.91.30.0 fix: honest eval failure reports, Bun floor with safe auto-update, make-pdf and /review on Codex (#3033)
* docs: approved reliability follow-ups plan (Oct 2026)
* D1: score overlay wrong answers as comparison results (contract v4)
A completed answer that fails the fixture's output check is now a valid
measurement with taskCorrect false and answerError set, in either arm.
Only execution, scope and metric failures invalidate a trial. ON
correctness still gates the case, so an OFF wrong answer stays in the
comparison. OVERLAY_CONTRACT moves to v4; v3 verdicts are not
reinterpreted. Replays the 37198445662 sonnet trials (OFF-0 wrong).
* C4: score auq-matrix substance with a 3-sample judge panel
Each skill keeps one native capture in the same execution; the
recommendation-substance grader now runs as a JUDGE_PANEL_SAMPLES panel on
that capture and gates on the mean against the unchanged minimum of 4. A
failed sample fails the panel without resampling. Approved at the gate
2026-10-04 for censuses after this change.
* E1: one Bun version source; setup refuses below the 1.3.3 floor, updater checks before advancing
bin/gstack-bun-version.sh holds the security floor (1.3.3) and the tested
version (1.4.0, the CI pin and engines.bun). setup refuses Bun below the
floor before writing anything, warns between floor and tested, warns and
continues on an unparseable version, and prints a 1.4.0 install hint.
Prereleases sort below their release (semver).
gstack-session-update now fetches, reads the incoming release's floor and
checks the bun setup will run before fast-forwarding to that exact
revision; below it the update is held with reason bun-too-old and a
resume command, leaving the revision and installs untouched.
Fixture copies of setup (team-mode, claude-code migration) now include the
sourced file; the slow-git lock test sleeps on fetch, the updater's
network step.
* E3: record the owned Chromium's PID over CDP so the #2709 reap runs on Playwright 1.62
Playwright 1.62's Browser has no process(), so chromiumProcInfo was always
null and browse stop never reaped a surviving headless Chromium. The owned
headless launch now reads the browser PID from CDP SystemInfo.getProcessInfo
(bounded at 2s; a hung or failed call logs once and records nothing). The
reaper's start-time and cmdline identity checks are unchanged, so a recycled
PID or an externally owned browser is never signalled.
* A1: ship docs gate records post-child hashes with gstack-docs-candidate, saves each record once
Stored transcripts (37198445662 gate-census-2) show the parent's reads of
release files take ~1 s; the parent's own model time dominates (completion
268 s of 392 s, store ~270 s of 387 s), and 45-78 s of it is retyping hash
maps, child output and records into several artifacts after the child
returns. The child contract is unchanged.
* A2: /plan-ceo-review sends the mode handoff line before any other tool call
plan-ceo-mode-routing's HOLD SCOPE session spends its 240 s post-selection
window on model work: 37176837432 skipped the handoff helper and paraphrased
the mode, 37198445662 ran it, then wrote the plan, loaded sections and probed
Codex for 114 s before the posture chat. The handoff line is the posture.
* B3: accept wrapped unchanged rows in the cropped Edit preview
PR run 37176835584 timed out with the floor's owned Edit menu on screen:
unchanged diff rows soft-wrap under their gutter with a blank marker
column, and a trimmed empty unchanged row keeps only its number. The
parser accepted wrapped +/- rows only, so it never verified the owned
edit. Both forms are now parsed and still compared with the exact source
lines; the captured viewport is replayed with negative controls.
* A3: record per-phase headroom diagnosis for docsync, plan-mode-no-op, mode-routing and gate slice 11
* E4: /review Step 1 keeps going when the base fetch fails and reports stale base coverage
Under Codex's read-only .git, 'git fetch && merge-base && diff --stat'
stopped at the fetch, so the empty output read as 'Nothing to review'.
The fetch is now separate, prints BASE_REFRESH fresh or stale with the
local base revision, and the review states stale coverage.
* D2: reword the format-approach question prompt
Both plan-ceo-review-format-approach refusals in census 37195203538
(Fable 5.1 model_refusal_no_fallback, category reasoning_extraction:
"duplicating model outputs") were anchored on the session prompt, whose
capture instruction asked for the "verbatim" question and "the exact
prose you would have shown" in a "format-capture test". The approach
case now asks for the same question in the same file as the step's
output, without that wording. The other three plan-format cases keep
their unchanged prompt (no refusals in 72 trials).
* C2: /qa defines its outcome filename fields; QA report templates name the measured timing fields
Low judge samples in failing panels named both: qa workflow (37186854666)
'an undefined {user} in the outcome filename', and qa-only workflow
(37198445662) 'template asks for Duration / durationMs totals while
reporting.md mandates Probe budget and Guarded command time'. The qa
union stays under its 1.11 size ratio (93292 of 93294 bytes).
* 0.1: session ledger written by the claude -p, Agent SDK, PTY and Codex runners
Each armed model session appends one best-effort line to
$GSTACK_EVAL_DIR/session-ledger.jsonl: a stable session key, the timeout it
armed, monotonic elapsed time, its end reason (completed, session_timeout,
observer_timeout, api_error, refusal, aborted, error), one sanitized evidence
line and, for stream-JSON runners, a compact liveness summary (longest
in-request silence with nothing outstanding, last event kind, tool open at
end). Partial messages now stream for every claude -p session and for SDK
sessions without a custom query provider (runSharedInteractive opts in); their
rows feed only the summary, so transcripts keep their size unless
publicStreamDiagnostics is set. Structured refusal and API-error signals win
over phrases, and tool results are never read for them.
* 0.2: failure_cause beside failure_class, from runner facts and the session ledger
One pure failureCauseOf() beside failureClassOf() sets failure_cause
(contract > pre_turn_infra > api_error > refusal > provider_stall >
session_timeout > observer_timeout > assertion > unknown; no machine-set
'detector') with one bounded evidence line, and failureDetailOf() keeps a
Bun matcher's Expected/Received or a judge's failing dimensions in
failure_detail. Both record builders (classifyTrialShard, junitCensus) call
them and attach the trial's ledger sessions. failure_class, error and
TRIAL_ERROR_MAX are computed exactly as before; the new fields are optional,
bounded and validated only when present, and a frozen copy of the v1 reader
accepts every new record. A provider stall needs partial messages and
STALL_WINDOW_MS (120 s, eval-budgets.ts beside HEADROOM_WARN/FAIL) of
in-request silence with nothing outstanding; it stays a failure under
EVAL_POLICY v1.
* E4: fingerprint under a read-only .git via a private object dir owned by review-log
Codex's sandbox mounts .git read-only, so gstack-wtree's git add -A could not
write blobs and /review lost its start capture. gstack-review-log now creates
a private 0700 object directory under TMPDIR, passes it to gstack-wtree (which
writes new objects there with the common-dir object store, resolved before
exporting, as an alternate, so linked worktrees work), exposes it to the
bindReview/checkSharedLibsReuse consumers that read the tree and blobs, and
removes it on EXIT/INT/TERM after they finish. gstack-wtree prints one stderr
line and exits 1 on a failed write or unwritable TMPDIR instead of printing a
partial hash; review-log no longer swallows that line. Standalone gstack-wtree
keeps writing to the repository store so git diff between saved snapshots
(/ship) still works.
* 0.3: red lines name their values, evidence and an after-a-repair command; one published-text sanitizer
Each red line in the census report keeps its stored error and adds the
machine cause with its evidence, the failure_detail values (Expected/Received,
or each failing judge dimension with its mean, sample count and the sample
sentence naming it), the slice artifact and shard directory that hold the
transcript, a fetch command (eval:pass-rates --run <id> --case <id>), and
'after a repair:' with a command that selects only that case (E2E ids via
--case, judges via their selection and an exact test-name filter). A
session that completed and failed a check says so instead of 'assertion
(success)'. Report text that reaches a PR comment, issue or committed file
goes through scripts/lib/published-text.ts: control and bidi characters
stripped, credential-scan patterns redacted, mentions neutralized, HTML and
link brackets escaped inline, blocks in a fence longer than any backtick run,
length capped. Verdicts and collector verdict objects are unchanged.
* E2: one make-pdf resolver for readiness and every $P block on env-var hosts
make-pdf exported $P from its readiness block, which Codex and the other
env-var hosts lose between blocks. binaryAssignment gains 'make-pdf': the
MAKE-PDF SETUP readiness check and the runtime prelude both use it, so every
block that runs $P derives the same make-pdf/dist/pdf binary, with
MAKE_PDF_BIN (dev builds) winning when it is executable. Claude output keeps
its repo-local-first probe; only the false 'subsequent blocks' comment
changed. Root plus P=... fits the 400-byte prelude budget; the fixed
MAKE_PDF_OVERRIDE segment is the one exemption, named in runtime-root.ts and
the conformance test.
* 0.1: eval:pass-rates --headroom, --reds and --run report views; 85% headroom alarm in --gate
A new pure module, scripts/lib/eval-history.ts, reports per case the slowest
armed session against the timeout it armed (censored when it timed out,
insufficient below 3 samples, unknown without a session ledger, with the
case wall shown separately as an upper bound), the census critical path from
the slice jobs' timings, the verdict red ledger per census by failure class
and machine cause with shrunk per-case rates, and the all-green probability
(an approximation: the product of 1 - p_i, with an interval). --run <id>
lists one census's reds with their values and history and downloads only the
slice artifacts those reds name, reporting any over 64 MB instead of skipping
them. eval-flake-rank.ts only routes flags: --help works offline, an unknown
flag or case id exits 2 before any fetch, --case narrows the JSON and the
alarms, every view prints its run/branch scope, and --gate also alarms on a
session above HEADROOM_FAIL (85%) of its budget.
* E5: GSTACK_DESIGN_IMAGE_MODEL override; the live model check always tests the defaults
The image_generation tool model gains an env override validated against a
conservative gpt-image-<version> pattern; a bad value fails before any
request and the error names the variable. Pairing and rejection hints name
the tool model actually requested. design/scripts/live-model-check.ts now
ignores GSTACK_DESIGN_MODEL and GSTACK_DESIGN_IMAGE_MODEL (and says so), so
an override cannot mask a retirement of a default, and its exit 2 (no key)
says nothing was checked.
* C3: /deslop-shared-libs names a concrete risk for each recommendation, never 'none'
Both shared-libs-opportunity-judgment reds (37193478719 t3, 37195203538 t1)
failed adoption_and_risk because the top recommendation said 'Main risk:
none found inside the repo' / 'none in behavior'. The skill asked to 'name
the main risk or uncertainty' without saying what counts.
* B4: --case selects loop-registered cases; an unknown id exits before any process
caseSelection() resolves the one paid file that registers a case and how a
trial selects it: by its literal Bun test name, by the id a registration loop
names its test at runtime (coverage audits, TPA, journeys), or by running the
whole file when it registers no other case (plan-mode-no-op, whose test names
are not the id, so the old name pattern ran nothing). A whole-file case runs
one file shard per trial and is judged by shard status, as CI judges it.
--list prints the selection. Every E2E_TIERS id now resolves except
carve-section-loading, which fifteen files register (CI runs each as its own
shard) and which still fails with the owning files named.
* 0.3: after-a-repair runs the shard file when --case cannot resolve a case
A case that several files register (carve-section-loading) cannot be selected
by --case; its red line now prints the bun test command for the shard file
that went red, so every printed command selects exactly one case.
* C1: pre-register workflow-judge control bundles for qa, qa-only, review and setup-deploy
Positive (production bundle), negative (headings only) and one targeted
execution-blocking defect per judge, one 3-sample panel each, built with the
eval's own reader, prompt builder, options and floors.
* F: cost_known: a trial whose harness captured no billing says so; totals never hide it
cost_usd stays a finite number. Trial and eval records gain an optional
cost_known boolean (absent means known, so old records still parse; the
validator accepts only a boolean). Session-ledger rows record whether the
session's terminal result carried total_cost_usd; PTY and Codex sessions
never do, and SDK captures without a terminal result do not. A trial is
cost-known only when it has eval records and every record and session
captured billing, so JUnit-only PTY cases (125 of 298 trials in run
37198445662 recorded 0) now read cost unknown. The census headline prints the
known sum plus the unknown count, never a bare total over unknown costs, and
no total when every cost is unknown; the history import keeps the flag.
PTY transcripts are not priced: test/helpers/pricing.ts estimates exclude
cache writes, so they cannot give the billed cost.
* B1: replay corpora for the five repeat-offender cases
shared-libs-plan-callers, plan-ceo-section-loading, auto-decide-preserved,
plan-design-review-plan-mode and ship-docsync-late-result each get an
inventory of every wave census in scope (all artifacts still available)
and minimal, redacted detector or actor inputs replayed through the real
detector in the free suite: census passes, every distinct census red, and
derived negative controls labelled as derived. Fixed false reds stay as
expected passes with their history; the two genuine auto-decide product
failures (37176837432, 37179171083) stay refused. A contract test bounds
each entry at 64 KB, scans it for HIGH credentials, requires a pass and a
refusal per case, and reports entries whose series identity is stale.
Building the auto-decide corpus exposed a latent native detector misread:
a bold 'Mode application and rationale.' heading was parsed as a
withdrawn Mode field (37186854666). A field status now ends at its
sentence.
* B2: read approved stale-fill decisions from more valid ledger records
The structured ledger reader is authoritative before the prose fallback,
but it missed two valid approved records in passing censuses
(37186854666: a camelCase writeProfile subject; 37195203538: a
slash-joined 'D1 / STALE-FILL' row id), which then rested on the prose
detector that caused both historical reds. Both forms are accepted; an
unresolved row, an unrelated decision, the copied fixture plan and
non-write camelCase words stay refused.
* C1: record workflow-judge control results (negatives separated; qa and review miss a removed step)
* C2: record every low workflow-judge sample in the wave censuses with its disposition
* 0.3: published lines cannot break a fixed fence
evals.yml wraps collector-outcomes failures in a fixed three-backtick fence
for the PR comment, so those lines (and eval:pass-rates output meant for the
tracking issue) go through sanitizeFixedFenceLine: cleaned like every
published line, with backticks turned into quotes. Inline published text
also neutralizes backticks.
* A1: fit the post-child record wording under ship's 1.428 parity size cap
* E4: inline the Step 1 base refresh to fit review's 1.199 parity size cap
* B2: inventory of gate and periodic phrase-only detector checks
docs/evals/detector-inventory.md lists every pass/fail decision in gate
and periodic E2E cases that rests on phrases in model free text, with
its structured source (if any), whether the contract is user-visible,
and a disposition: already structured-first, convert candidate, phrase
witness required, or no structured source. Convert candidates are
ordered by census red history.
* 0.2: failureCauseOf takes the stall window instead of importing it; the PTY ledger never throws
eval-store.ts imported STALL_WINDOW_MS from eval-budgets.ts, which free tests
replace with mock.module stubs (review-finalization-budget, office-hours,
qa-fix-loop, qa-only), so loading eval-store failed there. failureCauseOf is
now dependency-free: the record builders pass STALL_WINDOW_MS. A PTY session
whose handle cannot show its text (test doubles) records no panel instead of
throwing from runPtySession's cleanup. The 37186854666 replay also runs
through classifyTrialShard.
* A2/auto-decide: the CEO handoff chat repeats the helper's printed line verbatim
Claude Code 2.1.284 collapses the helper's Bash output ('Ran 1 shell
command'), so the 'Auto-decided review mode → …' line reaches the user only
through chat. In lane D's auto-decide-preserved corpus every passing capture's
chat starts with that line, and the reds paraphrased it ('HOLD SCOPE was
auto-selected…', 'Mode decision done…', "I'm defaulting to HOLD SCOPE…").
The template now says verbatim and why.
* C2: record the judge panels rerun on the final ship and review wording
* integrate: touchfiles and module-size ratchet entries for the wave's new modules
GLOBAL_TOUCHFILES gains session-ledger.ts, published-text.ts and
auq-substance-panel.ts; bin/gstack-bun-version.sh joins every entry that
lists setup or the updater; auq-matrix stays a rule case with its C4
evidence beside E2E_KINDS; eval-history, published-text and session-ledger
are registered new modules.
* E1: BROWSE SETUP install hint pins the tested Bun (1.4.0)
* E2: quote "$P" in make-pdf commands so install paths with spaces work
* E1: /gstack-upgrade checks the incoming Bun floor before pulling
gstack_bun_incoming_hold (bin/gstack-bun-version.sh) is now the one
pre-advance check: the auto-updater calls it before its fast-forward and
/gstack-upgrade's git path calls it after fetching. Below the incoming
release's floor the upgrade stops with BUN_TOO_OLD and the checkout,
VERSION and installs stay untouched. Vendored installs were already safe:
setup refuses before writing and the swap restores the previous install.
* 0.3: only main posts or closes tracking issues; reports carry alarms, red ledger, headroom and all-green
A branch-dispatched periodic or marathon run writes its report to the step
summary and a report artifact; the upsert and close steps run only on
refs/heads/main. The periodic body now keeps every pass-rate alarm line (the
old grep kept only the header), adds the red ledger and session headroom read
with an explicit --branch, the all-green estimate and the triage guide link,
and routes report text through scripts/lib/published-text.ts. The marathon
lane gains the matching close-on-green step.
Migrated: evals-workflow-wiring pinned report-summary.md in the upsert step;
the body is now written by 'Write the census report'.
* C3: prosons neutral-posture detector ignores a negated "taste call"
Both plan-review-prosons-neutral-neg reds (37179171083 t3, 37193478719 t3)
recommended A and closed with "a coverage call, not a taste call"; the
old /taste call/i matched the negation. The detector moves to a helper so
the two census captures replay free as negatives, with positive controls.
* B2: credit the CEO mode handoff only when the chat shows its line
Claude Code 2.1.284 collapses the gstack-ceo-mode-handoff Bash result, so
the printed AUTO_DECIDE line is user-visible only when a later chat message
of the session begins with it. The native detector now requires that
witness; the helper's tool result locates it but no longer replaces it, and
a generic mode statement cannot stand in once a handoff ran. Census
37182865432 is relabelled a genuine miss (expected: refused): the chat said
"I'm defaulting to HOLD SCOPE…" and "Mode decided: HOLD SCOPE". Corpus
replay: 5 credited, 3 refused. ceo-mode-handoff tests migrated to the
witness rule (deliberate behavior change).
* test: baseline the wave's approved detector and harness replay tests
B1 corpora, B3, C3, C4 and D2 replay captures through paid-case detectors,
actors or fixtures (test code by design); each entry carries its reason.
* E6: live Codex multi-block periodic case in the host-run Codex job
test/codex-e2e-multiblock-live.test.ts runs ./setup --host codex once from the
job's checkout into a fresh HOME and CODEX_HOME (auth.json copied from the CI
login home), asserts codex debug prompt-input lists the gstack router as a
real file, seeds one learning, and runs /learn's Show recent block through
codex exec: that later block must root itself through its prelude, run the
installed gstack-learnings-search and print the seeded learning; no block may
report a missing install or an unresolved gstack helper path.
The file joins CODEX_CI_FILES so the planner gives it to the host-run Codex
job. paid-test-durations.json carries a 180 s packing estimate (refreshed by
--write-durations after the first census) so the plan keeps every slice
within max-parallel. Oracles live in test/helpers/codex-multiblock.ts with
free replays in test/codex-multiblock.test.ts. Registration (E2E_TOUCHFILES,
E2E_TIERS periodic, E2E_KINDS rule) is the integrator's.
* E5: design live-model check as a periodic case
test/skill-e2e-design-model-smoke.test.ts runs design/scripts/live-model-check.ts
(default models only) under a temp HOME. Without OPENAI_API_KEY the case is a
Bun skip that the census lists as SKIPPED with no coverage credit; with a key,
exit 0 passes and any other exit fails with the script output (exit 2 under a
present key fails as 'nothing was checked'). paid-test-durations.json carries
a 90 s packing estimate. Registration (E2E_TOUCHFILES, E2E_TIERS periodic,
E2E_KINDS rule) is the integrator's.
* A1: docsync parent points children at a saved interface file
The /ship docsync fixture saved its observation interface only in the
parent prompt and told the parent to copy it into every child prompt,
about 30-40 s of parent generation per session (lane S, 37198445662).
docsSessionOptions now writes it to <home>/fixture-interface.md (0600) and
the parent tells each child to Read that file first. Fault-actor prompts,
which prepare already writes to a file, are unchanged. Fixture-only.
* A3: a JUnit case's recorded wall is capped by its shard's wall
junitCensus summed a case's JUnit test times as duration_ms; for
plan-mode-no-op's concurrent PTY tests that is 346 s inside a 176 s shard
(the "432-613 s" history). The case wall headroom shows is now
min(sum, shard elapsed), a true upper bound. Session headroom already
used the ledger's per-session clocks.
* A3: refresh paid-test durations from census 37198445662
Recorded 2026-09-29 estimates under-packed gate slice 11 (17 files
estimated 1078 s ran 1353 s; plan-review-report 76 s -> 198 s,
plan-ceo-plan-mode 37 s -> 112 s, safety-codex-consult 83 s -> 150 s).
Refreshed both tiers with --report --write-durations from the latest
census's slice walls. Old plan re-costed with observed walls: worst slice
~677 s (slice 11); refreshed plan: every slice estimated <= 539 s.
* C2: Browser Quick gets a feasible budget; QA loop names its directory and stop path
Judge rationales on the qa-only panels (2 of the last 3) cited the 30 s
Browser Quick budget: homepage + 5 targets, each a checkpoint write and a
guarded command, left ~5 s per probe. Quick is now 180 s (method text: 3
minutes); the qa-only-no-fix paid check expects the same native deadline.
PROBE_DIR points at its source line, a stopped loop goes to the final
report steps (annotations, materialize, Markdown), and /review and /ship
say that after smoke expiry only plan checks rerun in revalidation.
* C2: name the flows judges found implicit in review resolvers
Structured outside review: MISSING COVERAGE asks no fix question and is
never reported clean (was "preserve the existing user decision flow").
/review's HIGH-impact plan gate: spawned or non-interactive runs report
REQUIREMENTS MISSING and continue, writing no TODOs. Value cards: the byte
clamp applies to card lines; JSON records you write keep full values.
* C1: a free test catches a removed workflow step, which the judges missed
C1 controls showed the qa workflow judge passing /qa without Phase 10 and
the review judge passing /review without Step 3. The new test pins every
step/phase heading of qa, qa-only, review, ship, plan-ceo-review,
plan-eng-review and setup-deploy (derived from the templates' step
headings) and names any that goes missing or out of order; new step
headings must be pinned. The judge rubric is unchanged (EVAL_POLICY).
* size: fit the wave's wording under ship, review and qa parity caps
Measured after all lane merges: ship -79, review -210, qa -68 bytes over.
Trimmed the added wording first (post-expiry rule, MISSING COVERAGE,
plan-gate default, value-card JSON, stop line) and folded /review's
redundant INFORMATIONAL sentence and findings bullet. ship 268,036 (room
8), review 130,118 (room 1). qa needed 11 more bytes: cap 1.11 -> 1.1102
with the measured 93,305 / 84,049 = 1.11013 and its reason beside it.
* docs: troubleshooting sections for the wave's new messages
Bun floor (setup, auto-update, /gstack-upgrade), /review stale base and
fingerprint temp-dir errors, the design image model override, the
unrecorded Chromium PID, and a contributor section for census report
lines (pass-rates flags, failure causes, headroom, cost unknown, overlay
v4, stale corpora, auq substance). A free test checks every
troubleshooting.md#anchor printed by setup or bin scripts exists.
* TODOS: reliability follow-ups deferrals; close the Bun floor, make-pdf $P and #2709 reap
* 0.4: 'A census went red' guide; docs for causes, session ledger, headroom and new messages
docs/evals/census-red.md walks one red line to a diagnosis: prerequisites, the
report fields, eval:pass-rates --run/--reds/--headroom with expected output,
what unknown/insufficient/unrecorded data means, free replay, and the single
'after a repair' paid run allowed by TESTING_INTERNALS. Linked from
CONTRIBUTING, TESTING_INTERNALS and both census reports.
TESTING_INTERNALS documents the failure_cause glossary (one row per
FAILURE_CAUSES value, diagnostic only under EVAL_POLICY v1), failure_detail,
sessions[], cost_known, the session ledger, the pass-rates views and the 85%
headroom alarm, main-only tracking issues, overlay contract v4, auq-matrix's
3-sample substance panel and the E5/E6 periodic cases. troubleshooting.md
gains #bun-too-old and an anchor per new user-facing message from lanes P, D
and F. README names GSTACK_DESIGN_IMAGE_MODEL. test/census-red-docs.test.ts
pins the glossary to FAILURE_CAUSES, parses every guide command and checks the
links.
* v1.91.20.0: CHANGELOG and version for the reliability follow-ups
* regen: SKILL.md and sections for all hosts, digest, context-budget fixture
bun run gen:skill-docs --host all; goldens and llms.txt re-rendered
unchanged; bun test/helpers/capture-context-budget.ts.
* integrate: register codex-multiblock-live and design-model-smoke (periodic, rule)
* 0.4: census report and eval:pass-rates views link the census-red guide
* test: cookie approval fixture reverses the Bun install-hint pin (E1)
The historical approved cookie-workflow prompt carried BUN_VERSION=1.3.10;
the regenerated install hint now pins 1.4.0, so the fixture's reversal
list gains that edit like the earlier waves' entries.
* test: release floor counts the two new periodic cases (E5, E6)
* fix: /freeze, /guard, /unfreeze and /investigate work on Codex and other env-var hosts
Templates quoted "$HOME/.claude/skills/gstack/...", which the env-var hosts'
rewrites turned into "$HOME/.agents/skills/gstack/..." on Codex (a path a
global install never has; same class on the other hosts), and setup's
runtime roots had no freeze/bin. Quoted runtime paths now use
{{RUNTIME_ROOT}}: $GSTACK_ROOT on env-var hosts (the fence gets the
runtime prelude), the unchanged $HOME path on Claude. _link_runtime_dists
links freeze/bin and the careful/bin helper it sources into every
env-var host's root. Also fixes the same guess in /benchmark-models,
/retro and /diagram. test/env-host-runtime-assets.test.ts fails on the
previous code: no fence guesses a $HOME/<dir>/skills/gstack path, every
$GSTACK_ROOT path a fence uses is in the host's runtime root, and the
installed Codex /freeze and /unfreeze fences run from a fresh shell.
Template-reading tests now read the Claude render or expect the
placeholder (deliberate change).
* CHANGELOG/TODOS: lane W's workflow, cases and guide; the env-var host freeze fix
* E6: a block that reads a SKILL.md is not a missing install
The first live codex-multiblock-live run passed its real check (the /learn
block found the install and printed the seeded learning) but the oracle
flagged a sed of the learn SKILL.md, whose prelude source quotes the
'gstack: no install found (tried $_r)' text. Only an emitted line that
starts with the message and names a real root counts now; free replay
negative added.
* test: qa-only fixture keeps its caller-set 30 s deadline
The qa-only-no-fix fixture tells the model to start the guard with 30 s
(a caller limit), so its deadline assertion's default stays 30 000 ms; the
product's Browser Quick default (180 s) does not apply to it. Reverts the
helper and its test to their base.
* C2: keep PROBE_DIR 'owned' (qa-functional-prompt pins it); qa measured 93,311
* test: paid file count and periodic floor include the two new periodic cases
* test: ship docs freshness pin checks meaning after A1's wording trim
The snapshot still hashes selected release paths (--select narrows them),
generated outputs and docs; the pin now checks that instead of the
pre-trim sentence.
* E3: reap the recorded Chromium only after the daemon has exited
With the PID now recorded (E3), browse stop reaped Chromium while the
daemon was still shutting down; the daemon read that as a crash, exited 1
and left its state file (chromium-profile-isolation's headless stop test
failed 3/3 locally, passes on base). The CLI waits up to 5 s for the daemon
to exit before reaping a survivor.
* regen: qa and qa-only exploratory sections (PROBE_DIR wording)
* B1: batching counter reads a target declared mid-line; replay corpus
GSTA-23's census 37228573062 failed plan-eng-multi-finding-batching with
reviewCount=0 after 975 s although D1-D10 were asked one per finding,
each with its saved decision-ledger record. The report wrote 'Report
destination requested by the user. Review target (fixed): `PLAN.md` ...'
and the counter read target fields only at a line start, so no brief
inherited the plan identity. A target field may now open any sentence of
its line; a mid-sentence mention or another file still gives no identity.
Floor (3), budget and kind unchanged. Corpus: 9 earlier passing censuses,
the red (now credited at D5), and a derived batching refusal; 975 s noted
in the headroom diagnosis.
* B1: shared-libs corpus labels its git path and covers the supported shape
Every census entry was recorded in CI's eval image (git 2.43), where
gstack-safe-git falls back because --no-lazy-fetch is rejected; each entry
now says so (git_path: fallback). Two synthetic supported-path entries
(HEAD and a history citation in the brief, as git >= 2.44 produces) keep
the 37193478719 pass and the 37195203538 refusal, and a test asserts both
shapes are covered and the actor reads no git-path-specific text.
* v1.91.21.0: move off v1.91.20.0 (claimed by #3031)
* test: settle nav-guard commands before asserting; make two new tests Windows-safe
navigation-guard: Bun's expect(promise).rejects slows each Playwright round-trip
to ~1s while it waits (50 page.evaluate: 43ms plain vs 46.6s inside the matcher),
so the click's 5s timeout expired before the guard saw the link-local request
(CI: 'click: Timeout 5000ms exceeded ... waiting for scheduled navigations').
Settle the command first, then assert on the refusal; a source check keeps it so.
env-host-runtime-assets: PROBE_FILES now matches either path separator.
detector-corpus ship-docsync-late-result: JSON-escape the <ROOT> substitution so a
Windows temp path's backslashes do not break JSON.parse.
* fix: CI run/job ids are not phone numbers; context-recovery fixture gets the full runtime bin
pii.phone.e164 redacted the run id in the published eval report's fetch command
(--run 37235771700), breaking the command it prints. Digit-only spans in a run/job
flag, gh run argument, actions/runs/ or /job/ URL segment, or run_id/job_id key are
now exempt; the same digits elsewhere still report.
context-recovery-artifacts: the fixture's ./bin/ lacked gstack-skill-start and
gstack-context-recovery (the preamble's runtime scripts since v1.71), so the agent
spent its 10 turns hunting for them and hit error_max_turns with no final message
(run 37235771700, 11 turns). Link the whole runtime bin; diagnostic rerun passed
in 6 turns.
* test: replay run 37237194905's late-result capture through the docs authority detector
The only failure was call 20's write to a mistyped shard root (3GanyY for 3GanyU),
which the model disclosed; the detector is correct and the A1 post-snapshot
--out under the fixture home stays permitted. Named red recorded in TODOS.
* test: keep docsync-authority in the Windows profile
The replay test's literal helper path tripped the Windows curator's shebang-spawn
pattern, dropping the whole file from the Windows profile (test-free-shards
invariant). The snapshot --out case is already covered by docsync-command-grammar,
so it is removed here, and the Linux-path replay skips on win32.
* fix: gstack-docs-candidate --help works; docs gate red lines name the denying tool failure
PR run 37239016858 ship-docsync-completion: the parent ran
'gstack-docs-candidate --help' (the helper's usage error points at --help, which
failed). That undeclared command made docsToolFailures non-empty, which denied
the atomic-replacement proof of the child's legitimate native Edit, so the red
line blamed the Edit's temp file. A replay of the capture proves the atomic
replace once --help is declared.
- the helper prints its usage for --help/-h and exits 0
- the fixture interface and command grammar admit the helper's bare --help
- the atomic-proof denial names its check and the tool failures behind it
(deny labels renumbered to their lines)
- replay with negative controls (undeclared command, temp left behind, renamed
elsewhere); diagnostic rerun of ship-docsync-completion passed
* test: retro-base-branch fixture uses main and a shallow report path
PR run 37241493377: the agent wrote retro-output.md one directory above the
requested <tmp>/retro-base/retro-output.md (it dropped the retro-base segment),
so the report check found nothing. The fixture also ran plain git init, which
makes master on CI's git while the prompt says the repo is on main; the agent
logged that mismatch as a learning. Init on main and use the temp dir itself as
the repo. Diagnostic rerun passed (18 turns, 203s).
* test: qa-quick sets a caller probe limit and drains past its session budget
C2 raised Browser Quick's native budget to 180 s. qa-quick is the only gate case
on the native budget, and its wall went 194 s (main periodic, pre-C2) -> 254 s and
274 s (this PR's runs) -> 300 s session timeout (run 37242691913): the agent now
probes every navigation target with a checkpoint, then needs ~80 s to report.
Like qa-only-no-fix, the eval now sets a caller probe limit (30 s), scopes the
fixture, and gives bun's outer timeout the session drain margin instead of racing
the 300 s session. The product's 180 s budget and the session budget are unchanged.
Diagnostic rerun passed in 195 s.
* review/ship: use main's post-expiry smoke sentence; v1.91.24.0
GSTA-23's measurement on the review workflow judge: main's 'Post-expiry smoke
rechecks are not-run.' passed 10/10; the C2 wording passed 6/10 and the
'..., even in 4c.' variant failed in PR run 37248650879. qa-probe-gates pins
main's sentence before step 4c.
* test: record the plan-design corpus test in the test-of-test baseline
#3031's ratchet now counts tests that slice a paid test file's source; the B1
plan-design-review-plan-mode replay corpus reads test/skill-e2e-design.test.ts
like its sibling corpora already listed.
* fix: --case no longer deadlocks importing the running CLI
runCaseDiagnosis loaded runPaidShards with a dynamic import of
scripts/test-paid-shards.ts; when that file is the entry it is still in its
top-level await, so the import never resolved and `--case` hung after its
first line (on main since #3031, found by the per-case measurement runs). The
CLI now passes its runner in; library callers keep the lazy import.
* test: name the in-flight file and the native error when bun dies before its summary (#3037)
* Bun 1.4.2 everywhere CI and setup pin the tested version
Bun 1.4.0 on Windows aborts about 3% of free-suite shards with native
bad-handle errors (main too; #3037); the same head on 1.4.2 had none in 180
shard runs. Every current 1.4.0 pin moves together, as bun-version-drift
requires: setup-bun in every workflow (CSO orchestration included), the eval
image (sha256 from bun-v1.4.2's SHASUMS256.txt, re-hashed locally), GitLab,
the Ubicloud free-suite setup, gstack-bun-version.sh's tested version,
engines.bun and @types/bun, the install hint in generated skills, the Dia and
cookie native-qualification pins, and docs and tests that state the current
pin. Recorded 1.4.0 observations, fixtures, CHANGELOG history and the CSO
images' oven/bun:1.3.10 base stay.
* fix: --case runs a single-case file whole
A session option naming the case (testName: 'office-hours-section-loading')
made caseSelection pick test-name mode, whose filter matched no Bun test, so
every trial would run hollow. A file that registers only this case now runs
whole before any name match is tried.
* fix(office-hours): reviewer returns a hash-bound receipt instead of echoing its verdict
The spec reviewer now Writes its verdict once, seals it with
gstack-office-hours-review seal (validates the saved JSON and prints
OFFICE_HOURS_VERDICT round=N sha256=<hash> path=<verdict>), and returns
only that line. check --receipt recomputes the hash and fails closed on a
missing, malformed, stale-round, wrong-path or mismatched receipt. The
saved file stays the sole findings inventory; validators bind the receipt
to the saved bytes and the judge reads those bytes.
* fix(office-hours): only blocking findings send a design back for another review round
Review findings carry severity blocking|minor (verdict schema v2). PASS when
no blocking findings remain; minor findings are recorded in Reviewer
Concerns and the Spec Review table, never a reason for another round.
CONVERGENCE needs a persisting blocking prior. A persisting or unverified
blocking prior cannot be relabeled minor. The 3-round cap is unchanged.
* fix(office-hours): rounds 2 and 3 are delta re-reviews of the captured design diff
prepare saves each reviewed design as round-N.design.md and gives later
reviewers the exact line diff since the previous round. A new later-round
blocking finding must cite (changed_text) an excerpt of a changed line;
seal, check and the validators verify it against the snapshots and fail
closed when a citation or snapshot is missing. Other new observations are
minor. Severity, anti-downgrade and CONVERGENCE rules are unchanged.
* test(office-hours): preservation judge accepts COMPLETED with recorded minor findings
* fix(office-hours): parent fixes only blocking findings; callers may set --max-rounds
Between rounds the parent edits for blocking findings only; minor findings
stay recorded. gstack-office-hours-review takes --max-rounds 1|2|3 (default
3) on every command; reaching it stops at MAX_ITERATIONS. The marathon
fixture sets 2 rounds so round 1 (full) and round 2 (delta) run live.
* docs(changelog): office-hours review loop changes in v1.91.28.0
* fix(cso): helper JSON keeps run URLs only in the validated workflow slot
The shared redactor now treats GitHub Actions run ids as CI metadata
(c245d1bc2), so main's #3041 invariant that the same run digits stay
redacted outside qualification.workflow held only through the phone
false positive. Make it explicit in the CSO sanitizer.
* test(plan-ceo): split-overflow progress detector accepts the skill's candidate wording
Collection never ended when the model wrote 'E1 Slack:' / 'E1-SLACK:' or
labels like 'Include in this quarter' / 'Defer with flip trigger', so the
session ran the whole review (930-1464 s) and failed on its extra findings
or the case deadline (6 of 21 runs). The candidate check now accepts a
platform-bound E-id and the action is the bucket's leading verb, as the
skill names them (Include/Defer/Cut/Hold). The semantic judge, call floor
and include/defer/cut validation are unchanged. Replays the captured CI
37374526437, t2, t9 and t15 menus with negative controls.
* test(plan-ceo): split-overflow candidate identity is the one E1-E5 token
The skill fixes no header or question wording, and shape-matching kept
missing new forms (header 'D1.1 Slack' with 'SCOPE-E1:', 'D1.1 E1 Slack',
'E1 Slack DM bot ...:', a misspelled 'Mattermst'). Identity is now exactly
one E1-E5 token across the header and the question's first line, with no
other candidate's platform named. Options rules, judge, call floor and
include/defer/cut validation are unchanged. Replays all 31 measured runs;
the quoted-row controls now quote a second ledger id.
* v1.91.30.0: move past #3044's claimed 1.91.29.0
* test: register the split-overflow wording replay as a free fixture
* fix(skills): resolve judge-cited ambiguities in /qa and /plan-eng-review
- tasks JSONL keeps the raw branch as GIT_BRANCH instead of overwriting
gstack-slug's sanitized BRANCH used for artifact names
- the Codex outside-voice block runs for ready or unverified, matching
its routing rule
- plan-eng startup names the preamble's own setup prompts; the
cross-project question waits for Step 0
- /qa 8c: a fix that applies but leaves the finding reproducing gets one
revised fix, else it is reverted; browser verification names its before
screenshot; the exploratory loop drops its duplicate 3-step summary
* fix(plan-eng): keep the startup steps in their pinned order
* fix(plan-eng): explicit spawned sessions follow the spawned rule at the scope gate; one announcement rule
A missing tool still sends the prose menu whatever the session type, but a
session the caller or host explicitly marks spawned selects the menu's
Recommendation without prose or BLOCKED. The pre-load announcement
sentence now states Step 0's single check.
* eval policy v3: judge panels gate on the per-dimension median of 3 samples
A judge dimension now passes when at least 2 of its 3 samples meet the
unchanged threshold (judgePanelMedian); any other sample count or a
non-numeric sample fails closed. Thresholds, sample count, judge model and
prompts are unchanged; panel logs show every sample and the mean, and name
the median as the gate. Applies to the skill-llm-eval panels, the
workflow-judge cache, the auq-matrix substance panel and the calibration
replicas. EVAL_POLICY 2 -> 3 and HARNESS_VERSION 1 -> 2 (verdicts change),
so every case starts a new pass-rate series. Replays the captured panels
from jobs 112032744416 and 112049856586 as replayed evidence.
This commit is contained in:
Garry Tan authored and GitHub committed
1 parent
5885157a90
commit
55f8151fbe
322 files changed
+22550
-1051
No files matched your search
+47
-18
@@ -44,6 +44,8 @@
|
||||
* bun run eval:pass-rates --dir <path> # local eval dirs / downloaded artifacts (repeatable)
|
||||
* bun run eval:pass-rates --backfill # also import legacy slice artifacts, labeled pre-policy
|
||||
* bun run eval:pass-rates --json | --gate
|
||||
* bun run eval:pass-rates --headroom | --reds | --run <id> # report views (scripts/lib/eval-history.ts)
|
||||
* bun run eval:pass-rates --help # every flag, no network
|
||||
*/
|
||||
|
||||
import * as fs from 'node:fs';
|
||||
@@ -55,7 +57,9 @@ import { E2E_KINDS, E2E_TIERS, E2E_TOUCHFILES, GLOBAL_TOUCHFILES, LLM_JUDGE_TOUC
|
||||
import { CASE_QUARANTINE, EVAL_POLICY } from '../test/helpers/periodic-exclude-data';
|
||||
import { CASE_TEST_NAMES } from './test-paid-shards';
|
||||
import { resolveStateRoot } from '../lib/state-root';
|
||||
import { downloadRunArtifacts, isPooledTrialRun, isWeeklyHistoryRun, listWeeklyRuns, parseFlakeLedger, repoSlug, TRIAL_OUTCOMES_MAX_BYTES } from './lib/ci-history';
|
||||
import { GH_JOBS, PASS_RATES_USAGE, criticalPath, formatCriticalPath, formatHeadroom, formatRedLedger, headroom, headroomAlarms, parsePassRatesArgs,
|
||||
redLedger, triageRun, CENSUS_RED_GUIDE, type CriticalPath } from './lib/eval-history';
|
||||
import { downloadRunArtifacts, GH_HISTORY, isPooledTrialRun, isWeeklyHistoryRun, listWeeklyRuns, parseFlakeLedger, repoSlug, TRIAL_OUTCOMES_MAX_BYTES, type WeeklyRun } from './lib/ci-history';
|
||||
|
||||
interface TestSeries {
|
||||
name: string;
|
||||
@@ -237,7 +241,7 @@ export function backfillEvalFiles(files: string[], run?: { run_id: string; sha?:
|
||||
tier: caseTier(id, registry), kind: registry.kinds[id]!, trial: 1, panel: { n: 1, k: 1 }, attempt: 1,
|
||||
outcome, ...(outcome === 'failed' ? { failure_class: failureClassOf(entry) } : {}),
|
||||
exit_reason: entry.exit_reason, error: sanitizeTrialError(entry.error),
|
||||
duration_ms: Math.max(0, entry.duration_ms || 0), cost_usd: Math.max(0, entry.cost_usd || 0),
|
||||
duration_ms: Math.max(0, entry.duration_ms || 0), cost_usd: Math.max(0, entry.cost_usd || 0), ...(entry.cost_known === false ? { cost_known: false } : {}),
|
||||
model: entry.model, cli_version: result.claude_cli_version, policy_version: 0, quarantined: false,
|
||||
execution: entry.execution === 'reused' ? 'reused' : 'executed', source: 'backfill',
|
||||
run_id: run?.run_id ?? `local:${file}`, sha: run?.sha ?? result.git_sha, recorded_at: run?.timestamp ?? result.timestamp,
|
||||
@@ -292,7 +296,7 @@ export function holmRejections(pValues: number[], alpha: number): Set<number> {
|
||||
|
||||
export type PassRateLabel = 'INCONCLUSIVE' | 'BROKEN' | 'FLAKY' | 'FAILING' | 'PASSING';
|
||||
export type AlarmKind = 'drift' | 'rule-as-behavior' | 'regression' | 'quarantine-exit' | 'quarantine-expired'
|
||||
| 'quarantine-cap' | 'quarantine-invalid';
|
||||
| 'quarantine-cap' | 'quarantine-invalid' | 'headroom';
|
||||
|
||||
/** The EVAL_POLICY fields pass-rates reads (structural, so tests can vary them). */
|
||||
export interface PassRatePolicy {
|
||||
@@ -567,24 +571,22 @@ export { downloadRunArtifacts, GH_HISTORY, listWeeklyRuns, TRIAL_OUTCOMES_MAX_BY
|
||||
type HistoryFetcher, type RunArtifact, type WeeklyRun } from './lib/ci-history';
|
||||
|
||||
if (import.meta.main) {
|
||||
const argv = process.argv.slice(2);
|
||||
const flag = (name: string) => { const index = argv.indexOf(name); return index === -1 ? undefined : argv[index + 1]; };
|
||||
const dirs = argv.flatMap((arg, index) => arg === '--dir' && argv[index + 1] ? [argv[index + 1]!] : []);
|
||||
const asJson = argv.includes('--json');
|
||||
const gate = argv.includes('--gate');
|
||||
const backfill = argv.includes('--backfill');
|
||||
const caseFilter = flag('--case');
|
||||
const runsLimit = Number(flag('--runs')) || 10;
|
||||
const sinceDays = Number(flag('--since-days')) || 60;
|
||||
const repo = flag('--repo') ?? repoSlug();
|
||||
const workflow = flag('--workflow') ?? 'evals-periodic.yml';
|
||||
const branch = flag('--branch') ?? 'main';
|
||||
const parsed = parsePassRatesArgs(process.argv.slice(2), id => Object.hasOwn(E2E_TIERS, id) || Object.hasOwn(LLM_JUDGE_TOUCHFILES, id)
|
||||
|| (/^test\/[\w./-]+\.test\.ts$/.test(id) && fs.existsSync(id)));
|
||||
if ('help' in parsed) { console.log(PASS_RATES_USAGE); process.exit(0); }
|
||||
if ('error' in parsed) { console.error(`eval:pass-rates: ${parsed.error}\n${PASS_RATES_USAGE}`); process.exit(2); }
|
||||
const { dirs, json: asJson, gate, backfill, caseFilter, runs: runsLimit, sinceDays, view } = parsed;
|
||||
const repo = parsed.repo ?? repoSlug();
|
||||
const workflow = parsed.workflow ?? 'evals-periodic.yml';
|
||||
const branch = parsed.branch ?? 'main';
|
||||
|
||||
const records: TrialRecord[] = [];
|
||||
const unattributed = new Set<string>();
|
||||
const errors: string[] = [];
|
||||
let historyError: string | null = null;
|
||||
let weeklyRuns: string[] | undefined;
|
||||
let fetched: WeeklyRun[] = [];
|
||||
const cacheDir = path.join(path.resolve(resolveStateRoot()), 'eval-pass-rates-cache', repo.replace('/', '-'));
|
||||
let pooledRunIds: Set<string> | undefined;
|
||||
|
||||
const manualReviews: string[] = [];
|
||||
@@ -611,7 +613,8 @@ if (import.meta.main) {
|
||||
: [];
|
||||
pooledRunIds = new Set(pooledRuns.map(run => `${run.id}`));
|
||||
const runs = [...weekly, ...pooledRuns];
|
||||
const cacheDir = path.join(path.resolve(resolveStateRoot()), 'eval-pass-rates-cache', repo.replace('/', '-'));
|
||||
if (parsed.runId !== undefined && !runs.some(run => run.id === parsed.runId)) runs.push(GH_JOBS.getRun(repo, parsed.runId));
|
||||
fetched = runs;
|
||||
const match = backfill
|
||||
? (name: string) => name.startsWith('trial-outcomes') || /^(paid-slice-\d+|gate-census-\d+)(-a\d+)?$/.test(name)
|
||||
: (name: string) => name.startsWith('trial-outcomes');
|
||||
@@ -624,11 +627,37 @@ if (import.meta.main) {
|
||||
}
|
||||
}
|
||||
|
||||
const report = analyzePassRates(records, { weeklyRuns, pooledRunIds, unattributed: [...unattributed].sort(), errors, manualReviews });
|
||||
const scope = dirs.length ? `local dirs ${dirs.join(', ')}` : `${repo} ${workflow} on ${branch}${pooledRunIds?.size ? ` + ${pooledRunIds.size} pooled branch census run(s)` : ''}, `
|
||||
+ `last ${runsLimit} completed run(s): ${fetched.map(run => run.id).join(', ') || 'none'}`;
|
||||
const scoped = caseFilter ? records.filter(record => record.case === caseFilter) : records;
|
||||
const caseHeadroom = headroom(scoped);
|
||||
if (view !== 'rates') {
|
||||
const critical: CriticalPath[] = [];
|
||||
if (view !== 'reds' && !dirs.length) {
|
||||
for (const run of fetched.filter(run => view === 'headroom' || run.id === parsed.runId)) {
|
||||
try { critical.push(criticalPath(String(run.id), GH_JOBS.listJobs(repo, run.id))); } catch { critical.push({ run: String(run.id), jobs: 0, slowest: null, wallMs: null }); }
|
||||
}
|
||||
}
|
||||
const lines = view === 'headroom' ? formatHeadroom(caseHeadroom, critical) : view === 'reds' ? formatRedLedger(redLedger(scoped))
|
||||
: historyError ? [] : triageRun({ repo, run: fetched.find(run => run.id === parsed.runId)!, records, caseFilter, cacheDir, fetcher: GH_HISTORY,
|
||||
download: (match, maxBytes) => downloadRunArtifacts({ repo, run: fetched.find(run => run.id === parsed.runId)!, match, cacheDir, maxBytes }) })
|
||||
.concat(formatCriticalPath(critical), historyError ? [] : [`guide: ${CENSUS_RED_GUIDE}`]);
|
||||
if (asJson) console.log(JSON.stringify({ scope, historyError, view, lines }, null, 2));
|
||||
else {
|
||||
console.log(`scope: ${scope}`);
|
||||
if (historyError) console.log(`history unavailable (${historyError})`);
|
||||
for (const line of lines) console.log(line);
|
||||
}
|
||||
process.exit(historyError ? 1 : 0);
|
||||
}
|
||||
const full = analyzePassRates(records, { weeklyRuns, pooledRunIds, unattributed: [...unattributed].sort(), errors, manualReviews });
|
||||
const alarms = [...full.alarms, ...(gate ? headroomAlarms(caseHeadroom) : [])].filter(alarm => !caseFilter || alarm.case === caseFilter);
|
||||
const report = { ...full, cases: full.cases.filter(entry => !caseFilter || entry.case === caseFilter), alarms };
|
||||
const ledger = readFreeLedger();
|
||||
if (asJson) {
|
||||
console.log(JSON.stringify({ repo, workflow, branch, dirs, historyError, ...report, freeLedger: ledger }, null, 2));
|
||||
console.log(JSON.stringify({ repo, workflow, branch, scope, dirs, historyError, ...report, freeLedger: ledger }, null, 2));
|
||||
} else {
|
||||
console.log(`scope: ${scope}`);
|
||||
if (historyError) console.log(`pass-rates: history unavailable (${historyError}); every label below is INCONCLUSIVE`);
|
||||
console.log(formatPassRates(report, { caseFilter }));
|
||||
if (ledger.length > 0) {
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
{
|
||||
"version": 1,
|
||||
"version": 2,
|
||||
"patterns": [
|
||||
"test/helpers/agent-sdk-runner.ts",
|
||||
"test/helpers/codex-session-runner.ts",
|
||||
@@ -10,41 +10,44 @@
|
||||
"test/helpers/llm-judge.ts"
|
||||
],
|
||||
"files": {
|
||||
"scripts/lib/paid-cases.ts": {"blob":"cfb25a673c8f9c95aacfdbcd91b5dc471488da72","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/lib/paid-plan.ts": {"blob":"45721ce128a112666a0caed6c13c66637b74ca54","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/lib/paid-report.ts": {"blob":"72e2a3647f7e56f77ed27ced79ca9c4d239bb62c","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/lib/shard-engine.ts": {"blob":"3cd0b1aaa84018b00bf9ea0ce48093f3c8492f66","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/test-paid-shards.ts": {"blob":"93717f22a85d842152f49c4275367a1d943cf13e","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/test-pr-profile.ts": {"blob":"d7860897397de8fd605623570d1e7ebbbd5ec743","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"scripts/test-strict-output.ts": {"blob":"01185d20dc9761b04b9b2f751ec98664fe0eaf93","since":1},
|
||||
"test/helpers/agent-sdk-runner.ts": {"blob":"7edff19228222e3582f3efc0dbe12c35649e16b7","since":1},
|
||||
"test/helpers/claude-pty-runner.ts": {"blob":"20cc5894532873aa6145fc6617fdeec8cef94a2d","since":1},
|
||||
"test/helpers/codex-session-runner.ts": {"blob":"4eea35931a35a0afb006363402e3d4191e379737","since":1},
|
||||
"test/helpers/e2e-gate.ts": {"blob":"efb7e197dd5f66c3ad0e34074b14e640754f7a45","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"test/helpers/e2e-helpers.ts": {"blob":"2d4a2a53063cfc1f003db366d2c861ab1e379e6e","since":1},
|
||||
"test/helpers/eval-budgets.ts": {"blob":"824be22c1ddad036bdac9ad89367eb061c07475f","since":1},
|
||||
"test/helpers/eval-store.ts": {"blob":"23ec90ffdc235ecff27e734e4652e9ddda649cb1","since":1},
|
||||
"test/helpers/hermetic-env.ts": {"blob":"6473f78285384f4420f48cbe5928658919078e0b","since":1},
|
||||
"test/helpers/llm-judge.ts": {"blob":"453a4c1d355f3811deca53f953767109a9ec9292","since":1},
|
||||
"test/helpers/paid-test-set.ts": {"blob":"ff9bfe43c43ec2bdd11114f7b8ec7c5736116f75","since":1,"nonBehavioral":"audit-wave integration before the first v2 census (EVAL_POLICY v2 already starts fresh series)"},
|
||||
"test/helpers/pty-screen.ts": {"blob":"4f77569f9ce11f9cee9f347216e468cd7c956820","since":1},
|
||||
"test/helpers/pty/auq.ts": {"blob":"16e5ea2bb5b9f28f44808a1a09024dbd3e6f28c4","since":1},
|
||||
"test/helpers/pty/binary.ts": {"blob":"404d7515076d5c00804f89e23ea69296312c00e1","since":1},
|
||||
"test/helpers/pty/boundaries.ts": {"blob":"53eb76e57c0558c00254366d6334a1c833126575","since":1},
|
||||
"test/helpers/pty/classify.ts": {"blob":"3f2daa12ff6484f96b8d28d910a50231c4bb3ae8","since":1},
|
||||
"test/helpers/pty/fake-session.ts": {"blob":"554e064ef259b8444280eddfe9b0db93c66a8202","since":1},
|
||||
"test/helpers/pty/judge.ts": {"blob":"81c5bb4c48ab928aca2f186a1ade9df7169d7fbe","since":1},
|
||||
"test/helpers/pty/launch.ts": {"blob":"86136d044f496c11aaaaeeee9572b5831bd13208","since":1},
|
||||
"test/helpers/pty/plan-native.ts": {"blob":"ddba9cb68c7c0da935cff7c1cea258dee2dff740","since":1},
|
||||
"test/helpers/pty/runners/counting.ts": {"blob":"a4ac5961c136e3c18540d5b4075ef44c2b87a425","since":1},
|
||||
"test/helpers/pty/runners/floor.ts": {"blob":"a8d6d180280da349a9609af9141ccfd4205313d6","since":1},
|
||||
"test/helpers/pty/runners/observation.ts": {"blob":"032298c30ebeff0a9314749670895406169bde1d","since":1},
|
||||
"test/helpers/pty/screen.ts": {"blob":"ba86fa311c94401ecb7c1376a7a2bc7c1c516cf6","since":1},
|
||||
"test/helpers/pty/session.ts": {"blob":"68f56d907edc6ceaa731f06dbd99c99ee77bcfe6","since":1},
|
||||
"test/helpers/session-drain-policy.ts": {"blob":"6211239dd0485e6495a700b824a7e0e61d1bcd97","since":1},
|
||||
"test/helpers/session-runner.ts": {"blob":"702444c3b3f7e129d101ea3864752e09951adb77","since":1},
|
||||
"test/helpers/skill-fixture.ts": {"blob":"b188064594eb4201a54a80ba1bc6f28e31d7f49e","since":1},
|
||||
"test/helpers/test-selection.ts": {"blob":"51655f89f66be7a5ec487e2abfe4744300ebd5f6","since":1},
|
||||
"test/helpers/touchfiles.ts": {"blob":"15f369056d41b9dc42dcd86f81f60c878074704c","since":1}
|
||||
"scripts/lib/paid-cases.ts": {"blob":"9bacd52446192a7ecf3dfaae901ea443cf9c9ae2","since":2},
|
||||
"scripts/lib/paid-plan.ts": {"blob":"45721ce128a112666a0caed6c13c66637b74ca54","since":2},
|
||||
"scripts/lib/paid-report.ts": {"blob":"3fda96a8b1f7f3d55797934847353f254cc4d054","since":2},
|
||||
"scripts/lib/published-text.ts": {"blob":"3f2e4d616b5dd0204035710c66b64d49055c5714","since":2},
|
||||
"scripts/lib/shard-engine.ts": {"blob":"3cd0b1aaa84018b00bf9ea0ce48093f3c8492f66","since":2},
|
||||
"scripts/test-paid-shards.ts": {"blob":"8ddee5f11ed7856876135520dd8025ab989e2db9","since":2},
|
||||
"scripts/test-pr-profile.ts": {"blob":"d7860897397de8fd605623570d1e7ebbbd5ec743","since":2},
|
||||
"scripts/test-strict-output.ts": {"blob":"01185d20dc9761b04b9b2f751ec98664fe0eaf93","since":2},
|
||||
"test/helpers/agent-sdk-runner.ts": {"blob":"210bba92d241a90df578b2c7fc943ac5755d0025","since":2},
|
||||
"test/helpers/auq-substance-panel.ts": {"blob":"4881260ef793668e4963a062db6d06a5e1f7a40f","since":2},
|
||||
"test/helpers/claude-pty-runner.ts": {"blob":"20cc5894532873aa6145fc6617fdeec8cef94a2d","since":2},
|
||||
"test/helpers/codex-session-runner.ts": {"blob":"4eea35931a35a0afb006363402e3d4191e379737","since":2},
|
||||
"test/helpers/e2e-gate.ts": {"blob":"efb7e197dd5f66c3ad0e34074b14e640754f7a45","since":2},
|
||||
"test/helpers/e2e-helpers.ts": {"blob":"2d4a2a53063cfc1f003db366d2c861ab1e379e6e","since":2},
|
||||
"test/helpers/eval-budgets.ts": {"blob":"667d0e09b20fe5d7366f57b77b5fb9b04baa37ef","since":2},
|
||||
"test/helpers/eval-store.ts": {"blob":"609c6d099108ff2696eb014b47aa318a618dd2af","since":2},
|
||||
"test/helpers/hermetic-env.ts": {"blob":"6473f78285384f4420f48cbe5928658919078e0b","since":2},
|
||||
"test/helpers/llm-judge.ts": {"blob":"54e3306b494fb5d7f132ccdaaae5c680465028f9","since":2},
|
||||
"test/helpers/paid-test-set.ts": {"blob":"ff9bfe43c43ec2bdd11114f7b8ec7c5736116f75","since":2},
|
||||
"test/helpers/pty-screen.ts": {"blob":"4f77569f9ce11f9cee9f347216e468cd7c956820","since":2},
|
||||
"test/helpers/pty/auq.ts": {"blob":"16e5ea2bb5b9f28f44808a1a09024dbd3e6f28c4","since":2},
|
||||
"test/helpers/pty/binary.ts": {"blob":"404d7515076d5c00804f89e23ea69296312c00e1","since":2},
|
||||
"test/helpers/pty/boundaries.ts": {"blob":"53eb76e57c0558c00254366d6334a1c833126575","since":2},
|
||||
"test/helpers/pty/classify.ts": {"blob":"3f2daa12ff6484f96b8d28d910a50231c4bb3ae8","since":2},
|
||||
"test/helpers/pty/fake-session.ts": {"blob":"554e064ef259b8444280eddfe9b0db93c66a8202","since":2},
|
||||
"test/helpers/pty/judge.ts": {"blob":"81c5bb4c48ab928aca2f186a1ade9df7169d7fbe","since":2},
|
||||
"test/helpers/pty/launch.ts": {"blob":"4ef0bcb7b5e7c35d2aac0cbffd7f8a303440b6fe","since":2},
|
||||
"test/helpers/pty/plan-native.ts": {"blob":"ddba9cb68c7c0da935cff7c1cea258dee2dff740","since":2},
|
||||
"test/helpers/pty/runners/counting.ts": {"blob":"8c5c05e54abf961259d34d9c3d77db67138d7794","since":2},
|
||||
"test/helpers/pty/runners/floor.ts": {"blob":"465f01641fb3a18ae1c9d2613a9c2b26f8fb503e","since":2},
|
||||
"test/helpers/pty/runners/observation.ts": {"blob":"1c2b75fbb984708725abaa8939125f7cab969f9f","since":2},
|
||||
"test/helpers/pty/screen.ts": {"blob":"ee005f36ffa48617e6832ccbeb35e8e96873df9b","since":2},
|
||||
"test/helpers/pty/session.ts": {"blob":"58533cea9d81cc87bb3b499bdfe0db8e3a8709c5","since":2},
|
||||
"test/helpers/session-drain-policy.ts": {"blob":"6211239dd0485e6495a700b824a7e0e61d1bcd97","since":2},
|
||||
"test/helpers/session-ledger.ts": {"blob":"6e3c9e837e46f98addba2d12eaebff802da320bb","since":2},
|
||||
"test/helpers/session-runner.ts": {"blob":"d40eb29e8b8fdc0675b511693e4a90823b5963fc","since":2},
|
||||
"test/helpers/skill-fixture.ts": {"blob":"b188064594eb4201a54a80ba1bc6f28e31d7f49e","since":2},
|
||||
"test/helpers/test-selection.ts": {"blob":"51655f89f66be7a5ec487e2abfe4744300ebd5f6","since":2},
|
||||
"test/helpers/touchfiles.ts": {"blob":"15f369056d41b9dc42dcd86f81f60c878074704c","since":2}
|
||||
}
|
||||
}
|
||||
@@ -10,7 +10,7 @@ import {
|
||||
buildCrossSkillConsistencyJudgePrompt, buildVoiceDirectiveJudgePrompt,
|
||||
JUDGE_SCORE_SCHEMA, QA_ANTI_REFUSAL_JUDGE_SCHEMA, CROSS_SKILL_CONSISTENCY_JUDGE_SCHEMA,
|
||||
VOICE_DIRECTIVE_JUDGE_SCHEMA, VOICE_DIRECTIVE_DIMENSIONS, JUDGE_SCORE_DIMENSIONS, DEFAULT_JUDGE_MAX_TOKENS,
|
||||
judgePanelMean, judgePanelMajority,
|
||||
judgePanelMedian, judgePanelMajority,
|
||||
} from '../test/helpers/llm-judge';
|
||||
import { buildWorkflowJudgePrompt, WORKFLOW_JUDGE_RESPONSE_SCHEMA } from '../test/helpers/workflow-judge-input';
|
||||
import { CLAUDE_FRONTIER_EVAL_MODEL } from '../lib/eval-model';
|
||||
@@ -18,7 +18,7 @@ import type { CalibrationConfig, Verdict } from './judge-calibration';
|
||||
|
||||
type Sample = Record<string, unknown>;
|
||||
|
||||
/** What production rejects: judgePanelMean needs finite numbers, judgePanelMajority needs booleans. */
|
||||
/** What production rejects: judgePanelMedian needs finite numbers, judgePanelMajority needs booleans. */
|
||||
function requireFields(value: unknown, numbers: readonly string[], booleans: readonly string[] = []): void {
|
||||
const sample = (value ?? {}) as Sample;
|
||||
for (const key of numbers) {
|
||||
@@ -29,13 +29,13 @@ function requireFields(value: unknown, numbers: readonly string[], booleans: rea
|
||||
}
|
||||
}
|
||||
|
||||
const meets = (mean: Record<string, number>, floors: Record<string, number>): Verdict =>
|
||||
Object.entries(floors).every(([key, floor]) => mean[key]! >= floor) ? 'pass' : 'fail';
|
||||
const meets = (median: Record<string, number>, floors: Record<string, number>): Verdict =>
|
||||
Object.entries(floors).every(([key, floor]) => median[key]! >= floor) ? 'pass' : 'fail';
|
||||
|
||||
const QA_FLOORS = { clarity: 3, completeness: 3, actionability: 4 };
|
||||
const scoreJudge = (floors: Record<string, number>) => ({
|
||||
validate: (value: unknown) => requireFields(value, JUDGE_SCORE_DIMENSIONS),
|
||||
verdict: (samples: unknown[]) => meets(judgePanelMean<string>(samples as Sample[], JUDGE_SCORE_DIMENSIONS), floors),
|
||||
verdict: (samples: unknown[]) => meets(judgePanelMedian<string>(samples as Sample[], JUDGE_SCORE_DIMENSIONS), floors),
|
||||
});
|
||||
|
||||
const workflow = (id: string, priority: number, cases: string[], extra: { agentCapability?: 'frontier'; model?: string; floors?: Record<string, number> } = {}): CalibrationConfig => ({
|
||||
@@ -80,7 +80,7 @@ export const CALIBRATION_CONFIGS: CalibrationConfig[] = [
|
||||
oldOptions: {}, newOptions: { jsonSchema: QA_ANTI_REFUSAL_JUDGE_SCHEMA },
|
||||
validate: value => requireFields(value, ['confidence'], ['would_browse']),
|
||||
verdict: samples => judgePanelMajority<string>(samples as Sample[], 'would_browse')
|
||||
&& judgePanelMean<string>(samples as Sample[], ['confidence']).confidence >= 4 ? 'pass' : 'fail',
|
||||
&& judgePanelMedian<string>(samples as Sample[], ['confidence']).confidence >= 4 ? 'pass' : 'fail',
|
||||
estimatedOutputTokens: 1200,
|
||||
},
|
||||
{
|
||||
@@ -89,7 +89,7 @@ export const CALIBRATION_CONFIGS: CalibrationConfig[] = [
|
||||
oldOptions: {}, newOptions: { jsonSchema: CROSS_SKILL_CONSISTENCY_JUDGE_SCHEMA },
|
||||
validate: value => requireFields(value, ['score'], ['consistent']),
|
||||
verdict: samples => judgePanelMajority<string>(samples as Sample[], 'consistent')
|
||||
&& judgePanelMean<string>(samples as Sample[], ['score']).score >= 4 ? 'pass' : 'fail',
|
||||
&& judgePanelMedian<string>(samples as Sample[], ['score']).score >= 4 ? 'pass' : 'fail',
|
||||
estimatedOutputTokens: 1200,
|
||||
},
|
||||
{
|
||||
@@ -97,7 +97,7 @@ export const CALIBRATION_CONFIGS: CalibrationConfig[] = [
|
||||
build: inputs => buildVoiceDirectiveJudgePrompt(inputs.voiceSection!),
|
||||
oldOptions: {}, newOptions: { jsonSchema: VOICE_DIRECTIVE_JUDGE_SCHEMA },
|
||||
validate: value => requireFields(value, VOICE_DIRECTIVE_DIMENSIONS),
|
||||
verdict: samples => meets(judgePanelMean<string>(samples as Sample[], VOICE_DIRECTIVE_DIMENSIONS),
|
||||
verdict: samples => meets(judgePanelMedian<string>(samples as Sample[], VOICE_DIRECTIVE_DIMENSIONS),
|
||||
Object.fromEntries(VOICE_DIRECTIVE_DIMENSIONS.map(key => [key, 4]))),
|
||||
estimatedOutputTokens: 1200,
|
||||
},
|
||||
|
||||
@@ -0,0 +1,123 @@
|
||||
#!/usr/bin/env bun
|
||||
/**
|
||||
* C1 workflow-judge control bundles (pre-registered in docs/evals/judge-controls-2026-10.md).
|
||||
*
|
||||
* For each judge that produced low samples, builds the production bundle with the
|
||||
* eval's own reader and prompt builder, derives a negative and a targeted-defect
|
||||
* bundle, and runs one 3-sample panel per bundle with the eval's request options.
|
||||
*
|
||||
* Usage:
|
||||
* bun run scripts/judge-controls.ts --list (builds every bundle, no API calls)
|
||||
* bun run scripts/judge-controls.ts --run (36 judge calls; writes docs/evals/judge-controls/results.json)
|
||||
*/
|
||||
import * as fs from 'node:fs';
|
||||
import * as path from 'node:path';
|
||||
import { buildQaWorkflowJudgePrompt, callJudge, judgePanelMean, judgePanelMedian, DEFAULT_JUDGE_MAX_TOKENS, JUDGE_SCORE_DIMENSIONS, type CallJudgeOptions } from '../test/helpers/llm-judge';
|
||||
import { buildWorkflowJudgePrompt, readWorkflowJudgeInput, QA_DISCOVERY_REFERENCES, WORKFLOW_JUDGE_RESPONSE_SCHEMA } from '../test/helpers/workflow-judge-input';
|
||||
|
||||
const ROOT = path.resolve(import.meta.dir, '..');
|
||||
const FLOORS = { clarity: 3, completeness: 3, actionability: 4 } as const;
|
||||
const SAMPLES = 3;
|
||||
|
||||
type Judge = {
|
||||
name: string;
|
||||
read: () => string;
|
||||
prompt: (text: string) => string;
|
||||
options: CallJudgeOptions;
|
||||
defect: (text: string) => string;
|
||||
};
|
||||
|
||||
function cut(text: string, start: string, end: string): string {
|
||||
const from = text.indexOf(start);
|
||||
const to = text.indexOf(end, from + start.length);
|
||||
if (from < 0 || to < 0) throw new Error(`defect anchor missing: ${start} … ${end}`);
|
||||
return text.slice(0, from) + text.slice(to);
|
||||
}
|
||||
|
||||
function insertAfter(text: string, anchor: string, line: string): string {
|
||||
const at = text.indexOf(anchor);
|
||||
if (at < 0) throw new Error(`defect anchor missing: ${anchor}`);
|
||||
return text.slice(0, at + anchor.length) + '\n\n' + line + text.slice(at + anchor.length);
|
||||
}
|
||||
|
||||
const workflow = (skillPath: string, startMarker: string, endMarker: string | null, references: readonly string[] | undefined) =>
|
||||
() => readWorkflowJudgeInput({ root: ROOT, skillPath, startMarker, endMarker, references }).text;
|
||||
|
||||
const JUDGES: Judge[] = [
|
||||
{
|
||||
name: 'qa/SKILL.md workflow',
|
||||
read: workflow('qa/SKILL.md', '# /qa: Test', null,
|
||||
['qa/templates/functional-report-template.md', 'qa/templates/qa-report-template.md', 'qa/references/issue-taxonomy.md']),
|
||||
prompt: buildQaWorkflowJudgePrompt,
|
||||
options: {},
|
||||
defect: text => cut(text, '## Phase 10: Report', '## Phase 11:'),
|
||||
},
|
||||
{
|
||||
name: 'qa-only/SKILL.md workflow',
|
||||
read: workflow('qa-only/SKILL.md', '# /qa-only:', null, QA_DISCOVERY_REFERENCES.filter(file => file !== 'qa/sections/exploratory.md')),
|
||||
prompt: text => buildWorkflowJudgePrompt({ judgeContext: 'a report-only QA testing workflow',
|
||||
judgeGoal: 'how to select browser/native functional/mixed targets, explore safely with repository tools, report exact contract evidence and coverage limits, conditionally load browser/DX instructions and never mutate product/tests/Git through any tool' }, { files: [], text }),
|
||||
options: { max_tokens: DEFAULT_JUDGE_MAX_TOKENS, jsonSchema: WORKFLOW_JUDGE_RESPONSE_SCHEMA },
|
||||
defect: text => insertAfter(text, '# /qa-only: Report-Only QA Testing', 'After reporting each bug, fix it in the source and commit the fix.'),
|
||||
},
|
||||
{
|
||||
name: 'review/SKILL.md workflow',
|
||||
read: workflow('review/SKILL.md', '## Step 0: Detect platform and base branch', null,
|
||||
[...QA_DISCOVERY_REFERENCES, 'review/checklist.md', 'review/specialists/testing.md']),
|
||||
prompt: text => buildWorkflowJudgePrompt({ judgeContext: 'a pre-landing review with bounded exploratory QA',
|
||||
judgeGoal: 'how to review and explore changed behavior even for small diffs without a plan or server, preserve report-only discovery and the test_stub ASK gate, handle incomplete probes honestly, and rerun affected evidence after approved repairs',
|
||||
agentCapability: 'frontier' }, { files: [], text }),
|
||||
options: { max_tokens: DEFAULT_JUDGE_MAX_TOKENS },
|
||||
defect: text => cut(text, '## Step 3: Get the diff', '## Step 3.4'),
|
||||
},
|
||||
{
|
||||
name: 'setup-deploy/SKILL.md platform setup',
|
||||
read: workflow('setup-deploy/SKILL.md', '### Step 2: Detect platform', '## Important Rules', undefined),
|
||||
prompt: text => buildWorkflowJudgePrompt({ judgeContext: 'a deployment configuration setup workflow that detects deploy platforms and writes config to CLAUDE.md',
|
||||
judgeGoal: 'how to detect deploy platforms (Fly.io, Render, Vercel, Netlify, Heroku, GitHub Actions, custom), gather platform-specific configuration (URLs, status commands, health checks, custom hooks), and persist everything to CLAUDE.md for future automated use' }, { files: [], text }),
|
||||
options: { max_tokens: DEFAULT_JUDGE_MAX_TOKENS, jsonSchema: WORKFLOW_JUDGE_RESPONSE_SCHEMA },
|
||||
defect: text => {
|
||||
const step = text.indexOf('### Step 2: Detect platform');
|
||||
const open = text.indexOf('```bash', step);
|
||||
const close = text.indexOf('```', open + 7);
|
||||
if (step < 0 || open < 0 || close < 0) throw new Error('defect anchor missing: Step 2 bash block');
|
||||
return text.slice(0, open) + text.slice(close + 3);
|
||||
},
|
||||
},
|
||||
];
|
||||
|
||||
/** Keep only file boundaries and headings. */
|
||||
const negative = (text: string) => text.split('\n')
|
||||
.filter(line => /^--- (BEGIN|END) FILE /.test(line) || /^#{1,6} /.test(line)).join('\n');
|
||||
|
||||
function bundles(judge: Judge): Record<'positive' | 'negative' | 'defect', string> {
|
||||
const positive = judge.read();
|
||||
return { positive, negative: negative(positive), defect: judge.defect(positive) };
|
||||
}
|
||||
|
||||
const mode = process.argv[2];
|
||||
if (mode !== '--list' && mode !== '--run') {
|
||||
console.error('usage: bun run scripts/judge-controls.ts --list | --run');
|
||||
process.exit(2);
|
||||
}
|
||||
if (mode === '--list') {
|
||||
for (const judge of JUDGES) {
|
||||
const built = bundles(judge);
|
||||
console.log(judge.name, Object.fromEntries(Object.entries(built).map(([k, v]) => [k, v.length])));
|
||||
}
|
||||
process.exit(0);
|
||||
}
|
||||
|
||||
type Sample = { clarity: number; completeness: number; actionability: number; reasoning: string };
|
||||
const results = await Promise.all(JUDGES.flatMap(judge => Object.entries(bundles(judge)).map(async ([bundle, text]) => {
|
||||
const prompt = judge.prompt(text);
|
||||
const samples = await Promise.all(Array.from({ length: SAMPLES }, () => callJudge<Sample>(prompt, undefined, judge.options)));
|
||||
const mean = judgePanelMean(samples, JUDGE_SCORE_DIMENSIONS);
|
||||
const median = judgePanelMedian(samples, JUDGE_SCORE_DIMENSIONS);
|
||||
const verdict = Object.entries(FLOORS).every(([key, floor]) => median[key as keyof typeof FLOORS] >= floor) ? 'pass' : 'fail';
|
||||
return { judge: judge.name, bundle, verdict, median, mean, samples };
|
||||
})));
|
||||
const out = path.join(ROOT, 'docs/evals/judge-controls/results.json');
|
||||
fs.mkdirSync(path.dirname(out), { recursive: true });
|
||||
fs.writeFileSync(out, JSON.stringify({ recorded_at: new Date().toISOString(), results }, null, 2) + '\n');
|
||||
for (const r of results) console.log(`${r.judge} | ${r.bundle} | ${r.verdict} | median ${JSON.stringify(r.median)} mean ${JSON.stringify(r.mean)}`);
|
||||
@@ -0,0 +1,354 @@
|
||||
/**
|
||||
* eval:pass-rates report views over stored trial records (plan 0.1):
|
||||
* session headroom (--headroom), the verdict red ledger with the all-green
|
||||
* probability (--reds), and one census's triage view (--run <id>). Pure
|
||||
* functions plus the run-scoped fetches they need; scripts/eval-flake-rank.ts
|
||||
* parses flags, loads history and dispatches here.
|
||||
*
|
||||
* Headroom is per armed session (session ledger, `sessions[]` on a record):
|
||||
* elapsed over the timeout it armed. A case's headroom is its slowest
|
||||
* session; a timed-out sample is censored (its true duration is unknown).
|
||||
* Records without sessions are unknown, never scored; their case wall
|
||||
* (JUnit or shard clock) is shown separately as an upper bound.
|
||||
*
|
||||
* The all-green probability is an approximation that assumes independent
|
||||
* verdicts: P(green) = Π (1 - p_i) over the cases a census verdicts, where
|
||||
* p_i = (reds_i + K * p̄) / (verdicts_i + K) shrinks each case's red rate
|
||||
* toward the pooled rate p̄. Its interval recomputes the same product with p̄
|
||||
* replaced by the pooled rate's 95% Wilson bounds.
|
||||
*/
|
||||
import * as fs from 'node:fs';
|
||||
import * as path from 'node:path';
|
||||
import { spawnSync } from 'node:child_process';
|
||||
import { panelVerdict, type TrialOutcomeRecord } from '../../test/helpers/eval-store';
|
||||
import { HEADROOM_FAIL, HEADROOM_WARN } from '../../test/helpers/eval-budgets';
|
||||
import { sanitizeFixedFenceLine } from './published-text';
|
||||
import type { HistoryFetcher, WeeklyRun } from './ci-history';
|
||||
import { CASE_SHARDED_FILES, trialShardKey } from './paid-cases';
|
||||
import { shardSlug } from '../test-paid-shards';
|
||||
import { CENSUS_RED_GUIDE } from './paid-report';
|
||||
|
||||
export { CENSUS_RED_GUIDE };
|
||||
|
||||
/** Below this many session samples a case's headroom is insufficient, never a pass. */
|
||||
export const HEADROOM_MIN_SAMPLES = 3;
|
||||
/** Pseudo-verdicts of pooled rate added to each case's red rate. */
|
||||
export const RED_RATE_SHRINK = 4;
|
||||
/** Largest slice artifact --run downloads. */
|
||||
export const EVIDENCE_MAX_BYTES = 64 * 1024 * 1024;
|
||||
|
||||
type Rec = TrialOutcomeRecord;
|
||||
const pct = (x: number) => `${Math.round(x * 1000) / 10}%`;
|
||||
const secs = (ms: number) => `${Math.round(ms / 1000)}s`;
|
||||
|
||||
// --- Headroom ---
|
||||
|
||||
export interface SessionHeadroom { key: string; samples: number; censored: number; maxRatio: number; maxElapsedMs: number; budgetMs: number }
|
||||
export interface CaseHeadroom {
|
||||
case: string;
|
||||
status: 'over' | 'warn' | 'ok' | 'insufficient' | 'unknown';
|
||||
sessions: SessionHeadroom[];
|
||||
/** Slowest session's worst ratio, when any session was recorded. */
|
||||
worst?: SessionHeadroom;
|
||||
/** Largest case wall (JUnit or shard clock) over the window: an upper bound, never a session clock. */
|
||||
caseWallMs: number;
|
||||
runs: number;
|
||||
}
|
||||
|
||||
export function headroom(records: readonly Rec[]): CaseHeadroom[] {
|
||||
const byCase = new Map<string, Rec[]>();
|
||||
for (const r of records) if (r.outcome !== 'skipped' && r.execution !== 'reused') byCase.set(r.case, [...(byCase.get(r.case) ?? []), r]);
|
||||
return [...byCase.entries()].map(([id, rs]) => {
|
||||
const sessions = new Map<string, SessionHeadroom>();
|
||||
for (const s of rs.flatMap(r => r.sessions ?? [])) {
|
||||
if (!s.budget_ms) continue;
|
||||
const key = s.key;
|
||||
const entry = sessions.get(key) ?? { key, samples: 0, censored: 0, maxRatio: 0, maxElapsedMs: 0, budgetMs: s.budget_ms };
|
||||
entry.samples++;
|
||||
if (s.end === 'session_timeout' || s.end === 'observer_timeout') entry.censored++;
|
||||
const ratio = s.elapsed_ms / s.budget_ms;
|
||||
if (ratio > entry.maxRatio) Object.assign(entry, { maxRatio: ratio, maxElapsedMs: s.elapsed_ms, budgetMs: s.budget_ms });
|
||||
sessions.set(key, entry);
|
||||
}
|
||||
const list = [...sessions.values()].sort((a, b) => b.maxRatio - a.maxRatio);
|
||||
const worst = list[0];
|
||||
const status: CaseHeadroom['status'] = !worst ? 'unknown' : worst.samples < HEADROOM_MIN_SAMPLES ? 'insufficient'
|
||||
: worst.maxRatio > HEADROOM_FAIL ? 'over' : worst.maxRatio > HEADROOM_WARN ? 'warn' : 'ok';
|
||||
return { case: id, status, sessions: list, ...(worst ? { worst } : {}), caseWallMs: Math.max(0, ...rs.map(r => r.duration_ms)),
|
||||
runs: new Set(rs.map(r => r.run_id ?? 'local')).size };
|
||||
}).sort((a, b) => (b.worst?.maxRatio ?? -1) - (a.worst?.maxRatio ?? -1) || b.caseWallMs - a.caseWallMs);
|
||||
}
|
||||
|
||||
/** --gate alarms: a case whose slowest session ran above HEADROOM_FAIL of its armed budget. */
|
||||
export function headroomAlarms(cases: readonly CaseHeadroom[]): Array<{ kind: 'headroom'; case: string; message: string }> {
|
||||
return cases.filter(c => c.status === 'over').map(c => ({ kind: 'headroom' as const, case: c.case,
|
||||
message: `${c.case} session ${c.worst!.key}: max ${secs(c.worst!.maxElapsedMs)} of ${secs(c.worst!.budgetMs)} (${pct(c.worst!.maxRatio)})`
|
||||
+ `${c.worst!.censored ? `, ${c.worst!.censored} timed-out sample(s) censored` : ''} over ${c.worst!.samples} sample(s), above the ${pct(HEADROOM_FAIL)} cap. `
|
||||
+ 'Cut work in the skill or fixture; budgets are never raised (TESTING_INTERNALS "Timeout policy").' }));
|
||||
}
|
||||
|
||||
export function formatHeadroom(cases: readonly CaseHeadroom[], critical: readonly CriticalPath[] = []): string[] {
|
||||
const lines = [`headroom: slowest armed session per case (warn > ${pct(HEADROOM_WARN)}, alarm > ${pct(HEADROOM_FAIL)}; `
|
||||
+ `insufficient below ${HEADROOM_MIN_SAMPLES} samples; censored = timed out)`,
|
||||
' status session max/budget samples case wall (upper bound, not a session clock) case'];
|
||||
for (const c of cases) {
|
||||
const w = c.worst;
|
||||
const session = w ? `${pct(w.maxRatio)} ${secs(w.maxElapsedMs)}/${secs(w.budgetMs)}` : 'unknown (no session ledger)';
|
||||
lines.push(` ${c.status.padEnd(12)} ${session.padEnd(26)} ${(w ? `${w.samples}${w.censored ? ` (${w.censored} censored)` : ''}` : '-').padEnd(7)} `
|
||||
+ `${secs(c.caseWallMs).padEnd(45)} ${c.case}${w ? ` [${w.key}]` : ''}`);
|
||||
}
|
||||
return [...lines, ...formatCriticalPath(critical)];
|
||||
}
|
||||
|
||||
export function formatCriticalPath(critical: readonly CriticalPath[]): string[] {
|
||||
return critical.map(cp => `critical path ${cp.run}: ${cp.slowest ? `${cp.slowest.name} ${secs(cp.slowest.ms)}` : 'unknown'}`
|
||||
+ ` of ${cp.jobs} slice job(s), run wall ${cp.wallMs === null ? 'unknown' : secs(cp.wallMs)}${cp.files?.length ? `; slice files: ${cp.files.join(', ')}` : ''}`);
|
||||
}
|
||||
|
||||
// --- Critical path from runner slice timings ---
|
||||
|
||||
export interface JobTiming { name: string; startedAt: string | null; completedAt: string | null }
|
||||
export interface CriticalPath { run: string; jobs: number; slowest: { name: string; ms: number } | null; wallMs: number | null; files?: string[] }
|
||||
|
||||
const SLICE_JOB = /^(eval-slices|eval-codex-slices|gate-census) \((\d+)\)$/;
|
||||
|
||||
export function criticalPath(run: string, jobs: readonly JobTiming[], plan?: { entries: Array<{ file: string; slice?: number; status?: string }> }): CriticalPath {
|
||||
const timed = jobs.filter(j => SLICE_JOB.test(j.name) && j.startedAt && j.completedAt)
|
||||
.map(j => ({ name: j.name, ms: Date.parse(j.completedAt!) - Date.parse(j.startedAt!) }));
|
||||
const slowest = timed.sort((a, b) => b.ms - a.ms)[0] ?? null;
|
||||
const all = jobs.filter(j => j.startedAt && j.completedAt);
|
||||
const wallMs = all.length ? Math.max(...all.map(j => Date.parse(j.completedAt!))) - Math.min(...all.map(j => Date.parse(j.startedAt!))) : null;
|
||||
const slice = slowest ? Number(SLICE_JOB.exec(slowest.name)![2]) : null;
|
||||
const files = plan && slice !== null && slowest!.name.startsWith('gate-census')
|
||||
? plan.entries.filter(e => e.slice === slice && e.status === 'planned').map(e => e.file) : undefined;
|
||||
return { run, jobs: timed.length, slowest, wallMs, ...(files ? { files } : {}) };
|
||||
}
|
||||
|
||||
// --- Red ledger ---
|
||||
|
||||
export interface VerdictRed { run: string; lane: string; case: string; status: string; trials: Rec[] }
|
||||
export interface CensusReds { run: string; lane: string; verdicts: number; reds: VerdictRed[]; byClass: Record<string, number>; byCause: Record<string, number> }
|
||||
export interface RedLedger {
|
||||
censuses: CensusReds[];
|
||||
pooled: { verdicts: number; reds: number; rate: number; lo: number; hi: number };
|
||||
perCase: Array<{ case: string; verdicts: number; reds: number; rate: number }>;
|
||||
allGreen: { p: number; lo: number; hi: number; verdicts: number } | null;
|
||||
}
|
||||
|
||||
function wilson(k: number, n: number, z = 1.96): { lo: number; hi: number } {
|
||||
if (n === 0) return { lo: 0, hi: 1 };
|
||||
const p = k / n, d = 1 + z * z / n, c = p + z * z / (2 * n), m = z * Math.sqrt(p * (1 - p) / n + z * z / (4 * n * n));
|
||||
return { lo: Math.max(0, (c - m) / d), hi: Math.min(1, (c + m) / d) };
|
||||
}
|
||||
|
||||
/** One panelVerdict per case per census lane (earliest attempt), exactly as the report computes it. */
|
||||
export function verdictsOf(records: readonly Rec[]): Array<{ run: string; lane: string; case: string; status: string; failsLane: boolean; trials: Rec[] }> {
|
||||
const groups = new Map<string, Rec[]>();
|
||||
for (const r of records) {
|
||||
if (r.policy_version < 1 || r.execution === 'reused') continue;
|
||||
// A JUnit case is one verdict per shard file, as the report counts it; trial shards form one panel.
|
||||
const key = `${r.run_id ?? 'local'}\0${r.lane ?? r.tier}\0${r.case}\0${r.source === 'shard' ? '' : r.file}`;
|
||||
groups.set(key, [...(groups.get(key) ?? []), r]);
|
||||
}
|
||||
return [...groups.entries()].map(([key, rs]) => {
|
||||
const [run, lane, id] = key.split('\0') as [string, string, string, string];
|
||||
const attempt = Math.min(...rs.map(r => r.attempt));
|
||||
const trials = rs.filter(r => r.attempt === attempt);
|
||||
const v = panelVerdict({ case: id, kind: trials[0]!.kind, panel: trials[0]!.panel, quarantined: trials[0]!.quarantined,
|
||||
trials: trials.map(t => ({ trial: t.trial, outcome: t.outcome, attempt, ...(t.failure_class ? { failure_class: t.failure_class } : {}) })) });
|
||||
return { run, lane, case: id, status: v.status, failsLane: v.failsLane, trials };
|
||||
});
|
||||
}
|
||||
|
||||
export function redLedger(records: readonly Rec[]): RedLedger {
|
||||
const verdicts = verdictsOf(records).filter(v => v.status !== 'SKIPPED');
|
||||
const censusKey = (v: { run: string; lane: string }) => `${v.run}\0${v.lane}`;
|
||||
const censuses = new Map<string, CensusReds>();
|
||||
for (const v of verdicts) {
|
||||
const c = censuses.get(censusKey(v)) ?? { run: v.run, lane: v.lane, verdicts: 0, reds: [], byClass: {}, byCause: {} };
|
||||
c.verdicts++;
|
||||
if (v.failsLane) {
|
||||
c.reds.push({ run: v.run, lane: v.lane, case: v.case, status: v.status, trials: v.trials });
|
||||
for (const t of v.trials.filter(t => t.outcome === 'failed')) {
|
||||
c.byClass[t.failure_class ?? 'assertion'] = (c.byClass[t.failure_class ?? 'assertion'] ?? 0) + 1;
|
||||
c.byCause[t.failure_cause ?? 'unrecorded'] = (c.byCause[t.failure_cause ?? 'unrecorded'] ?? 0) + 1;
|
||||
}
|
||||
}
|
||||
censuses.set(censusKey(v), c);
|
||||
}
|
||||
const reds = verdicts.filter(v => v.failsLane).length;
|
||||
const rate = verdicts.length ? reds / verdicts.length : 0;
|
||||
const pooled = { verdicts: verdicts.length, reds, rate, ...wilson(reds, verdicts.length) };
|
||||
const byCase = new Map<string, { verdicts: number; reds: number }>();
|
||||
for (const v of verdicts) {
|
||||
const e = byCase.get(v.case) ?? { verdicts: 0, reds: 0 };
|
||||
e.verdicts++; if (v.failsLane) e.reds++;
|
||||
byCase.set(v.case, e);
|
||||
}
|
||||
const perCase = [...byCase.entries()].map(([id, e]) => ({ case: id, ...e, rate: (e.reds + RED_RATE_SHRINK * rate) / (e.verdicts + RED_RATE_SHRINK) }))
|
||||
.sort((a, b) => b.rate - a.rate || (a.case < b.case ? -1 : 1));
|
||||
const ordered = [...censuses.values()];
|
||||
const latestRun = ordered.map(c => c.run).sort().at(-1);
|
||||
const census = verdicts.filter(v => v.run === latestRun);
|
||||
const green = (prior: number) => census.reduce((p, v) => {
|
||||
const e = byCase.get(v.case)!;
|
||||
return p * (1 - (e.reds + RED_RATE_SHRINK * prior) / (e.verdicts + RED_RATE_SHRINK));
|
||||
}, 1);
|
||||
const allGreen = census.length ? { p: green(rate), lo: green(pooled.hi), hi: green(pooled.lo), verdicts: census.length } : null;
|
||||
return { censuses: ordered.sort((a, b) => (a.run < b.run ? -1 : a.run > b.run ? 1 : a.lane < b.lane ? -1 : 1)), pooled, perCase, allGreen };
|
||||
}
|
||||
|
||||
const tally = (counts: Record<string, number>) => Object.entries(counts).sort(([a, x], [b, y]) => y - x || (a < b ? -1 : 1)).map(([k, n]) => `${k} ${n}`).join(', ') || 'none';
|
||||
|
||||
export function formatRedLedger(ledger: RedLedger, opts: { top?: number } = {}): string[] {
|
||||
const lines = ['reds: verdict reds per census lane by failure class and machine cause (cause "unrecorded" = written before failure_cause existed)'];
|
||||
for (const c of ledger.censuses) {
|
||||
lines.push(` ${c.run} ${c.lane}: ${c.reds.length} red of ${c.verdicts} verdicts${c.reds.length ? ` — class: ${tally(c.byClass)}; cause: ${tally(c.byCause)}` : ''}`);
|
||||
for (const red of c.reds) {
|
||||
const failed = red.trials.filter(t => t.outcome === 'failed');
|
||||
lines.push(` ✗ ${red.case} ${red.status} ${failed.map(t => `t${t.trial} ${t.failure_class ?? 'assertion'}/${t.failure_cause ?? 'unrecorded'}`
|
||||
+ `${t.error ? ` — ${sanitizeFixedFenceLine(t.error, 160)}` : ''}`).join('; ')}`);
|
||||
}
|
||||
}
|
||||
const p = ledger.pooled;
|
||||
lines.push(`pooled: ${p.reds} red of ${p.verdicts} verdicts = ${pct(p.rate)} per verdict [95% ${pct(p.lo)}–${pct(p.hi)}]`);
|
||||
const top = ledger.perCase.filter(c => c.reds > 0).slice(0, opts.top ?? 15);
|
||||
if (top.length) lines.push(`per-case red rates (shrunk toward the pooled rate, K=${RED_RATE_SHRINK}):`, ...top.map(c => ` ${pct(c.rate).padStart(6)} ${c.reds}/${c.verdicts} ${c.case}`));
|
||||
const g = ledger.allGreen;
|
||||
lines.push(g ? `all-green probability (approximation, assumes independent verdicts): ${pct(g.p)} for the latest census's ${g.verdicts} verdicts [${pct(g.lo)}–${pct(g.hi)} with the pooled rate's 95% interval as prior]; `
|
||||
+ 'formula Π(1 - p_i) over its cases; rerun: bun run eval:pass-rates --reds' : 'all-green probability: no post-policy verdicts');
|
||||
lines.push(`guide: ${CENSUS_RED_GUIDE}`);
|
||||
return lines;
|
||||
}
|
||||
|
||||
// --- One census's triage view ---
|
||||
|
||||
/** Fetches the report views make beyond the history loader: a named run and its slice jobs. */
|
||||
export interface RunJobsFetcher { listJobs(repo: string, runId: number): JobTiming[]; getRun(repo: string, runId: number): WeeklyRun }
|
||||
|
||||
function ghJson(args: string[]): string[] {
|
||||
const result = spawnSync('gh', args, { timeout: 120_000, maxBuffer: 64 * 1024 * 1024 });
|
||||
if (result.status !== 0) throw new Error(`gh ${args.slice(0, 2).join(' ')} failed: ${String(result.stderr || result.error || '').trim()}`);
|
||||
return result.stdout.toString('utf8').split('\n').filter(Boolean);
|
||||
}
|
||||
|
||||
export const GH_JOBS: RunJobsFetcher = {
|
||||
listJobs: (repo, runId) => ghJson(['api', `repos/${repo}/actions/runs/${runId}/jobs?per_page=100`, '--paginate',
|
||||
'--jq', '.jobs[] | {name, startedAt: .started_at, completedAt: .completed_at}']).map(line => JSON.parse(line) as JobTiming),
|
||||
getRun: (repo, runId) => JSON.parse(ghJson(['api', `repos/${repo}/actions/runs/${runId}`,
|
||||
'--jq', '{id, attempt: .run_attempt, sha: .head_sha, branch: .head_branch, createdAt: .created_at}'])[0]!) as WeeklyRun,
|
||||
};
|
||||
|
||||
/** The manifest shard key a record ran under: its trial shard, its case shard, or its file. */
|
||||
export function recordShardKey(r: Pick<Rec, 'file' | 'case' | 'trial' | 'source'>): string {
|
||||
return r.source === 'shard' ? trialShardKey(r.file, r.case, r.trial) : CASE_SHARDED_FILES.includes(r.file) ? `${r.file}#${r.case}` : r.file;
|
||||
}
|
||||
|
||||
/**
|
||||
* The run's reds with their values and history, and the slice artifacts that
|
||||
* hold their transcripts. Only the slices a red names are downloaded; a slice
|
||||
* over EVIDENCE_MAX_BYTES is reported, never silently skipped.
|
||||
*/
|
||||
export function triageRun(opts: { repo: string; run: WeeklyRun; records: readonly Rec[]; caseFilter?: string; cacheDir: string;
|
||||
fetcher: HistoryFetcher; download: (match: (name: string) => boolean, maxBytes: number) => string[] }): string[] {
|
||||
const runId = String(opts.run.id);
|
||||
const mine = opts.records.filter(r => r.run_id === runId);
|
||||
const reds = verdictsOf(mine).filter(v => v.failsLane && (!opts.caseFilter || v.case === opts.caseFilter));
|
||||
const lines = [`run ${runId} (${opts.run.branch} @ ${opts.run.sha.slice(0, 9)}, ${opts.run.createdAt}): ${reds.length} verdict red(s)${opts.caseFilter ? ` for ${opts.caseFilter}` : ''}`];
|
||||
const history = verdictsOf(opts.records);
|
||||
const plans = new Map<string, Map<string, number>>();
|
||||
const planFor = (lane: string) => {
|
||||
const name = lane.startsWith('gate') && lane.includes('census') ? 'gate-census-plan' : 'paid-plan';
|
||||
if (!plans.has(name)) {
|
||||
const files = new Map<string, number>();
|
||||
for (const dir of opts.download(n => n === name, 8 * 1024 * 1024)) {
|
||||
try {
|
||||
const manifest = JSON.parse(fs.readFileSync(path.join(dir, 'manifest.json'), 'utf8'));
|
||||
for (const e of manifest.entries ?? []) if (e.status === 'planned' && Number.isInteger(e.slice)) files.set(e.file, e.slice);
|
||||
} catch { /* no manifest: slices stay unknown */ }
|
||||
}
|
||||
plans.set(name, files);
|
||||
}
|
||||
return { prefix: name === 'gate-census-plan' ? 'gate-census' : 'paid-slice', files: plans.get(name)! };
|
||||
};
|
||||
const artifacts = opts.fetcher.listArtifacts(opts.repo, opts.run.id);
|
||||
for (const red of reds) {
|
||||
const runs = history.filter(v => v.case === red.case);
|
||||
lines.push(`✗ ${red.case} ${red.lane} ${red.status} (history: ${runs.filter(v => !v.failsLane).length}/${runs.length} verdicts green)`);
|
||||
for (const t of red.trials.filter(t => t.outcome === 'failed')) {
|
||||
const d = t.failure_detail;
|
||||
const detail = !d ? '' : 'judge' in d ? ` ${d.judge.map(j => `${j.dimension} ${j.mean} < ${j.threshold} (${j.samples} samples)${j.rationale ? ` "${j.rationale}"` : ''}`).join('; ')}`
|
||||
: ` Expected: ${d.expected} · Received: ${d.received}`;
|
||||
lines.push(sanitizeFixedFenceLine(` t${t.trial}: ${t.failure_class ?? 'assertion'} / cause ${t.failure_cause ?? 'unrecorded'}`
|
||||
+ `${t.failure_cause_evidence ? `: ${t.failure_cause_evidence}` : ''}${t.error ? ` — ${t.error}` : ''}${detail}`, 2000));
|
||||
const plan = planFor(red.lane);
|
||||
const slice = plan.files.get(recordShardKey(t));
|
||||
if (slice === undefined) { lines.push(' evidence: slice unknown (no plan artifact for this lane)'); continue; }
|
||||
const name = `${plan.prefix}-${slice}-a${t.attempt}`;
|
||||
const artifact = artifacts.find(a => a.name === name);
|
||||
if (!artifact) { lines.push(` evidence: ${name} (expired or missing)`); continue; }
|
||||
if (artifact.size > EVIDENCE_MAX_BYTES) { lines.push(` evidence: ${name} too large to fetch (${Math.round(artifact.size / 1024 / 1024)} MB)`); continue; }
|
||||
const [dir] = opts.download(n => n === name, EVIDENCE_MAX_BYTES);
|
||||
lines.push(` evidence: ${dir ? path.join(dir, 'shards', shardSlug([recordShardKey(t)])) : `${name} (download failed)`}`);
|
||||
}
|
||||
}
|
||||
return lines;
|
||||
}
|
||||
|
||||
// --- Flags ---
|
||||
|
||||
export const PASS_RATES_USAGE = `Usage: bun run eval:pass-rates [options]
|
||||
(no view flag) per-case trial pass rates with 95% Wilson intervals
|
||||
--headroom slowest armed session per case vs its budget, case wall upper bound, census critical path
|
||||
--reds verdict reds per census by failure class and cause, pooled and per-case rates, all-green probability
|
||||
--run <id> one census: each red with its values, history and fetched transcript evidence
|
||||
--case <id> restrict every view, the JSON and the alarms to one registered case
|
||||
--runs <n> completed runs per branch to read (default 10)
|
||||
--branch <name> branch whose runs to read beside main (default: current branch)
|
||||
--repo <owner/repo> --workflow <file> --dir <path> (repeatable, local artifacts; no network)
|
||||
--backfill also import legacy slice artifacts, labeled pre-policy
|
||||
--since-days <n> legacy eval files window for --dir (default 60)
|
||||
--json machine-readable output
|
||||
--gate exit 1 on ACTION REQUIRED (pass-rate alarms and sessions above ${pct(HEADROOM_FAIL)} of budget)
|
||||
--help this text (no network)`;
|
||||
|
||||
export interface PassRatesArgs {
|
||||
view: 'rates' | 'headroom' | 'reds' | 'run';
|
||||
runId?: number; caseFilter?: string; runs: number; sinceDays: number; dirs: string[];
|
||||
json: boolean; gate: boolean; backfill: boolean; repo?: string; workflow?: string; branch?: string;
|
||||
}
|
||||
|
||||
/** Parse eval:pass-rates flags; unknown flags, bad values and unknown case ids are errors (exit 2, before any fetch). */
|
||||
export function parsePassRatesArgs(argv: readonly string[], isKnownCase: (id: string) => boolean): { help: true } | { error: string } | PassRatesArgs {
|
||||
const out: PassRatesArgs = { view: 'rates', runs: 10, sinceDays: 60, dirs: [], json: false, gate: false, backfill: false };
|
||||
const views: string[] = [];
|
||||
for (let i = 0; i < argv.length; i++) {
|
||||
const arg = argv[i]!;
|
||||
const value = () => { const v = argv[++i]; if (v === undefined || v.startsWith('--')) throw new Error(`${arg} needs a value`); return v; };
|
||||
const count = () => { const v = value(); if (!/^[1-9]\d*$/.test(v)) throw new Error(`${arg} needs a positive integer, got ${v}`); return Number(v); };
|
||||
try {
|
||||
switch (arg) {
|
||||
case '--help': case '-h': return { help: true };
|
||||
case '--headroom': views.push('headroom'); break;
|
||||
case '--reds': views.push('reds'); break;
|
||||
case '--run': views.push('run'); out.runId = count(); break;
|
||||
case '--case': out.caseFilter = value(); break;
|
||||
case '--runs': out.runs = count(); break;
|
||||
case '--since-days': out.sinceDays = count(); break;
|
||||
case '--dir': out.dirs.push(value()); break;
|
||||
case '--repo': out.repo = value(); break;
|
||||
case '--workflow': out.workflow = value(); break;
|
||||
case '--branch': out.branch = value(); break;
|
||||
case '--json': out.json = true; break;
|
||||
case '--gate': out.gate = true; break;
|
||||
case '--backfill': out.backfill = true; break;
|
||||
default: return { error: `unknown flag ${arg}` };
|
||||
}
|
||||
} catch (error) { return { error: (error as Error).message }; }
|
||||
}
|
||||
if (views.length > 1) return { error: `choose one of --headroom, --reds, --run (got ${views.join(', ')})` };
|
||||
if (views[0]) out.view = views[0] as PassRatesArgs['view'];
|
||||
if (out.caseFilter !== undefined && !isKnownCase(out.caseFilter)) return { error: `unknown case ${out.caseFilter} (not an E2E, judge or paid test file id)` };
|
||||
if (out.view === 'run' && out.dirs.length) return { error: '--run reads one GitHub run; it cannot combine with --dir' };
|
||||
return out;
|
||||
}
|
||||
@@ -251,6 +251,7 @@ export const CODEX_CI_FILES: readonly string[] = [
|
||||
'test/skill-e2e-outside-voice.test.ts',
|
||||
'test/skill-e2e-outside-plan-disabled.test.ts',
|
||||
'test/skill-e2e-safety-codex-boundary.test.ts',
|
||||
'test/codex-e2e-multiblock-live.test.ts',
|
||||
];
|
||||
|
||||
/**
|
||||
|
||||
+158
-42
@@ -5,15 +5,19 @@ import * as fs from 'node:fs';
|
||||
import * as path from 'node:path';
|
||||
import { normalizeRelativePath } from './shard-engine';
|
||||
import { EVAL_POLICY } from '../../test/helpers/periodic-exclude-data';
|
||||
import { STALL_WINDOW_MS } from '../../test/helpers/eval-budgets';
|
||||
import {
|
||||
isFinalizedEvalResultFile, failureClassOf, panelVerdict, sanitizeTrialError, formatTrialOutcomes, TRIAL_OUTCOME_SCHEMA, TRIAL_OUTCOMES_FILE,
|
||||
type EvalCaseKind, type PanelVerdict, type TrialFailureClass, type TrialOutcome, type TrialOutcomeRecord,
|
||||
trialCostKnown, trialFailureFields, trialSessions, type TrialFailureDetail,
|
||||
} from '../../test/helpers/eval-store';
|
||||
import { readSessionLedger, type SessionLedgerRow } from '../../test/helpers/session-ledger';
|
||||
import { publishedFence, sanitizeFixedFenceLine } from './published-text';
|
||||
import { E2E_KINDS } from '../../test/helpers/touchfiles-data';
|
||||
import { manualReviewProblem } from '../../test/helpers/cookie-workflow-manual-review';
|
||||
import { writeNegativeReceipt, writePanelReceipt } from '../e2e-shard-reuse';
|
||||
import { E2E_TOUCHFILES, E2E_TIERS, LLM_JUDGE_TOUCHFILES } from '../../test/helpers/touchfiles';
|
||||
import { CASE_TEST_NAMES, type CaseTrialPlan, caseTrialPlan, fileCaseRegistration, shardCaseId, shardFile, trialShardKey } from './paid-cases';
|
||||
import { CASE_TEST_NAMES, type CaseTrialPlan, caseTestNamePattern, caseTrialPlan, fileCaseRegistration, shardCaseId, shardFile, trialShardKey } from './paid-cases';
|
||||
import { type ManifestEntry, PAID_TEST_DURATIONS_FILE, type PaidRunManifest, type SliceResult, collectorOutcomeCounts, formatProfileCoverage, loadPaidTestDurations, mergePaidTestDurations, parseRunManifest, trialPanelKey, verifySliceResults, writePaidTestDurations } from './paid-plan';
|
||||
import { DEFAULT_JOBS, type PaidCaseSelection, type PaidTier, ROOT, type ShardTrialRecord } from './paid-types';
|
||||
import { collectPaidTestFiles, isAllSkippedPass, shardSlug, paidSelectionEnv } from './paid-select';
|
||||
@@ -21,14 +25,39 @@ import type { RunShardsOptions } from '../test-paid-shards';
|
||||
|
||||
// ─── Local diagnosis: one case through the CI panel runner (A9) ────────────
|
||||
|
||||
/** The one paid file that statically registers `id`, else a thrown reason. */
|
||||
/** How --case runs exactly one case: trials selected by its Bun test name, or its whole file when that file registers no other case. */
|
||||
export interface CaseSelection { file: string; mode: 'name' | 'file'; reason: string }
|
||||
|
||||
const escapeRe = (text: string) => text.replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
|
||||
|
||||
/**
|
||||
* The paid file that registers `id` and how a trial selects it. A file whose
|
||||
* registration is complete wins over one that also lists the id; a test
|
||||
* registered literally under the id (or its CASE_TEST_NAMES label) is
|
||||
* selected by name; a loop that registers ids at runtime is selected by the
|
||||
* id it names its test; a file registering only this case runs whole (its
|
||||
* test names are not the id). Anything else throws before any process starts.
|
||||
*/
|
||||
export function caseSelection(id: string, rootDir = ROOT, discovered = collectPaidTestFiles(rootDir)): CaseSelection {
|
||||
const files = discovered.map(file => {
|
||||
const source = fs.readFileSync(path.join(rootDir, file), 'utf8');
|
||||
return { file, source, ...fileCaseRegistration(file, source) };
|
||||
}).filter(f => f.registered.includes(id));
|
||||
const owners = files.filter(f => f.known).length ? files.filter(f => f.known) : files;
|
||||
if (owners.length !== 1) throw new Error(`--case ${id}: ${owners.length ? `registered by ${owners.map(f => f.file).join(', ')}` : 'no paid file statically registers it'}; it needs exactly one`);
|
||||
const owner = owners[0]!;
|
||||
const name = escapeRe(CASE_TEST_NAMES[id] ?? id);
|
||||
if (owner.registered.length === 1) return { file: owner.file, mode: 'file', reason: 'the whole file (it registers no other case)' };
|
||||
if (new RegExp(`\\b(?:test|it|test\\.\\w+|test(?:Concurrent)?IfSelected)\\s*\\(\\s*(['"\`])${name}\\1|testName\\s*:\\s*(['"\`])${name}\\2`).test(owner.source)) {
|
||||
return { file: owner.file, mode: 'name', reason: `its Bun test is named ${CASE_TEST_NAMES[id] ?? id}` };
|
||||
}
|
||||
if (owner.computed) return { file: owner.file, mode: 'name', reason: `its loop names the Bun test ${id} at runtime` };
|
||||
throw new Error(`--case ${id}: ${owner.file} registers several cases and none of its tests is named ${id}; add its test name to CASE_TEST_NAMES`);
|
||||
}
|
||||
|
||||
/** The one paid file that registers `id`, else a thrown reason. */
|
||||
export function caseFile(id: string, rootDir = ROOT, discovered = collectPaidTestFiles(rootDir)): string {
|
||||
const owners = discovered.filter(file => {
|
||||
const { registered, known } = fileCaseRegistration(file, fs.readFileSync(path.join(rootDir, file), 'utf8'));
|
||||
return known && registered.includes(id);
|
||||
});
|
||||
if (owners.length !== 1) throw new Error(`--case ${id}: ${owners.length ? `registered by ${owners.join(', ')}` : 'no paid file statically registers it'}; it needs exactly one`);
|
||||
return owners[0]!;
|
||||
return caseSelection(id, rootDir, discovered).file;
|
||||
}
|
||||
|
||||
/**
|
||||
@@ -41,15 +70,13 @@ export function caseFile(id: string, rootDir = ROOT, discovered = collectPaidTes
|
||||
export async function runCaseDiagnosis(id: string, options: {
|
||||
trials?: number; jobs?: number; withinShardConcurrency?: number; timeoutMs?: number; rootDir?: string; env?: NodeJS.ProcessEnv;
|
||||
evalDirBase?: string; commandFor?: RunShardsOptions['commandFor']; log?: (line: string) => void; file?: string;
|
||||
/** The CLI passes its own runner: importing the entry module during its top-level await never resolves. */
|
||||
runShards?: typeof import('../test-paid-shards').runPaidShards;
|
||||
} = {}): Promise<PanelVerdict> {
|
||||
const rootDir = options.rootDir ?? ROOT;
|
||||
const log = options.log ?? ((line: string) => console.log(line));
|
||||
const file = options.file ?? caseFile(id, rootDir);
|
||||
const testName = CASE_TEST_NAMES[id] ?? id;
|
||||
if (!options.commandFor && !fs.readFileSync(path.join(rootDir, file), 'utf8').includes(testName)) {
|
||||
throw new Error(`--case ${id}: ${file} has no literal Bun test named "${testName}", so a trial could not select it. `
|
||||
+ 'Run the whole file (bun test <file> with EVALS=1) or add its literal name to CASE_TEST_NAMES.');
|
||||
}
|
||||
const selection = options.file ? { file: options.file, mode: 'name' as const } : caseSelection(id, rootDir);
|
||||
const file = selection.file;
|
||||
const policy = caseTrialPlan(id);
|
||||
const n = options.trials ?? policy.panel.n;
|
||||
const plan: CaseTrialPlan = { ...policy, panel: { n, k: policy.panel.k === policy.panel.n ? n : Math.min(policy.panel.k, n) } };
|
||||
@@ -57,14 +84,27 @@ export async function runCaseDiagnosis(id: string, options: {
|
||||
const tier = E2E_TIERS[id] as PaidTier;
|
||||
log(`[test:paid] --case ${id}: ${n} trial(s) of ${file} (kind ${plan.kind}, PASS at ${plan.panel.k}/${n}${plan.quarantined ? ', quarantined' : ''}), tier=${tier}`);
|
||||
// The runner is the CLI module; load it lazily so this library never imports it statically (no cycle).
|
||||
const { runPaidShards } = await import('../test-paid-shards');
|
||||
const summary = await runPaidShards(keys.map(key => [key]), {
|
||||
jobs: Math.min(options.jobs ?? DEFAULT_JOBS, n), withinShardConcurrency: options.withinShardConcurrency, timeoutMs: options.timeoutMs,
|
||||
rootDir, log, commandFor: options.commandFor, evalDirBase: options.evalDirBase,
|
||||
trials: Object.fromEntries(keys.map(key => [key, plan])),
|
||||
const runPaidShards = options.runShards ?? (await import('../test-paid-shards')).runPaidShards;
|
||||
const shardOptions = { withinShardConcurrency: options.withinShardConcurrency, timeoutMs: options.timeoutMs, rootDir, log, commandFor: options.commandFor,
|
||||
env: { ...(options.env ?? process.env), EVALS: '1', EVALS_TIER: tier, EVALS_PREFLIGHT_OK: '1', EVALS_ALL: '1',
|
||||
...paidSelectionEnv('full', { e2e: [id], judges: [] }, `--case ${id}`) },
|
||||
});
|
||||
...paidSelectionEnv('full', { e2e: [id], judges: [] }, `--case ${id}`) } };
|
||||
if (selection.mode === 'file') {
|
||||
// The whole file is the case: one file shard per trial, each with its own eval dir, judged by shard status as CI does.
|
||||
const fileTrials = [];
|
||||
for (let trial = 1; trial <= n; trial++) {
|
||||
const [outcome] = (await runPaidShards([[file]], { ...shardOptions, jobs: 1,
|
||||
...(options.evalDirBase ? { evalDirBase: path.join(options.evalDirBase, `t${trial}`) } : {}) })).outcomes;
|
||||
if (!outcome || outcome.runnerError !== undefined || isAllSkippedPass(outcome) || !['passed', 'failed', 'timed-out'].includes(outcome.status)) continue;
|
||||
fileTrials.push({ trial, outcome: outcome.status === 'passed' ? 'passed' as const : 'failed' as const,
|
||||
...(outcome.status === 'passed' ? {} : { failure_class: outcome.status === 'timed-out' ? 'timeout' as const : 'assertion' as const }) });
|
||||
}
|
||||
const verdict = panelVerdict({ case: id, kind: plan.kind, panel: plan.panel, trials: fileTrials, quarantined: plan.quarantined });
|
||||
log(formatPanelLine({ ...verdict, file, slices: {} }, tier));
|
||||
return verdict;
|
||||
}
|
||||
const summary = await runPaidShards(keys.map(key => [key]), { ...shardOptions,
|
||||
jobs: Math.min(options.jobs ?? DEFAULT_JOBS, n), evalDirBase: options.evalDirBase,
|
||||
trials: Object.fromEntries(keys.map(key => [key, plan])) });
|
||||
const trials = summary.outcomes.flatMap(outcome => outcome.trial && outcome.trial.outcome !== null ? [{
|
||||
trial: outcome.trial.trial, outcome: outcome.trial.outcome, ...(outcome.trial.failure_class ? { failure_class: outcome.trial.failure_class } : {}),
|
||||
...(outcome.trial.exit_reason ? { exit_reason: outcome.trial.exit_reason } : {}), ...(outcome.trial.error ? { error: outcome.trial.error } : {}),
|
||||
@@ -187,6 +227,20 @@ export interface JUnitCensus {
|
||||
|
||||
const bump = (counts: Map<string, number>, key: string) => counts.set(key, (counts.get(key) ?? 0) + 1);
|
||||
|
||||
/** Ledger rows of one JUnit case: its own id or test name, else the shard's only case. */
|
||||
function caseSessions(rows: SessionLedgerRow[], id: string, onlyCase: boolean): SessionLedgerRow[] {
|
||||
return rows.filter(row => row.case === id || (!row.case && (onlyCase || row.test_name === id || row.test_name === CASE_TEST_NAMES[id])));
|
||||
}
|
||||
|
||||
/** The artifact holding a slice's shards: the downloaded directory's own name when it is one. */
|
||||
/** The red-census guide the report and every eval:pass-rates view link. */
|
||||
export const CENSUS_RED_GUIDE = 'docs/evals/census-red.md';
|
||||
|
||||
export function sliceArtifactName(root: string, slice: number, attempt: number): string {
|
||||
const name = path.basename(root);
|
||||
return /^(?:paid-slice|gate-census)-\d+-a\d+$/.test(name) ? name : `slice ${slice} attempt ${attempt}`;
|
||||
}
|
||||
|
||||
export function junitCensus(manifest: PaidRunManifest, artifacts: ReportArtifact[], recordsByShard: Map<string, any[]>,
|
||||
base: Record<string, unknown>, cliVersion?: string): JUnitCensus {
|
||||
const census: JUnitCensus = { ruleCases: [], history: [], failedShards: new Set(), skipped: [], deselected: new Map(), shardSkipReasons: new Map(), unattributed: [] };
|
||||
@@ -207,6 +261,7 @@ export function junitCensus(manifest: PaidRunManifest, artifacts: ReportArtifact
|
||||
else census.unattributed.push(`${key} :: ${tc.name || '(no name)'} (${tc.outcome})`);
|
||||
}
|
||||
const records = recordsByShard.get(key) ?? [];
|
||||
const ledger = readSessionLedger(path.join(root, 'shards', shardSlug([key])));
|
||||
for (const [id, cases] of byCase) {
|
||||
const kind = (E2E_KINDS[id] ?? 'rule') as EvalCaseKind;
|
||||
const failed = cases.find(tc => tc.outcome === 'failed');
|
||||
@@ -216,18 +271,24 @@ export function junitCensus(manifest: PaidRunManifest, artifacts: ReportArtifact
|
||||
const failureClass: TrialFailureClass | undefined = !failed ? undefined
|
||||
: failed.failureType === 'TimeoutError' ? 'timeout' : failedRecord ? failureClassOf(failedRecord) : 'assertion';
|
||||
const error = sanitizeTrialError(failedRecord?.error ?? failed?.message);
|
||||
const rerun = Object.hasOwn(E2E_TIERS, id) ? rerunCommand(manifest.tier, id, 1) : `EVALS=1 EVALS_TIER=${manifest.tier} bun test ${shardFile(key)}`;
|
||||
const sessions = caseSessions(ledger, id, byCase.size === 1);
|
||||
// A case's JUnit times sum its concurrent tests (plan-mode-no-op: 346 s in a 176 s shard), so its wall is capped by the shard's.
|
||||
const diagnosis = failed && failureClass ? trialFailureFields({ failure_class: failureClass, exit_reason: failedRecord?.exit_reason,
|
||||
error: failedRecord?.error ?? failed.message, sessions, record: failedRecord }, STALL_WINDOW_MS) : undefined;
|
||||
if (failed) census.failedShards.add(key);
|
||||
if (caseOutcome === 'skipped') {
|
||||
bump(reasons, RUNTIME_SKIP_REASON);
|
||||
census.skipped.push(`◌ ${id} ${key} (${cases.map(tc => tc.name).join('; ')}): ${RUNTIME_SKIP_REASON}`);
|
||||
}
|
||||
census.ruleCases.push({ id, kind, outcome: caseOutcome,
|
||||
...(failed ? { line: `✗ ${id} ${kind} FAIL ${failureClass}${failedRecord?.exit_reason === 'timeout' && failedRecord?.timeout_at_turn !== undefined ? ` at turn ${failedRecord.timeout_at_turn}` : ''}${error ? ` — ${error}` : ''} [slice ${result.sliceIndex}, attempt ${result.attempt ?? 1}] rerun: ${rerun}` } : {}) });
|
||||
...(failed ? { line: `✗ ${id} ${kind} FAIL ${failureClass}${failedRecord?.exit_reason === 'timeout' && failedRecord?.timeout_at_turn !== undefined ? ` at turn ${failedRecord.timeout_at_turn}` : ''}${error ? ` — ${error}` : ''}`
|
||||
+ `${diagnosisText(diagnosis, error)} [slice ${result.sliceIndex}, attempt ${result.attempt ?? 1}]`
|
||||
+ `${evidencePointer(sliceArtifactName(root, result.sliceIndex, result.attempt ?? 1), key, typeof base.run_id === 'string' ? base.run_id : undefined, id)}`
|
||||
+ ` after a repair: ${afterRepairCommand(manifest.tier, id, shardFile(key))}` } : {}) });
|
||||
census.history.push({ ...base, case: id, file: shardFile(key), kind, trial: 1, panel: { n: 1, k: 1 }, outcome: caseOutcome,
|
||||
...(failureClass ? { failure_class: failureClass } : {}), ...(failedRecord?.exit_reason ? { exit_reason: String(failedRecord.exit_reason) } : {}),
|
||||
...(error && failed ? { error } : {}), duration_ms: cases.reduce((sum, tc) => sum + tc.timeMs, 0),
|
||||
cost_usd: Math.round(mine.reduce((sum: number, r: any) => sum + (Number(r.cost_usd) || 0), 0) * 100) / 100,
|
||||
...(error && failed ? { error } : {}), ...diagnosis, ...trialSessions(sessions), duration_ms: Math.min(cases.reduce((sum, tc) => sum + tc.timeMs, 0), outcome.elapsedMs),
|
||||
cost_usd: Math.round(mine.reduce((sum: number, r: any) => sum + (Number(r.cost_usd) || 0), 0) * 100) / 100, ...trialCostKnown(mine, sessions),
|
||||
...(typeof mine[0]?.model === 'string' ? { model: mine[0].model } : {}), ...(cliVersion ? { cli_version: cliVersion } : {}),
|
||||
quarantined: false, execution: outcome.reused ? 'reused' : 'executed', source: 'junit' } as TrialOutcomeRecord);
|
||||
}
|
||||
@@ -290,7 +351,9 @@ export function panelReports(manifest: PaidRunManifest, results: SliceResult[],
|
||||
return [{ trial: t.trial, outcome: t.outcome, attempt, ...(t.failure_class ? { failure_class: t.failure_class } : {}),
|
||||
...(t.exit_reason ? { exit_reason: t.exit_reason } : {}), ...(t.error ? { error: t.error } : {}),
|
||||
...(got.outcome.reused ? { execution: 'reused' as const } : {}),
|
||||
...(t.timeout_at_turn !== undefined ? { timeout_at_turn: t.timeout_at_turn } : {}) }];
|
||||
...(t.timeout_at_turn !== undefined ? { timeout_at_turn: t.timeout_at_turn } : {}),
|
||||
...(t.failure_cause ? { failure_cause: t.failure_cause } : {}), ...(t.failure_cause_evidence ? { failure_cause_evidence: t.failure_cause_evidence } : {}),
|
||||
...(t.failure_detail ? { failure_detail: t.failure_detail } : {}) }];
|
||||
});
|
||||
const verdict = panelVerdict({ case: shardCaseId(key)!, kind: plan.kind, panel: plan.panel, trials, quarantined: plan.quarantined });
|
||||
// Reuse is whole-panel only: every trial reused from one run, or none.
|
||||
@@ -303,29 +366,69 @@ export function panelReports(manifest: PaidRunManifest, results: SliceResult[],
|
||||
});
|
||||
}
|
||||
|
||||
type TrialDiagnosis = { failure_cause?: string; failure_cause_evidence?: string; failure_detail?: TrialFailureDetail };
|
||||
|
||||
/** Cause (when it says more than the class), its evidence and the failure detail, appended to a red line. */
|
||||
function diagnosisText(d: TrialDiagnosis | undefined, error?: string): string {
|
||||
if (!d?.failure_cause) return '';
|
||||
const evidence = d.failure_cause_evidence && d.failure_cause_evidence !== error ? `: ${d.failure_cause_evidence}` : '';
|
||||
const detail = !d.failure_detail ? ''
|
||||
: 'judge' in d.failure_detail ? ` ${d.failure_detail.judge.map(j => `${j.dimension} ${j.mean} < ${j.threshold} (${j.samples} samples)${j.rationale ? ` "${j.rationale}"` : ''}`).join('; ')}`
|
||||
: ` Expected: ${d.failure_detail.expected} · Received: ${d.failure_detail.received}`;
|
||||
return ` · cause ${d.failure_cause}${evidence}${detail}`;
|
||||
}
|
||||
|
||||
/** Where a red's transcript lives and how to fetch it. */
|
||||
function evidencePointer(artifact: string, key: string, runId: string | undefined, id: string): string {
|
||||
const arg = /^[\w./-]+$/.test(id) ? id : `'${id.replace(/'/g, `'\\''`)}'`;
|
||||
return ` evidence: ${artifact}/shards/${shardSlug([key])}${runId ? ` (fetch: bun run eval:pass-rates --run ${runId} --case ${arg})` : ''}`;
|
||||
}
|
||||
|
||||
/** Why one failed trial failed, in one line: a case timeout names its turn. */
|
||||
function trialCause(trial: PanelVerdict['trials'][number] & { timeout_at_turn?: number }): string {
|
||||
function trialCause(trial: PanelVerdict['trials'][number] & TrialDiagnosis & { timeout_at_turn?: number }): string {
|
||||
if (trial.outcome === 'skipped') return `t${trial.trial}: skipped`;
|
||||
const cls = trial.failure_class ?? 'assertion';
|
||||
const head = cls === 'timeout' || trial.exit_reason === 'timeout'
|
||||
? `timeout${trial.timeout_at_turn !== undefined ? ` at turn ${trial.timeout_at_turn}` : ''}`
|
||||
: cls;
|
||||
return `t${trial.trial}: ${head}${trial.exit_reason && trial.exit_reason !== 'timeout' ? ` (${trial.exit_reason})` : ''}${trial.error ? ` — ${trial.error}` : ''}`;
|
||||
const exit = trial.exit_reason === 'success' ? ' (session completed; check failed)'
|
||||
: trial.exit_reason && trial.exit_reason !== 'timeout' ? ` (${trial.exit_reason})` : '';
|
||||
return `t${trial.trial}: ${head}${exit}${trial.error ? ` — ${trial.error}` : ''}${diagnosisText(trial, trial.error)}`;
|
||||
}
|
||||
|
||||
export function rerunCommand(tier: PaidTier, id: string, trials: number): string {
|
||||
return `bun run scripts/test-paid-shards.ts --tier ${tier} --case ${id}${trials > 1 ? ` --trials ${trials}` : ''}`;
|
||||
/**
|
||||
* A paid run of exactly one case, for after a repair (never on unchanged
|
||||
* inputs): E2E cases through --case, judges through their selection and an
|
||||
* exact test-name filter, a keyless file (its own case) by file.
|
||||
*/
|
||||
export function afterRepairCommand(tier: PaidTier, id: string, file: string, trials = 1): string {
|
||||
if (Object.hasOwn(E2E_TIERS, id) && resolvesAlone(id)) return `bun run scripts/test-paid-shards.ts --tier ${tier} --case ${id}${trials > 1 ? ` --trials ${trials}` : ''}`;
|
||||
if (Object.hasOwn(LLM_JUDGE_TOUCHFILES, id)) {
|
||||
return `EVALS=1 EVALS_TIER=${tier} EVALS_JUDGE_SELECTION_JSON='${JSON.stringify({ version: 1, selected: [id], reason: 'after a repair' })}' bun test ${file} -t '${caseTestNamePattern([id])}'`;
|
||||
}
|
||||
return `EVALS=1 EVALS_TIER=${tier} bun test ${file}`;
|
||||
}
|
||||
|
||||
/** One line per non-PASS or split panel verdict. */
|
||||
export function formatPanelLine(panel: PanelReport, tier: PaidTier): string {
|
||||
const resolved = new Map<string, boolean>();
|
||||
/** Whether --case can select this E2E id (a case several files register runs by its shard file instead). */
|
||||
function resolvesAlone(id: string): boolean {
|
||||
if (!resolved.has(id)) { try { caseSelection(id); resolved.set(id, true); } catch { resolved.set(id, false); } }
|
||||
return resolved.get(id)!;
|
||||
}
|
||||
|
||||
/** One line per non-PASS or split panel verdict. `artifacts` names each slice's artifact (slice -> name). */
|
||||
export function formatPanelLine(panel: PanelReport, tier: PaidTier, ctx: { artifacts?: Map<number, string>; runId?: string } = {}): string {
|
||||
const mark = panel.status === 'PASS' ? '⚠' : panel.failsLane ? '✗' : '◌';
|
||||
const label = panel.status === 'PASS' ? `PASS ${panel.passed}/${panel.panel.n}` : `${panel.status} ${panel.passed}/${panel.panel.n}`;
|
||||
const causes = panel.trials.filter(t => t.outcome !== 'passed').map(t => trialCause(t as any));
|
||||
const where = Object.entries(panel.slices).map(([trial, slice]) => `t${trial}@slice ${slice}`).join(', ');
|
||||
const red = panel.trials.find(t => t.outcome === 'failed');
|
||||
const slice = red ? panel.slices[red.trial] : undefined;
|
||||
const evidence = red && slice !== undefined
|
||||
? evidencePointer(ctx.artifacts?.get(slice) ?? `slice ${slice} attempt ${panel.attempt}`, trialShardKey(panel.file, panel.case, red.trial), ctx.runId, panel.case) : '';
|
||||
return `${mark} ${panel.case} ${panel.kind}${panel.quarantined ? ' (quarantined)' : ''} ${label} (${panel.marks})`
|
||||
+ `${causes.length ? ` ${causes.join('; ')}` : ''}${panel.status === 'INCOMPLETE' || panel.status === 'SKIPPED' ? ` [${panel.reason}]` : ''}`
|
||||
+ `${where ? ` [${where}, attempt ${panel.attempt}]` : ''} rerun: ${rerunCommand(tier, panel.case, panel.panel.n)}`;
|
||||
+ `${where ? ` [${where}, attempt ${panel.attempt}]` : ''}${evidence} after a repair: ${afterRepairCommand(tier, panel.case, panel.file, panel.panel.n)}`;
|
||||
}
|
||||
|
||||
export interface ReportHeadline {
|
||||
@@ -348,7 +451,13 @@ export interface ReportHeadline {
|
||||
redispatchEligible: boolean;
|
||||
}
|
||||
|
||||
export function formatHeadline(h: ReportHeadline): string[] {
|
||||
/** Known cost, never a bare total when any trial's cost is unknown, and no total when none is known. */
|
||||
export function formatCost(knownUsd: number, unknownTrials: number, trials: number): string {
|
||||
if (trials > 0 && unknownTrials === trials) return `cost unknown (no billing captured for any of ${trials} trial(s))`;
|
||||
return `cost $${knownUsd.toFixed(2)}${unknownTrials ? ` known + ${unknownTrials} of ${trials} trial(s) cost unknown` : ''}`;
|
||||
}
|
||||
|
||||
export function formatHeadline(h: ReportHeadline, cost: { unknownTrials: number; trials: number } = { unknownTrials: 0, trials: 0 }): string[] {
|
||||
const c = h.counts;
|
||||
const minutes = h.wallMs === null ? 'unknown' : `${Math.floor(h.wallMs / 60_000)}m${String(Math.round((h.wallMs % 60_000) / 1000)).padStart(2, '0')}s`;
|
||||
return [
|
||||
@@ -356,7 +465,7 @@ export function formatHeadline(h: ReportHeadline): string[] {
|
||||
` rule ${c.rule.passed}/${c.rule.total} · behavior ${c.behavior.passed}/${c.behavior.total}${c.behavior.split ? ` (${c.behavior.split} split)` : ''}`
|
||||
+ ` · judge ${c.judge.passed}/${c.judge.total} · quarantined ${c.quarantined.total} (${c.quarantined.failingLane} failing the lane)`,
|
||||
` SKIPPED ${c.skipped} · INFRA ${c.infra} · INCOMPLETE ${c.incomplete} · unattributed ${c.unattributed} · ACTION REQUIRED ${h.actionRequired}`,
|
||||
` wall ${minutes} · cost $${h.costUsd.toFixed(2)}${h.redispatchEligible ? ' · every red is machine-classified INFRA/INCOMPLETE: eligible for ONE re-dispatch as a new run (EVAL_POLICY.infraRedispatch); report both runs' : ''}`,
|
||||
` wall ${minutes} · ${formatCost(h.costUsd, cost.unknownTrials, cost.trials)}${h.redispatchEligible ? ' · every red is machine-classified INFRA/INCOMPLETE: eligible for ONE re-dispatch as a new run (EVAL_POLICY.infraRedispatch); report both runs' : ''}`,
|
||||
];
|
||||
}
|
||||
|
||||
@@ -538,11 +647,16 @@ export function runPaidReport(reportDir: string, options: { writeDurations?: boo
|
||||
history.push({ ...common(attempt), case: t.case, file: shardFile(outcome.files[0]!), kind: t.kind, trial: t.trial, panel: t.panel,
|
||||
outcome: t.outcome, ...(t.outcome === 'failed' ? { failure_class: t.failure_class ?? 'assertion' } : {}),
|
||||
...(t.exit_reason ? { exit_reason: t.exit_reason } : {}), ...(t.error ? { error: t.error } : {}),
|
||||
duration_ms: t.duration_ms, cost_usd: t.cost_usd, ...(t.model ? { model: t.model } : {}),
|
||||
...(t.outcome === 'failed' && t.failure_cause ? { failure_cause: t.failure_cause } : {}),
|
||||
...(t.outcome === 'failed' && t.failure_cause_evidence ? { failure_cause_evidence: t.failure_cause_evidence } : {}),
|
||||
...(t.outcome === 'failed' && t.failure_detail ? { failure_detail: t.failure_detail } : {}), ...(t.sessions ? { sessions: t.sessions } : {}),
|
||||
duration_ms: t.duration_ms, cost_usd: t.cost_usd, ...(t.cost_known === false ? { cost_known: false } : {}), ...(t.model ? { model: t.model } : {}),
|
||||
...(outcome.reused ? { input_identity: outcome.reused.inputKey } : {}),
|
||||
quarantined: t.quarantined, execution: outcome.reused ? 'reused' : 'executed', source: 'shard' } as TrialOutcomeRecord);
|
||||
}
|
||||
}
|
||||
const sliceArtifacts = new Map(artifacts.filter(a => (a.result.attempt ?? 1) === primary)
|
||||
.map(a => [a.result.sliceIndex, sliceArtifactName(a.root, a.result.sliceIndex, primary)] as const));
|
||||
const census = junitCensus(manifest, artifacts.filter(a => (a.result.attempt ?? 1) === primary), recordsByShard,
|
||||
common(primary), env.GSTACK_CLAUDE_CLI_VERSION);
|
||||
history.push(...census.history);
|
||||
@@ -568,7 +682,7 @@ export function runPaidReport(reportDir: string, options: { writeDurations?: boo
|
||||
if (got && got.o.status === 'passed') return [];
|
||||
if (got && census.failedShards.has(normalizeRelativePath(entry.file))) return [];
|
||||
const id = shardCaseId(entry.file);
|
||||
return [`✗ ${entry.file} rule shard ${got ? got.o.status : 'NOT REPORTED'}${got?.o.runnerError ? ` — ${sanitizeTrialError(got.o.runnerError)}` : ''} [slice ${entry.slice}, attempt ${primary}]${id ? ` rerun: ${rerunCommand(manifest.tier, id, 1)}` : ''}`];
|
||||
return [`✗ ${entry.file} rule shard ${got ? got.o.status : 'NOT REPORTED'}${got?.o.runnerError ? ` — ${sanitizeTrialError(got.o.runnerError)}` : ''} [slice ${entry.slice}, attempt ${primary}]${id ? ` after a repair: ${afterRepairCommand(manifest.tier, id, shardFile(entry.file))}` : ''}`];
|
||||
});
|
||||
const behaviorPanels = panels.filter(p => !p.quarantined && p.kind === 'behavior');
|
||||
const lanePanels = panels.filter(p => !p.quarantined);
|
||||
@@ -582,7 +696,7 @@ export function runPaidReport(reportDir: string, options: { writeDurations?: boo
|
||||
const failureLines = [
|
||||
...ruleShardFailures,
|
||||
...ruleCases.filter(c => c.line).map(c => c.line!),
|
||||
...panels.filter(p => p.status !== 'PASS' || p.split).map(p => formatPanelLine(p, manifest.tier)),
|
||||
...panels.filter(p => p.status !== 'PASS' || p.split).map(p => formatPanelLine(p, manifest.tier, { artifacts: sliceArtifacts, runId })),
|
||||
];
|
||||
const red = verdict.problems.length > 0;
|
||||
const headline: ReportHeadline = {
|
||||
@@ -603,19 +717,20 @@ export function runPaidReport(reportDir: string, options: { writeDurations?: boo
|
||||
costUsd: Math.round(costUsd * 100) / 100,
|
||||
redispatchEligible: red && infraOnly(verdict.problems),
|
||||
};
|
||||
const headlineLines = formatHeadline(headline);
|
||||
const executed = history.filter(r => r.attempt === primary && r.execution === 'executed' && r.outcome !== 'skipped');
|
||||
const headlineLines = formatHeadline(headline, { unknownTrials: executed.filter(r => r.cost_known === false).length, trials: executed.length });
|
||||
for (const line of headlineLines) console.log(line);
|
||||
if (failureLines.length) {
|
||||
console.log('[test:paid] failures and split verdicts:');
|
||||
for (const line of failureLines) console.log(` ${line}`);
|
||||
}
|
||||
for (const panel of laterPanels) console.log(` attempt ${panel.attempt} (re-run; reported, never replacing attempt ${primary}): ${formatPanelLine(panel, manifest.tier)}`);
|
||||
const fence = (lines: string[]) => ['```', ...lines.map(line => line.replace(/```/g, "'''")), '```'];
|
||||
const fence = (lines: string[]) => publishedFence(lines);
|
||||
fs.writeFileSync(summaryMdPath, [
|
||||
...fence(headlineLines),
|
||||
...(failureLines.length ? ['', '**Failures and split verdicts**', '', ...fence(failureLines)] : []),
|
||||
...(censusLines.length ? ['', '**Skipped, deselected and unattributed testcases**', '', ...fence(censusLines)] : []),
|
||||
...(verdict.problems.length ? ['', `**ACTION REQUIRED (${verdict.problems.length})**`, '', ...fence(verdict.problems.map(p => sanitizeTrialError(p) ?? p))] : []),
|
||||
...(verdict.problems.length ? ['', `**ACTION REQUIRED (${verdict.problems.length})**`, '', ...fence(verdict.problems.map(p => sanitizeTrialError(p) ?? p)), '', `Guide: ${CENSUS_RED_GUIDE}`] : []),
|
||||
].join('\n') + '\n');
|
||||
if (!manualProblems.length) fs.writeFileSync(summaryPath, JSON.stringify({ version: 2, files, totals: {
|
||||
...evidence, total: evidence.passed + evidence.failed + evidence.manual_accepted,
|
||||
@@ -626,7 +741,8 @@ export function runPaidReport(reportDir: string, options: { writeDurations?: boo
|
||||
marks: p.marks, split: p.split, quarantined: p.quarantined, failsLane: p.failsLane, redClass: p.redClass, reason: p.reason,
|
||||
trials: p.trials.map(t => ({ trial: t.trial, outcome: t.outcome, ...(t.failure_class ? { failure_class: t.failure_class } : {}),
|
||||
...(t.exit_reason ? { exit_reason: t.exit_reason } : {}), ...(t.error ? { error: t.error } : {}) })) })),
|
||||
failures: failureLines.map(line => sanitizeTrialError(line) ?? line) }, null, 2) + '\n');
|
||||
// evals.yml wraps these in a fixed fence for the PR comment.
|
||||
failures: failureLines.map(line => sanitizeFixedFenceLine(line)) }, null, 2) + '\n');
|
||||
if (verdict.problems.length) {
|
||||
console.error(`[test:paid] report: ${verdict.problems.length} problem(s):`);
|
||||
for (const problem of verdict.problems) console.error(` ✗ ${problem}`);
|
||||
|
||||
@@ -4,7 +4,7 @@
|
||||
*/
|
||||
import * as path from 'node:path';
|
||||
import { normalizeRelativePath, type LanePolicy } from './shard-engine';
|
||||
import type { EvalCaseKind, PanelShape, TrialFailureClass, TrialOutcome } from '../../test/helpers/eval-store';
|
||||
import type { EvalCaseKind, PanelShape, TrialFailureCause, TrialFailureClass, TrialFailureDetail, TrialOutcome, TrialSessionSummary } from '../../test/helpers/eval-store';
|
||||
|
||||
export const ROOT = path.resolve(import.meta.dir, '..', '..');
|
||||
|
||||
@@ -119,6 +119,11 @@ export interface ShardTrialRecord {
|
||||
cost_usd: number;
|
||||
duration_ms: number;
|
||||
model?: string;
|
||||
failure_cause?: TrialFailureCause;
|
||||
failure_cause_evidence?: string;
|
||||
failure_detail?: TrialFailureDetail;
|
||||
sessions?: TrialSessionSummary[];
|
||||
cost_known?: false;
|
||||
}
|
||||
|
||||
/** Upload reserve between a slice's in-process deadline and its CI job ceiling (ENG-2). */
|
||||
|
||||
@@ -0,0 +1,41 @@
|
||||
/**
|
||||
* The one sanitizer for model- or fixture-written text that reaches a public
|
||||
* issue, PR comment or committed file (census report, tracking issue,
|
||||
* eval:pass-rates output). Control and bidi characters are stripped, the
|
||||
* repo's credential-scan patterns (lib/redact-engine, the engine the quality
|
||||
* gate's credential scan runs) are redacted, @-mentions are neutralized and
|
||||
* length is capped. Inline text also has HTML and Markdown link brackets
|
||||
* escaped; blocks go inside a code fence longer than any backtick run they
|
||||
* contain, so their content stays literal and cannot close the fence.
|
||||
*/
|
||||
import { redactFindingSpans } from '../../lib/redact-engine';
|
||||
|
||||
export const PUBLISHED_TEXT_MAX = 1000;
|
||||
|
||||
const cap = (text: string, max: number) => (text.length > max ? `${text.slice(0, max - 1)}…` : text);
|
||||
|
||||
function clean(text: string): string {
|
||||
// eslint-disable-next-line no-control-regex
|
||||
const stripped = text.replace(/[\u0000-\u0008\u000a-\u001f\u007f-\u009f\u200b-\u200f\u202a-\u202e\u2066-\u2069]/g, ' ').slice(0, 64 * 1024);
|
||||
const redacted = redactFindingSpans(stripped) ?? '[redacted: credential-shaped text that could not be isolated]';
|
||||
return redacted.replace(/@(?=[A-Za-z0-9_-])/g, '@\u200b');
|
||||
}
|
||||
|
||||
/** One inline line: cleaned, HTML, Markdown link brackets and backticks neutralized, capped. */
|
||||
export function sanitizePublishedText(text: string, max = PUBLISHED_TEXT_MAX): string {
|
||||
return cap(clean(text).replace(/&/g, '&').replace(/</g, '<').replace(/>/g, '>')
|
||||
.replace(/\[/g, '[').replace(/\]/g, ']').replace(/`/g, "'"), max);
|
||||
}
|
||||
|
||||
/** One line for a consumer that wraps it in a fixed three-backtick fence (the PR comment): backticks become quotes. */
|
||||
export function sanitizeFixedFenceLine(text: string, max = PUBLISHED_TEXT_MAX): string {
|
||||
return cap(clean(text).replace(/`/g, "'"), max);
|
||||
}
|
||||
|
||||
/** Cleaned, capped lines inside a fence longer than the longest backtick run they contain. */
|
||||
export function publishedFence(lines: readonly string[], max = PUBLISHED_TEXT_MAX): string[] {
|
||||
const body = lines.map(line => cap(clean(line), max));
|
||||
const longest = Math.max(0, ...body.flatMap(line => (line.match(/`+/g) ?? []).map(run => run.length)));
|
||||
const fence = '`'.repeat(Math.max(3, longest + 1));
|
||||
return [fence, ...body, fence];
|
||||
}
|
||||
+140
-138
@@ -1,109 +1,109 @@
|
||||
{
|
||||
"version": 2,
|
||||
"recordedAt": "2026-09-29T19:29:33.835Z",
|
||||
"recordedAt": "2026-10-04T18:45:27.037Z",
|
||||
"tiers": {
|
||||
"gate": {
|
||||
"test/llm-judge-recommendation.test.ts": 1000,
|
||||
"test/skill-e2e-ask-user-question-format-compliance.test.ts": 63904,
|
||||
"test/skill-e2e-ask-user-question-format-compliance.test.ts": 54768,
|
||||
"test/skill-e2e-autoplan-dual-voice.test.ts": 1000,
|
||||
"test/skill-e2e-bws.test.ts": 101092,
|
||||
"test/skill-e2e-bws.test.ts": 74261,
|
||||
"test/skill-e2e-context-skills.test.ts": 1000,
|
||||
"test/skill-e2e-coverage-audit.test.ts": 50849,
|
||||
"test/skill-e2e-cso.test.ts": 227394,
|
||||
"test/skill-e2e-deploy.test.ts#benchmark-workflow": 62156,
|
||||
"test/skill-e2e-deploy.test.ts#canary-workflow": 50673,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-first-run": 125350,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-review-gate": 177755,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-workflow": 200984,
|
||||
"test/skill-e2e-deploy.test.ts#setup-deploy-workflow": 51171,
|
||||
"test/skill-e2e-coverage-audit.test.ts": 65669,
|
||||
"test/skill-e2e-cso.test.ts": 229620,
|
||||
"test/skill-e2e-deploy.test.ts#benchmark-workflow": 120646,
|
||||
"test/skill-e2e-deploy.test.ts#canary-workflow": 42804,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-first-run": 103709,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-review-gate": 147499,
|
||||
"test/skill-e2e-deploy.test.ts#land-and-deploy-workflow": 148339,
|
||||
"test/skill-e2e-deploy.test.ts#setup-deploy-workflow": 36798,
|
||||
"test/skill-e2e-design.test.ts": 261000,
|
||||
"test/skill-e2e-design.test.ts#design-review-detector-shim": 42811,
|
||||
"test/skill-e2e-design.test.ts#design-review-detector-shim-dom": 81914,
|
||||
"test/skill-e2e-design.test.ts#design-review-plugin-handoff": 126642,
|
||||
"test/skill-e2e-design.test.ts#plan-design-review-no-ui-scope": 24431,
|
||||
"test/skill-e2e-diagram.test.ts": 29998,
|
||||
"test/skill-e2e-docsync-spawned.test.ts": 137929,
|
||||
"test/skill-e2e-design.test.ts#design-review-detector-shim": 45224,
|
||||
"test/skill-e2e-design.test.ts#design-review-detector-shim-dom": 87823,
|
||||
"test/skill-e2e-design.test.ts#design-review-plugin-handoff": 74874,
|
||||
"test/skill-e2e-design.test.ts#plan-design-review-no-ui-scope": 39970,
|
||||
"test/skill-e2e-diagram.test.ts": 30124,
|
||||
"test/skill-e2e-docsync-spawned.test.ts": 108078,
|
||||
"test/skill-e2e-gbrain-roundtrip-local.test.ts": 1000,
|
||||
"test/skill-e2e-hermetic-canary.test.ts": 8813,
|
||||
"test/skill-e2e-investigate-owned-completion.test.ts": 45178,
|
||||
"test/skill-e2e-investigate-owned-termination.test.ts": 48608,
|
||||
"test/skill-e2e-hermetic-canary.test.ts": 4951,
|
||||
"test/skill-e2e-investigate-owned-completion.test.ts": 36179,
|
||||
"test/skill-e2e-investigate-owned-termination.test.ts": 48428,
|
||||
"test/skill-e2e-ios-device.test.ts": 1000,
|
||||
"test/skill-e2e-learnings.test.ts": 28899,
|
||||
"test/skill-e2e-office-hours-auto-mode.test.ts": 83101,
|
||||
"test/skill-e2e-learnings.test.ts": 33590,
|
||||
"test/skill-e2e-office-hours-auto-mode.test.ts": 78399,
|
||||
"test/skill-e2e-office-hours-brain-writeback.test.ts": 1000,
|
||||
"test/skill-e2e-office-hours-phase4.test.ts": 1000,
|
||||
"test/skill-e2e-office-hours.test.ts": 1000,
|
||||
"test/skill-e2e-plan-ceo-finding-floor.test.ts": 297738,
|
||||
"test/skill-e2e-plan-ceo-plan-mode.test.ts": 36812,
|
||||
"test/skill-e2e-plan-ceo-finding-floor.test.ts": 248087,
|
||||
"test/skill-e2e-plan-ceo-plan-mode.test.ts": 112067,
|
||||
"test/skill-e2e-plan-decision-classification.test.ts": 1000,
|
||||
"test/skill-e2e-plan-design-with-ui.test.ts": 240000,
|
||||
"test/skill-e2e-plan-devex-finding-floor.test.ts": 207635,
|
||||
"test/skill-e2e-plan-design-with-ui.test.ts": 285788,
|
||||
"test/skill-e2e-plan-devex-finding-floor.test.ts": 181446,
|
||||
"test/skill-e2e-plan-devex-peer-comparison-classification.test.ts": 1000,
|
||||
"test/skill-e2e-plan-devex-plan-mode.test.ts": 115679,
|
||||
"test/skill-e2e-plan-devex-plan-mode.test.ts": 105953,
|
||||
"test/skill-e2e-plan-format.test.ts": 1000,
|
||||
"test/skill-e2e-plan-mode-no-op.test.ts": 217184,
|
||||
"test/skill-e2e-plan-mode-no-op.test.ts": 271596,
|
||||
"test/skill-e2e-plan-prosons.test.ts": 1000,
|
||||
"test/skill-e2e-plan-tune.test.ts": 63925,
|
||||
"test/skill-e2e-plan-tune.test.ts": 70702,
|
||||
"test/skill-e2e-plan.test.ts": 251000,
|
||||
"test/skill-e2e-plan.test.ts#plan-review-report": 76406,
|
||||
"test/skill-e2e-plan.test.ts#plan-review-report": 197666,
|
||||
"test/skill-e2e-qa-bugs.test.ts": 1000,
|
||||
"test/skill-e2e-qa-callers.test.ts": 737164,
|
||||
"test/skill-e2e-qa-callers.test.ts#review-exploratory-small-cli": 294865,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-late-input": 294865,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-plan-checks": 294865,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-small-cli": 294865,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-unavailable": 294865,
|
||||
"test/skill-e2e-qa-functional-fix.test.ts": 235133,
|
||||
"test/skill-e2e-qa-functional.test.ts": 356641,
|
||||
"test/skill-e2e-qa-callers.test.ts#review-exploratory-small-cli": 224426,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-late-input": 234381,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-plan-checks": 192841,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-small-cli": 147399,
|
||||
"test/skill-e2e-qa-callers.test.ts#ship-exploratory-unavailable": 191231,
|
||||
"test/skill-e2e-qa-functional-fix.test.ts": 200616,
|
||||
"test/skill-e2e-qa-functional.test.ts": 188766,
|
||||
"test/skill-e2e-qa-workflow.test.ts#qa-bootstrap": 35863,
|
||||
"test/skill-e2e-qa-workflow.test.ts#qa-only-no-fix": 201333,
|
||||
"test/skill-e2e-qa-workflow.test.ts#qa-quick": 169852,
|
||||
"test/skill-e2e-retro.test.ts": 141804,
|
||||
"test/skill-e2e-retro.test.ts": 200434,
|
||||
"test/skill-e2e-review-army.test.ts": 520000,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-delivery-audit": 50219,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-json-findings": 21332,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-migration-safety": 145112,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-perf-n-plus-one": 241299,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-quality-score": 86836,
|
||||
"test/skill-e2e-review-attribution.test.ts": 77592,
|
||||
"test/skill-e2e-review.test.ts": 136132,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-delivery-audit": 39313,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-json-findings": 22950,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-migration-safety": 162617,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-perf-n-plus-one": 114005,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-quality-score": 55496,
|
||||
"test/skill-e2e-review-attribution.test.ts": 80562,
|
||||
"test/skill-e2e-review.test.ts": 132863,
|
||||
"test/skill-e2e-safety-codex-boundary.test.ts": 1000,
|
||||
"test/skill-e2e-safety-codex-consult.test.ts": 82940,
|
||||
"test/skill-e2e-safety-codex-consult.test.ts": 150155,
|
||||
"test/skill-e2e-safety-design-risk.test.ts": 1000,
|
||||
"test/skill-e2e-safety-ios-demo.test.ts": 1000,
|
||||
"test/skill-e2e-safety-pair-agent.test.ts": 39860,
|
||||
"test/skill-e2e-safety-ship-evidence.test.ts": 106130,
|
||||
"test/skill-e2e-session-intelligence.test.ts": 55332,
|
||||
"test/skill-e2e-safety-pair-agent.test.ts": 36028,
|
||||
"test/skill-e2e-safety-ship-evidence.test.ts": 100833,
|
||||
"test/skill-e2e-session-intelligence.test.ts": 53080,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts": 680000,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-index-flags": 190202,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-path-eligibility": 205615,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-prior-coverage": 214728,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-index-flags": 145972,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-path-eligibility": 170775,
|
||||
"test/skill-e2e-shared-libs-paths.test.ts#shared-libs-review-prior-coverage": 172422,
|
||||
"test/skill-e2e-shared-libs.test.ts": 658000,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-read-only": 197454,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-review-lifecycle": 267080,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-review-revalidation": 316752,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-unsupported-git": 170776,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-read-only": 160556,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-review-lifecycle": 164017,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-review-revalidation": 281097,
|
||||
"test/skill-e2e-shared-libs.test.ts#shared-libs-unsupported-git": 87581,
|
||||
"test/skill-e2e-ship-docsync.test.ts": 129000,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-completion": 366116,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-current": 434135,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-failure": 164667,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-late-result": 214218,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-launch-failure": 140987,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-missing-asset": 142564,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-missing-marker": 117648,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-recovery": 225432,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-stale-after": 269572,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-stale-before": 197633,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-store": 396588,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-timeout-unsettled": 152278,
|
||||
"test/skill-e2e-ship-hook-consent.test.ts": 69995,
|
||||
"test/skill-e2e-ship-hook-refresh.test.ts": 70007,
|
||||
"test/skill-e2e-ship-skip.test.ts": 80318,
|
||||
"test/skill-e2e-skillify.test.ts": 191127,
|
||||
"test/skill-e2e-test-value.test.ts": 332000,
|
||||
"test/skill-e2e-third-party-actions.test.ts": 82102,
|
||||
"test/skill-e2e-triage.test.ts": 54088,
|
||||
"test/skill-e2e-workflow.test.ts": 184925,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-completion": 398349,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-current": 382630,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-failure": 238690,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-late-result": 132034,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-launch-failure": 53611,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-missing-asset": 106550,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-missing-marker": 59963,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-recovery": 132137,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-stale-after": 155419,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-stale-before": 164443,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-store": 392835,
|
||||
"test/skill-e2e-ship-docsync.test.ts#ship-docsync-timeout-unsettled": 63645,
|
||||
"test/skill-e2e-ship-hook-consent.test.ts": 49572,
|
||||
"test/skill-e2e-ship-hook-refresh.test.ts": 40853,
|
||||
"test/skill-e2e-ship-skip.test.ts": 81807,
|
||||
"test/skill-e2e-skillify.test.ts": 177262,
|
||||
"test/skill-e2e-test-value.test.ts": 216843,
|
||||
"test/skill-e2e-third-party-actions.test.ts": 71961,
|
||||
"test/skill-e2e-triage.test.ts": 64816,
|
||||
"test/skill-e2e-workflow.test.ts": 265175,
|
||||
"test/skill-llm-eval.test.ts": 354000,
|
||||
"test/skill-routing-e2e.test.ts": 1000
|
||||
},
|
||||
@@ -123,98 +123,100 @@
|
||||
"test/carve-section-loading.test.ts#carve-section-loading-review": 247902,
|
||||
"test/carve-section-loading.test.ts#carve-section-loading-setup-gbrain": 80994,
|
||||
"test/carve-section-loading.test.ts#carve-section-loading-spec": 176639,
|
||||
"test/codex-e2e-recommendation-substance.test.ts": 17984,
|
||||
"test/codex-e2e-shared-libs.test.ts": 53442,
|
||||
"test/codex-e2e-sol-scope.test.ts": 37063,
|
||||
"test/codex-e2e.test.ts": 120390,
|
||||
"test/llm-judge-recommendation.test.ts": 15334,
|
||||
"test/skill-e2e-arm-benchmark.test.ts": 75588,
|
||||
"test/codex-e2e-multiblock-live.test.ts": 180000,
|
||||
"test/codex-e2e-recommendation-substance.test.ts": 14617,
|
||||
"test/codex-e2e-shared-libs.test.ts": 49563,
|
||||
"test/codex-e2e-sol-scope.test.ts": 32228,
|
||||
"test/codex-e2e.test.ts": 110677,
|
||||
"test/llm-judge-recommendation.test.ts": 21562,
|
||||
"test/skill-e2e-arm-benchmark.test.ts": 72714,
|
||||
"test/skill-e2e-aside.test.ts": 1000,
|
||||
"test/skill-e2e-auq-consistency.test.ts": 70429,
|
||||
"test/skill-e2e-auq-matrix.test.ts": 165796,
|
||||
"test/skill-e2e-auq-verbose-vs-carved-ab.test.ts": 58547,
|
||||
"test/skill-e2e-auto-decide-preserved.test.ts": 154931,
|
||||
"test/skill-e2e-autoplan-dual-voice.test.ts": 167918,
|
||||
"test/skill-e2e-benchmark-providers.test.ts": 9495,
|
||||
"test/skill-e2e-auq-consistency.test.ts": 63226,
|
||||
"test/skill-e2e-auq-matrix.test.ts": 145925,
|
||||
"test/skill-e2e-auq-verbose-vs-carved-ab.test.ts": 65480,
|
||||
"test/skill-e2e-auto-decide-preserved.test.ts": 234151,
|
||||
"test/skill-e2e-autoplan-dual-voice.test.ts": 154207,
|
||||
"test/skill-e2e-benchmark-providers.test.ts": 8847,
|
||||
"test/skill-e2e-bws.test.ts": 1000,
|
||||
"test/skill-e2e-context-skills.test.ts": 169497,
|
||||
"test/skill-e2e-context-skills.test.ts": 144874,
|
||||
"test/skill-e2e-coverage-audit.test.ts": 1000,
|
||||
"test/skill-e2e-cso.test.ts": 253498,
|
||||
"test/skill-e2e-cso.test.ts": 303949,
|
||||
"test/skill-e2e-design-model-smoke.test.ts": 90000,
|
||||
"test/skill-e2e-design.test.ts": 817000,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-core": 159257,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-existing": 175607,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-preview": 114638,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-research": 109573,
|
||||
"test/skill-e2e-design.test.ts#design-html-slop-gate": 158970,
|
||||
"test/skill-e2e-design.test.ts#plan-design-review-plan-mode": 300166,
|
||||
"test/skill-e2e-diagram.test.ts": 55187,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-core": 173162,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-existing": 225306,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-preview": 107037,
|
||||
"test/skill-e2e-design.test.ts#design-consultation-research": 99890,
|
||||
"test/skill-e2e-design.test.ts#design-html-slop-gate": 139668,
|
||||
"test/skill-e2e-design.test.ts#plan-design-review-plan-mode": 169183,
|
||||
"test/skill-e2e-diagram.test.ts": 52819,
|
||||
"test/skill-e2e-gbrain-roundtrip-local.test.ts": 1000,
|
||||
"test/skill-e2e-health.test.ts": 188584,
|
||||
"test/skill-e2e-health.test.ts": 164674,
|
||||
"test/skill-e2e-hermetic-canary.test.ts": 1000,
|
||||
"test/skill-e2e-ios-device.test.ts": 1000,
|
||||
"test/skill-e2e-learnings.test.ts": 1000,
|
||||
"test/skill-e2e-office-hours-brain-writeback.test.ts": 203044,
|
||||
"test/skill-e2e-office-hours-design-draft.test.ts": 285868,
|
||||
"test/skill-e2e-office-hours-design-draft.test.ts": 210162,
|
||||
"test/skill-e2e-office-hours-phase4.test.ts": 41421,
|
||||
"test/skill-e2e-office-hours-section-loading.test.ts": 1200000,
|
||||
"test/skill-e2e-office-hours.test.ts": 136970,
|
||||
"test/skill-e2e-outside-plan-disabled.test.ts": 54255,
|
||||
"test/skill-e2e-outside-voice.test.ts": 186881,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash-opus-5-5.test.ts": 56080,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash-sonnet.test.ts": 85282,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash.test.ts": 80465,
|
||||
"test/skill-e2e-overlay-harness-opus-4-7-effort-match-trivial.test.ts": 49282,
|
||||
"test/skill-e2e-overlay-harness-opus-4-7-literal-interpretation.test.ts": 318532,
|
||||
"test/skill-e2e-plan-ceo-mode-routing.test.ts": 443504,
|
||||
"test/skill-e2e-plan-ceo-review-section-loading.test.ts": 342250,
|
||||
"test/skill-e2e-outside-plan-disabled.test.ts": 72070,
|
||||
"test/skill-e2e-outside-voice.test.ts": 195221,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash-opus-5-5.test.ts": 57374,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash-sonnet.test.ts": 77984,
|
||||
"test/skill-e2e-overlay-harness-claude-dedicated-tools-vs-bash.test.ts": 55431,
|
||||
"test/skill-e2e-overlay-harness-opus-4-7-effort-match-trivial.test.ts": 34905,
|
||||
"test/skill-e2e-overlay-harness-opus-4-7-literal-interpretation.test.ts": 226920,
|
||||
"test/skill-e2e-plan-ceo-mode-routing.test.ts": 378316,
|
||||
"test/skill-e2e-plan-ceo-review-section-loading.test.ts": 302668,
|
||||
"test/skill-e2e-plan-ceo-split-overflow.test.ts": 504266,
|
||||
"test/skill-e2e-plan-decision-classification.test.ts": 93469,
|
||||
"test/skill-e2e-plan-design-finding-floor.test.ts": 155538,
|
||||
"test/skill-e2e-plan-design-plan-mode.test.ts": 106608,
|
||||
"test/skill-e2e-plan-devex-peer-comparison-classification.test.ts": 79786,
|
||||
"test/skill-e2e-plan-decision-classification.test.ts": 88190,
|
||||
"test/skill-e2e-plan-design-finding-floor.test.ts": 131042,
|
||||
"test/skill-e2e-plan-design-plan-mode.test.ts": 111838,
|
||||
"test/skill-e2e-plan-devex-peer-comparison-classification.test.ts": 77201,
|
||||
"test/skill-e2e-plan-eng-artifact-full.test.ts": 595009,
|
||||
"test/skill-e2e-plan-eng-artifact.test.ts": 498474,
|
||||
"test/skill-e2e-plan-eng-finding-floor.test.ts": 388666,
|
||||
"test/skill-e2e-plan-eng-multi-finding-batching.test.ts": 597500,
|
||||
"test/skill-e2e-plan-eng-plan-mode.test.ts": 229226,
|
||||
"test/skill-e2e-plan-eng-artifact.test.ts": 406795,
|
||||
"test/skill-e2e-plan-eng-finding-floor.test.ts": 234397,
|
||||
"test/skill-e2e-plan-eng-multi-finding-batching.test.ts": 409161,
|
||||
"test/skill-e2e-plan-eng-plan-mode.test.ts": 170914,
|
||||
"test/skill-e2e-plan-format.test.ts": 219670,
|
||||
"test/skill-e2e-plan-prosons.test.ts": 163659,
|
||||
"test/skill-e2e-plan-tune.test.ts": 1000,
|
||||
"test/skill-e2e-plan.test.ts": 824000,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review": 136101,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review-expansion-energy": 91984,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review-selective": 267212,
|
||||
"test/skill-e2e-plan.test.ts#plan-eng-review": 128506,
|
||||
"test/skill-e2e-qa-bugs.test.ts": 333847,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review": 274619,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review-expansion-energy": 100000,
|
||||
"test/skill-e2e-plan.test.ts#plan-ceo-review-selective": 265163,
|
||||
"test/skill-e2e-plan.test.ts#plan-eng-review": 169400,
|
||||
"test/skill-e2e-qa-bugs.test.ts": 355966,
|
||||
"test/skill-e2e-qa-workflow.test.ts#qa-fix-loop": 411728,
|
||||
"test/skill-e2e-retro.test.ts": 166711,
|
||||
"test/skill-e2e-retro.test.ts": 150139,
|
||||
"test/skill-e2e-review-army.test.ts": 511000,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-consensus": 277908,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-red-team": 88882,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-simplification": 136644,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-simplification-precision": 23884,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-consensus": 107375,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-red-team": 110743,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-simplification": 140434,
|
||||
"test/skill-e2e-review-army.test.ts#review-army-simplification-precision": 25020,
|
||||
"test/skill-e2e-review-attribution.test.ts": 1000,
|
||||
"test/skill-e2e-review.test.ts": 179362,
|
||||
"test/skill-e2e-safety-codex-boundary.test.ts": 100310,
|
||||
"test/skill-e2e-safety-codex-boundary.test.ts": 102851,
|
||||
"test/skill-e2e-safety-codex-consult.test.ts": 1000,
|
||||
"test/skill-e2e-safety-design-risk.test.ts": 52700,
|
||||
"test/skill-e2e-safety-ios-demo.test.ts": 53270,
|
||||
"test/skill-e2e-safety-design-risk.test.ts": 54889,
|
||||
"test/skill-e2e-safety-ios-demo.test.ts": 100092,
|
||||
"test/skill-e2e-safety-pair-agent.test.ts": 1000,
|
||||
"test/skill-e2e-safety-ship-evidence.test.ts": 1000,
|
||||
"test/skill-e2e-session-intelligence.test.ts": 1000,
|
||||
"test/skill-e2e-setup-gbrain-bad-token.test.ts": 42912,
|
||||
"test/skill-e2e-setup-gbrain-path4-local-pglite.test.ts": 165743,
|
||||
"test/skill-e2e-setup-gbrain-remote.test.ts": 121127,
|
||||
"test/skill-e2e-shared-libs-periodic.test.ts": 369660,
|
||||
"test/skill-e2e-ship-section-loading.test.ts": 255273,
|
||||
"test/skill-e2e-setup-gbrain-bad-token.test.ts": 39292,
|
||||
"test/skill-e2e-setup-gbrain-path4-local-pglite.test.ts": 128414,
|
||||
"test/skill-e2e-setup-gbrain-remote.test.ts": 89701,
|
||||
"test/skill-e2e-shared-libs-periodic.test.ts": 217221,
|
||||
"test/skill-e2e-ship-section-loading.test.ts": 189463,
|
||||
"test/skill-e2e-skillify.test.ts": 58522,
|
||||
"test/skill-e2e-sync-gbrain-readiness.test.ts": 60234,
|
||||
"test/skill-e2e-sync-gbrain-readiness.test.ts": 47502,
|
||||
"test/skill-e2e-test-value.test.ts": 332000,
|
||||
"test/skill-e2e-third-party-actions.test.ts": 1000,
|
||||
"test/skill-e2e-triage.test.ts": 1000,
|
||||
"test/skill-e2e-workflow.test.ts": 110357,
|
||||
"test/skill-llm-eval.test.ts": 383454,
|
||||
"test/skill-routing-e2e.test.ts": 68843
|
||||
"test/skill-e2e-workflow.test.ts": 121594,
|
||||
"test/skill-llm-eval.test.ts": 339704,
|
||||
"test/skill-routing-e2e.test.ts": 81360
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -133,7 +133,7 @@ If \`NEEDS_SETUP\`:
|
||||
3. If \`bun\` is not installed:
|
||||
\`\`\`bash
|
||||
if ! command -v bun >/dev/null 2>&1; then
|
||||
BUN_VERSION="1.3.10"
|
||||
BUN_VERSION="1.4.2"
|
||||
BUN_INSTALL_SHA="bab8acfb046aac8c72407bdcce903957665d655d7acaa3e11c7c4616beae68dd"
|
||||
tmpfile=$(mktemp "\${TMPDIR:-/tmp}/bun-install.XXXXXX")
|
||||
curl -fsSL "https://bun.sh/install" -o "$tmpfile"
|
||||
|
||||
@@ -140,6 +140,10 @@ export const RESOLVERS: Record<string, ResolverFn> = {
|
||||
DX_FRAMEWORK: generateDxFramework,
|
||||
TASTE_PROFILE: generateTasteProfile,
|
||||
BIN_DIR: (ctx) => ctx.paths.binDir,
|
||||
// Quotable runtime root for paths under gstack's install (freeze/bin, bin, lib):
|
||||
// $GSTACK_ROOT on env-var hosts (each fence gets the runtime prelude), the
|
||||
// literal $HOME path on Claude, so a quoted path keeps working with spaces.
|
||||
RUNTIME_ROOT: (ctx) => getHostConfig(ctx.host).usesEnvVars ? '$GSTACK_ROOT' : `$HOME/${getHostConfig(ctx.host).globalRoot}`,
|
||||
// Literal bin dir for skills that skip the shared preamble, where
|
||||
// $GSTACK_BIN is never set (#2906): the per-install root when rendering for
|
||||
// one, else the host's default global root.
|
||||
|
||||
@@ -1,29 +1,26 @@
|
||||
import { type TemplateContext, toShellPath } from './types';
|
||||
import { type TemplateContext } from './types';
|
||||
import { binaryAssignment } from './runtime-root';
|
||||
|
||||
/**
|
||||
* {{MAKE_PDF_SETUP}} — emits the shell preamble that resolves $P to the
|
||||
* make-pdf binary. Mirrors generateDesignSetup.
|
||||
* {{MAKE_PDF_SETUP}} — emits the readiness check that resolves $P to the
|
||||
* make-pdf binary (make-pdf/dist/pdf). Mirrors generateDesignSetup: the
|
||||
* assignment comes from binaryAssignment(ctx, 'make-pdf'), the same resolver
|
||||
* the runtime prelude puts into every later block that uses $P on env-var
|
||||
* hosts, so readiness and use always name the same binary.
|
||||
*
|
||||
* $P = make-pdf/dist/pdf.
|
||||
*
|
||||
* Resolution order:
|
||||
* 1. Local skill root: $_ROOT/{localSkillRoot}/make-pdf/dist/pdf
|
||||
* 2. Global: ~/{globalRoot}/make-pdf/dist/pdf
|
||||
* 3. Env override (MAKE_PDF_BIN) — for contributor dev builds
|
||||
* Resolution order: MAKE_PDF_BIN (contributor dev builds), then the
|
||||
* repo-local install, then the global install (env-var hosts: $GSTACK_ROOT,
|
||||
* which the prelude resolves repo-local first).
|
||||
*/
|
||||
export function generateMakePdfSetup(ctx: TemplateContext): string {
|
||||
return `## MAKE-PDF SETUP (run this check BEFORE any make-pdf command)
|
||||
|
||||
\`\`\`bash
|
||||
_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
|
||||
P=""
|
||||
[ -n "$MAKE_PDF_BIN" ] && [ -x "$MAKE_PDF_BIN" ] && P="$MAKE_PDF_BIN"
|
||||
[ -z "$P" ] && [ -n "$_ROOT" ] && [ -x "$_ROOT/${ctx.paths.localSkillRoot}/make-pdf/dist/pdf" ] && P="$_ROOT/${ctx.paths.localSkillRoot}/make-pdf/dist/pdf"
|
||||
[ -z "$P" ] && P="${toShellPath(ctx.paths.skillRoot)}/make-pdf/dist/pdf"
|
||||
${binaryAssignment(ctx, 'make-pdf')}
|
||||
if [ -x "$P" ]; then
|
||||
echo "MAKE_PDF_READY: $P"
|
||||
alias _p_="$P" # shellcheck alias helper (not exported)
|
||||
export P # available as $P in subsequent blocks within the same skill invocation
|
||||
export P # env-var hosts re-derive $P in every later block (runtime prelude)
|
||||
else
|
||||
echo "MAKE_PDF_NOT_AVAILABLE (run './setup' in the gstack repo to build it)"
|
||||
fi
|
||||
@@ -36,12 +33,12 @@ If \`MAKE_PDF_READY\` is printed: \`$P\` is the binary path for the rest of
|
||||
the skill. Use \`$P\` (not an explicit path) so the skill body stays portable.
|
||||
|
||||
Core commands:
|
||||
- \`$P generate <input.md> [output.pdf]\` — render markdown to PDF (80% use case)
|
||||
- \`$P generate --cover --toc essay.md out.pdf\` — full publication layout
|
||||
- \`$P generate --watermark DRAFT memo.md draft.pdf\` — diagonal DRAFT watermark
|
||||
- \`$P preview <input.md>\` — render HTML and open in browser (fast iteration)
|
||||
- \`$P setup\` — verify the browser (Aside, or gstack's own headless fallback) + pdftotext and run a smoke test
|
||||
- \`$P --help\` — full flag reference
|
||||
- \`"$P" generate <input.md> [output.pdf]\` — render markdown to PDF (80% use case)
|
||||
- \`"$P" generate --cover --toc essay.md out.pdf\` — full publication layout
|
||||
- \`"$P" generate --watermark DRAFT memo.md draft.pdf\` — diagonal DRAFT watermark
|
||||
- \`"$P" preview <input.md>\` — render HTML and open in browser (fast iteration)
|
||||
- \`"$P" setup\` — verify the browser (Aside, or gstack's own headless fallback) + pdftotext and run a smoke test
|
||||
- \`"$P" --help\` — full flag reference
|
||||
|
||||
Output contract:
|
||||
- \`stdout\`: ONLY the output path on success. One line.
|
||||
|
||||
@@ -176,7 +176,7 @@ ${outsideVoiceInvocation(ctx, { timeoutMs: 540000, nativeAlreadyRequired: true,
|
||||
${outsideVoiceFor(ctx).id === 'codex' ? 'The Codex backend uses `codex review --base` without a positional prompt: those arguments are mutually exclusive. Never drop --base to resolve an argv error; prompt-only review changes the diff scope.' : 'The Claude Code backend receives the parent-captured base diff, including committed and working-tree changes, because review mode cannot execute git.'}
|
||||
|
||||
Present output under \`${outsideVoiceFor(ctx).label.toUpperCase()} SAYS (code review):\` inside a \`tool-output\` fence.
|
||||
Only a completed response with severity tags or an explicit no-findings conclusion establishes the gate. P0/P1 findings (\`[P0]\`/\`[P1]\` or native \`P0:\`/\`P1:\` labels; \`VERDICT: findings\`) → GATE: FAIL. Completed without P0/P1 → GATE: PASS. Refusal, failure, missing markers or \`OUTSIDE_STATUS: unverified\` → GATE: MISSING COVERAGE; preserve the existing user decision flow.
|
||||
Only a completed response with severity tags or an explicit no-findings conclusion establishes the gate. P0/P1 findings (\`[P0]\`/\`[P1]\` or native \`P0:\`/\`P1:\` labels; \`VERDICT: findings\`) → GATE: FAIL. Completed without P0/P1 → GATE: PASS. Refusal, failure, missing markers or \`OUTSIDE_STATUS: unverified\` → GATE: MISSING COVERAGE; no fix question.
|
||||
|
||||
If GATE is FAIL, use AskUserQuestion:
|
||||
\`\`\`
|
||||
@@ -409,7 +409,7 @@ THE PLAN:
|
||||
function codexPlanReviewRun(ctx: TemplateContext, ceo: boolean, needsApprovalReadiness: boolean): string {
|
||||
return `**If \`CODEX_MODE: ready\` (or \`unverified\`) — run ${outsideVoiceFor(ctx).label}:**
|
||||
|
||||
${['plan-ceo-review', 'plan-eng-review'].includes(ctx.skillName) ? `Run this block only for \`ready\`, in the one foreground Bash call described below.
|
||||
${['plan-ceo-review', 'plan-eng-review'].includes(ctx.skillName) ? `Run this block for \`ready\` or \`unverified\`, in the one foreground Bash call described below.
|
||||
Its opening harness guard rechecks the fresh shell: exit 78 uses the same Native
|
||||
fallback below, never a replacement provider. Finish termination before fallback and consume only
|
||||
completed output. Use private temporary paths, with no background jobs.` : `Run the selected backend in the one foreground Bash invocation described below.
|
||||
|
||||
@@ -467,13 +467,13 @@ The plan completion results augment the existing Scope Drift Detection. If a pla
|
||||
|
||||
- **NOT DONE items** become additional evidence for **MISSING REQUIREMENTS** in the scope drift report.
|
||||
- **Items in the diff that don't match any plan item** become evidence for **SCOPE CREEP** detection.
|
||||
- **HIGH-impact plan-file discrepancies** trigger AskUserQuestion:
|
||||
- Show the investigation findings
|
||||
- **HIGH-impact plan-file discrepancies** trigger AskUserQuestion showing the investigation findings:
|
||||
- Options: A) Stop this review for implementation, B) Continue this review with P1 TODOs, C) Record the items as intentionally dropped
|
||||
- A ends this invocation before code review or implementation. List the missing work; after implementation, start a fresh /review.
|
||||
- B queues the approved TODO changes for Step 5, not this read-only audit. B/C continue to the final Scope Check and Step 2. None of these choices authorizes shipping or waives required verification.
|
||||
- Spawned or non-interactive: report REQUIREMENTS MISSING, continue; record nothing.
|
||||
|
||||
This is **INFORMATIONAL** unless HIGH-impact plan-file discrepancies are found (then it gates via AskUserQuestion).
|
||||
Otherwise the audit is **INFORMATIONAL**.
|
||||
Discrepancies derived only from fallback sources (commit messages, TODOS.md, PR description) never trigger
|
||||
this question, whatever their IMPACT: report them in the Scope Check as lower-confidence missing requirements.
|
||||
|
||||
|
||||
@@ -97,9 +97,6 @@ Never reset DEADLINE_FILE/bypass DEADLINE_TOOL. Expiry or an invalid/missing DEA
|
||||
|
||||
Each probe is one native command/interaction plus checks, excluding bookkeeping.
|
||||
Never batch probes.
|
||||
How one probe works: (1) if bounded, check the clock and choose the next command;
|
||||
(2) publish a checkpoint naming it (functional: \`capture --after\` publishes it);
|
||||
(3) run exactly that command via DEADLINE_TOOL or EVIDENCE_TOOL, read its result, repeat.
|
||||
|
||||
1. First demonstrate success: output AND durable effects. Guard if bounded; await completion.
|
||||
2. **Decide whether another probe is needed.** If bounded, run \`bun DEADLINE_TOOL status DEADLINE_FILE\`.
|
||||
|
||||
@@ -2,8 +2,8 @@
|
||||
* Runtime root for env-var hosts (C1, #1159): the one owner of how generated
|
||||
* bash finds gstack on Codex and every other `usesEnvVars` host.
|
||||
*
|
||||
* Those hosts run each fenced block in a fresh shell, so `GSTACK_ROOT`, `B`
|
||||
* and `D` set by an earlier block are gone. The preamble's root resolution and
|
||||
* Those hosts run each fenced block in a fresh shell, so `GSTACK_ROOT`, `B`,
|
||||
* `D` and `P` set by an earlier block are gone. The preamble's root resolution and
|
||||
* the per-fence prelude come from `runtimeRootPrelude()`; `insertRuntimePreludes()`
|
||||
* is the single post-render pass (gen-skill-docs, before rewriteInstallRoot)
|
||||
* that puts it into every fence that uses these variables without assigning
|
||||
@@ -19,6 +19,13 @@ import { toShellPath, type TemplateContext } from './types';
|
||||
|
||||
/** Per-fence byte budget for the prelude (T-ENG1), asserted by the INV-3 test. */
|
||||
export const PRELUDE_BYTE_BUDGET = 400;
|
||||
/**
|
||||
* make-pdf's dev-build override (E2), appended to `P`'s assignment. It is the
|
||||
* one fixed segment outside PRELUDE_BYTE_BUDGET: root plus `P=...` fits the
|
||||
* budget like `B`/`D` do, and the override is needed in every block that runs
|
||||
* `$P` so readiness and use name the same binary.
|
||||
*/
|
||||
export const MAKE_PDF_OVERRIDE = ';[ -x "${MAKE_PDF_BIN:-}" ]&&P=$MAKE_PDF_BIN||:';
|
||||
|
||||
const ROOT_VARS = /\$\{?GSTACK_(?:ROOT|BIN|BROWSE|DESIGN|MAKE_PDF)\b/;
|
||||
const assigns = (name: string) => new RegExp(`(?:^|[\\s;&|(])${name}=`, 'm');
|
||||
@@ -48,13 +55,23 @@ export function runtimeRootPrelude(ctx: TemplateContext): string {
|
||||
}
|
||||
|
||||
/**
|
||||
* Lines that assign `B` (browse) or `D` (design). BROWSE SETUP, DESIGN SETUP and
|
||||
* the prelude all use this, so a later fence runs the binary setup checked.
|
||||
* Env-var hosts derive from `GSTACK_ROOT`, which the prelude already points at
|
||||
* the repo-local install when one exists; other hosts keep their
|
||||
* repo-local-first probe.
|
||||
* Lines that assign `B` (browse), `D` (design) or `P` (make-pdf). BROWSE SETUP,
|
||||
* DESIGN SETUP, MAKE-PDF SETUP and the prelude all use this, so a later fence
|
||||
* runs the binary setup checked. Env-var hosts derive from `GSTACK_ROOT`, which
|
||||
* the prelude already points at the repo-local install when one exists; other
|
||||
* hosts keep their repo-local-first probe. make-pdf's binary is
|
||||
* `make-pdf/dist/pdf`, and `MAKE_PDF_BIN` (a contributor dev build) wins on
|
||||
* every host.
|
||||
*/
|
||||
export function binaryAssignment(ctx: TemplateContext, tool: 'browse' | 'design'): string {
|
||||
export function binaryAssignment(ctx: TemplateContext, tool: 'browse' | 'design' | 'make-pdf'): string {
|
||||
if (tool === 'make-pdf') {
|
||||
if (getHostConfig(ctx.host).usesEnvVars) return `P=$GSTACK_ROOT/make-pdf/dist/pdf${MAKE_PDF_OVERRIDE}`;
|
||||
return `_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
|
||||
P=""
|
||||
[ -n "$MAKE_PDF_BIN" ] && [ -x "$MAKE_PDF_BIN" ] && P="$MAKE_PDF_BIN"
|
||||
[ -z "$P" ] && [ -n "$_ROOT" ] && [ -x "$_ROOT/${ctx.paths.localSkillRoot}/make-pdf/dist/pdf" ] && P="$_ROOT/${ctx.paths.localSkillRoot}/make-pdf/dist/pdf"
|
||||
[ -z "$P" ] && P="${toShellPath(ctx.paths.skillRoot)}/make-pdf/dist/pdf"`;
|
||||
}
|
||||
const v = tool === 'browse' ? 'B' : 'D';
|
||||
const dir = tool === 'browse' ? ctx.paths.browseDir : ctx.paths.designDir;
|
||||
if (getHostConfig(ctx.host).usesEnvVars) return `${v}=$GSTACK_ROOT/${tool}/dist/${tool}`;
|
||||
@@ -69,8 +86,9 @@ export function fencePrelude(ctx: TemplateContext, body: string): string {
|
||||
if (!getHostConfig(ctx.host).usesEnvVars) return '';
|
||||
const needB = uses('B').test(body) && !assigns('B').test(body);
|
||||
const needD = uses('D').test(body) && !assigns('D').test(body);
|
||||
const needP = uses('P').test(body) && !assigns('P').test(body);
|
||||
if (assigns('GSTACK_ROOT').test(body)) return '';
|
||||
const needRoot = needB || needD || ROOT_VARS.test(body);
|
||||
const needRoot = needB || needD || needP || ROOT_VARS.test(body);
|
||||
const derived = Object.entries(DERIVED)
|
||||
.filter(([name]) => uses(`GSTACK_${name}`).test(body) && !assigns(`GSTACK_${name}`).test(body))
|
||||
.map(([name, dir]) => `GSTACK_${name}=$GSTACK_ROOT/${dir}`);
|
||||
@@ -79,6 +97,7 @@ export function fencePrelude(ctx: TemplateContext, body: string): string {
|
||||
derived.join(' '),
|
||||
needB ? binaryAssignment(ctx, 'browse') : '',
|
||||
needD ? binaryAssignment(ctx, 'design') : '',
|
||||
needP ? binaryAssignment(ctx, 'make-pdf') : '',
|
||||
].filter(Boolean).join('\n');
|
||||
}
|
||||
|
||||
|
||||
@@ -47,37 +47,54 @@ all preceding valid round files in order (omit them for round 1):
|
||||
~/.claude/skills/gstack/bin/gstack-office-hours-review prepare --design "<design-path>" --out-dir "<review-directory>" "<round-1.json if present>" "<round-2.json if present>"
|
||||
\`\`\`
|
||||
|
||||
Omit absent arguments rather than passing placeholders. The helper chooses the next
|
||||
Omit absent arguments rather than passing placeholders. Users get 3 rounds; only
|
||||
when a caller sets a lower review round limit, add \`--max-rounds <N>\` with that
|
||||
same value to every prepare, check, and finalize command. The helper chooses the next
|
||||
round and writes \`round-N.prompt.md\`. It includes the full finding schema, all five
|
||||
review dimensions (Completeness, Consistency, Clarity, Scope, Feasibility), the
|
||||
office-hours coaching contract, and the COMPLETE preceding JSON verdict.
|
||||
office-hours coaching contract, and the COMPLETE preceding JSON verdict. It also
|
||||
saves the design as reviewed (\`round-N.design.md\`). Round 1 is a full review;
|
||||
rounds 2 and 3 are delta re-reviews of the exact design diff since the last round,
|
||||
so prepare every round in the same review directory and do not edit the design
|
||||
between prepare and its review.
|
||||
|
||||
Use the Agent tool with \`run_in_background: false\` and its returned \`dispatch\`
|
||||
string unchanged as the prompt. The reviewer must Read the entire prepared prompt
|
||||
file before reviewing the design. Do not recreate the prompt, copy selected fields,
|
||||
or summarize prior findings. A parent Read does not deliver the file to the reviewer.
|
||||
The reviewer has fresh context and cannot see the brainstorming conversation.
|
||||
Its prepared contract requires a complete JSON Write and an identical JSON response.
|
||||
Its prepared contract requires a complete JSON Write, sealed by the dispatch's
|
||||
\`Seal:\` command, and a one-line \`OFFICE_HOURS_VERDICT\` receipt as its entire response.
|
||||
It protects the required coaching and Assignment sections, distinguishes unknown
|
||||
customer facts from committed behavior, and requires evidence for every prior status.
|
||||
|
||||
**Step 2: Check stop conditions, then fix and re-dispatch**
|
||||
|
||||
After each verdict, BEFORE fixing any findings or dispatching again, validate the
|
||||
saved files with the helper (list every completed round in order):
|
||||
saved files with the helper. Pass the reviewer's entire response unchanged as the
|
||||
receipt and list every completed round in order:
|
||||
|
||||
\`\`\`bash
|
||||
~/.claude/skills/gstack/bin/gstack-office-hours-review check "<round-1.json>" "<round-2.json if present>" "<round-3.json if present>"
|
||||
~/.claude/skills/gstack/bin/gstack-office-hours-review check --receipt "<reviewer response>" "<round-1.json>" "<round-2.json if present>" "<round-3.json if present>"
|
||||
\`\`\`
|
||||
|
||||
A missing, malformed, or mismatched receipt fails the check: that attempt is a failed review.
|
||||
|
||||
Omit absent arguments rather than passing placeholders.
|
||||
**Convergence guard and stopping rules:** Read its stop reason:
|
||||
- PASS: no unresolved findings; proceed to Step 3.
|
||||
- CONVERGENCE: the reviewer explicitly marked a prior obligation persisting with
|
||||
a concrete prior/current finding pair and document evidence. Stop even if new
|
||||
findings appear. Shared topic labels or new refinements alone are insufficient.
|
||||
- MAX_ITERATIONS: round 3 completed; stop.
|
||||
- CONTINUE: fix the listed findings in the design, then return to Step 1 to prepare and dispatch the next review.
|
||||
**Convergence guard and stopping rules:** Each finding is blocking or minor. Only
|
||||
blocking findings require another round. Read its stop reason:
|
||||
- PASS: no blocking findings remain. Any minor findings are recorded, not fixed,
|
||||
and never justify another round; proceed to Step 3.
|
||||
- CONVERGENCE: the reviewer explicitly marked a blocking prior obligation
|
||||
persisting with a concrete prior/current finding pair and document evidence.
|
||||
Stop even if new findings appear. Shared topic labels or new refinements alone
|
||||
are insufficient.
|
||||
- MAX_ITERATIONS: round 3 completed (or the caller's lower round limit); stop.
|
||||
- CONTINUE: fix only the blocking findings in the design, then return to Step 1
|
||||
to prepare and dispatch the next review. Do not edit for minor findings
|
||||
mid-loop: they stay recorded for the user, and new text only gives the next
|
||||
diff more to review. Its reviewer checks every prior finding and raises new
|
||||
blocking findings only for problems your changes introduced or exposed.
|
||||
|
||||
On a stop, do not fix again or re-dispatch. Run the finalizer before approval:
|
||||
|
||||
@@ -119,7 +136,7 @@ mkdir -p "$GSTACK_STATE_ROOT/analytics"
|
||||
echo '{"skill":"office-hours","ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","iterations":ITERATIONS,"issues_found":FOUND,"issues_fixed":FIXED,"remaining":REMAINING,"quality_score":SCORE}' >> "$GSTACK_STATE_ROOT/analytics/spec-review.jsonl" 2>/dev/null || true
|
||||
\`\`\`
|
||||
Use iterations, issues_found, issues_fixed, remaining, and quality_score from the
|
||||
helper. FOUND counts finding observations across rounds; FIXED counts only
|
||||
helper; its remaining_blocking and remaining_minor split the remaining count. FOUND counts finding observations across rounds; FIXED counts only
|
||||
reviewer-confirmed resolutions. An unavailable score is null, never invented.`;
|
||||
}
|
||||
|
||||
|
||||
@@ -60,13 +60,13 @@ backslashes serialize cleanly — never use hand-rolled \`echo\` / \`printf\`.
|
||||
|
||||
\`\`\`bash
|
||||
GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "\${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
|
||||
BRANCH=$(~/.claude/skills/gstack/bin/gstack-slug --get BRANCH 2>/dev/null)
|
||||
SLUG=$(~/.claude/skills/gstack/bin/gstack-slug --get SLUG 2>/dev/null)
|
||||
TASKS_DIR="$GSTACK_STATE_ROOT/projects/\${SLUG:-unknown}"
|
||||
mkdir -p "$TASKS_DIR"
|
||||
TASKS_FILE="$TASKS_DIR/tasks-${phase}-$(date +%Y%m%d-%H%M%S).jsonl"
|
||||
COMMIT=$(git rev-parse HEAD 2>/dev/null || echo unknown)
|
||||
BRANCH=$(git branch --show-current 2>/dev/null || echo unknown)
|
||||
# The task record keeps the raw git branch; artifact filenames keep gstack-slug's sanitized BRANCH.
|
||||
GIT_BRANCH=$(git branch --show-current 2>/dev/null || echo unknown)
|
||||
RUN_ID="$(date -u +%Y%m%dT%H%M%SZ)-$$"
|
||||
|
||||
# Repeat ONE jq invocation per task identified during this review.
|
||||
@@ -77,7 +77,7 @@ RUN_ID="$(date -u +%Y%m%dT%H%M%SZ)-$$"
|
||||
jq -nc \\
|
||||
--arg phase '${phase}' \\
|
||||
--arg run_id "$RUN_ID" \\
|
||||
--arg branch "$BRANCH" \\
|
||||
--arg branch "$GIT_BRANCH" \\
|
||||
--arg commit "$COMMIT" \\
|
||||
--arg id "$TASK_ID" \\
|
||||
--arg priority "$PRIORITY" \\
|
||||
|
||||
@@ -129,7 +129,7 @@ function cardRules(mode: TestValueBarMode, skillName: string): string {
|
||||
qa: skillName === 'qa-only' ? 'Put it under each proposed test.' : 'Put it in the 8e.5 record.',
|
||||
audit: 'Read cards from test header comments when present.',
|
||||
}[mode];
|
||||
return `Value card: \`Value: protects=<...>; fails_when=<...>; why_new=<...>; seam=none\` (seam: \`none\` or its name); each field at most ${CARD_FIELD_MAX_BYTES} UTF-8 bytes here (clamp to 157 plus \`...\`; JSON keeps full values). ${where} A missing upstream card never blocks: derive it; ignore unknown fields.
|
||||
return `Value card: \`Value: protects=<...>; fails_when=<...>; why_new=<...>; seam=none\` (seam: \`none\` or its name); each field at most ${CARD_FIELD_MAX_BYTES} UTF-8 bytes here (clamp to 157 plus \`...\`; written JSON keeps full values). ${where} A missing upstream card never blocks: derive it; ignore unknown fields.
|
||||
|
||||
Example: ${EXAMPLE_CARD}
|
||||
${EXAMPLE_REJECTED}`;
|
||||
|
||||
@@ -676,6 +676,8 @@ export interface FreeRunReport {
|
||||
inFlight: string[];
|
||||
/** Planned files never observed in the stream (silent passers + never-flushed wedges). */
|
||||
filesWithNoOutput: number;
|
||||
/** The last nonblank line either stream printed: a native abort's message when bun dies without a summary. */
|
||||
lastOutputLine: string | null;
|
||||
}
|
||||
|
||||
interface FileProgress {
|
||||
@@ -711,6 +713,7 @@ export class FreeRunReporter {
|
||||
private testsRan: number | null = null;
|
||||
private filesRan: number | null = null;
|
||||
private sawSummary = false;
|
||||
private lastOutputLine: string | null = null;
|
||||
|
||||
constructor(
|
||||
private readonly plannedFiles: string[],
|
||||
@@ -758,6 +761,7 @@ export class FreeRunReporter {
|
||||
unhandledErrors: [...this.unhandled],
|
||||
inFlight,
|
||||
filesWithNoOutput: this.plannedFiles.filter((f) => !this.progress.has(normalizeRelativePath(f))).length,
|
||||
lastOutputLine: this.lastOutputLine,
|
||||
};
|
||||
}
|
||||
|
||||
@@ -768,6 +772,7 @@ export class FreeRunReporter {
|
||||
// (fail) lines land under a second phantom file (observed on the first
|
||||
// Linux run: 5 real failures reported as 10 across 2 files).
|
||||
const line = stripAnsiLine(rawLine).replace(/^::group::/, '');
|
||||
if (line.trim() !== '') this.lastOutputLine = line.trim();
|
||||
let visible = false;
|
||||
const failedCount = this.failureSummary.consume(line, origin);
|
||||
if (failedCount !== null) {
|
||||
@@ -903,6 +908,10 @@ export function buildRunEpilogue(
|
||||
for (const u of report.unhandledErrors) {
|
||||
lines.push(` ⚠ unhandled error between tests (around ${u.file ?? 'unknown file'})`);
|
||||
}
|
||||
if (status === 'failed' && !report.sawTerminalSummary) {
|
||||
lines.push(` ⚠ test process ended before its summary; in flight: ${report.inFlight.length > 0 ? report.inFlight.join(', ') : 'unknown'}`);
|
||||
if (report.lastOutputLine) lines.push(` ⚠ last output: ${report.lastOutputLine}`);
|
||||
}
|
||||
if (status === 'timed-out') {
|
||||
if (report.inFlight.length > 0) {
|
||||
lines.push(` ⏱ in flight at kill: ${report.inFlight.join(', ')}`);
|
||||
|
||||
+25
-16
@@ -66,13 +66,16 @@ import { EVAL_POLICY, PERIODIC_CI_EXCLUDE } from '../test/helpers/periodic-exclu
|
||||
import {
|
||||
getProjectEvalDir, getClaudeCliVersion, isFinalizedEvalResultFile, failureClassOf, panelVerdict,
|
||||
sanitizeTrialError, CONTRACT_VIOLATIONS_FILE, TRIAL_ENV, type TrialFailureClass,
|
||||
trialCostKnown, trialFailureFields, trialSessions,
|
||||
} from '../test/helpers/eval-store';
|
||||
import { STALL_WINDOW_MS } from '../test/helpers/eval-budgets';
|
||||
import { readSessionLedger, type SessionLedgerRow } from '../test/helpers/session-ledger';
|
||||
import { preflightAnthropicApi } from '../test/helpers/anthropic-preflight';
|
||||
import { e2eReuseLaneProblem, prepareE2EShardReuse, selectPlanReceipts } from './e2e-shard-reuse';
|
||||
import { E2E_TOUCHFILES, E2E_TIERS } from '../test/helpers/touchfiles';
|
||||
import { scopeCodexAccess, shardFile, shardCaseId, shardTrial, trialShardKey, type CaseTrialPlan, caseTrialPlan, excludedCasesNamePattern, caseTestNamePattern, expandCaseShards, expandTrialShards, fileCaseRegistration, partitionCaseExclusions } from './lib/paid-cases';
|
||||
import { retriesForFiles, trialPanelKey, sliceExecutionOrder, buildRunManifest, parseRunManifest, type SliceResult, sliceExitCode, guardTrialRecords, formatSlicePlan, formatCapacityPreflight, sliceDeadlineMs } from './lib/paid-plan';
|
||||
import { caseFile, runCaseDiagnosis, formatPanelLine, runPaidReport } from './lib/paid-report';
|
||||
import { caseSelection, runCaseDiagnosis, formatPanelLine, runPaidReport } from './lib/paid-report';
|
||||
import {
|
||||
DEFAULT_JOBS, DEFAULT_MAX_FILES_PER_SHARD, DEFAULT_SHARD_TIMEOUT_MS, DEFAULT_TIER, DEFAULT_WITHIN_SHARD_CONCURRENCY, OVERLAY_MAX_ACTIVE_SHARDS,
|
||||
PAID_LANE_POLICY, PAID_TIERS, ROOT, SLICE_JOB_STARTED_AT_ENV, SLICE_UPLOAD_RESERVE_MS, isOverlayTestFile,
|
||||
@@ -94,8 +97,10 @@ export * from './lib/paid-cases';
|
||||
export * from './lib/paid-plan';
|
||||
export * from './lib/paid-report';
|
||||
|
||||
/** Records and contract evidence an isolated shard left in its eval dir. */
|
||||
export function readTrialEvidence(evalDir: string | undefined): { records: any[]; contract: string | null } {
|
||||
type TrialEvidence = { records: any[]; contract: string | null; sessions?: SessionLedgerRow[] };
|
||||
|
||||
/** Records, contract evidence and session-ledger rows an isolated shard left in its eval dir. */
|
||||
export function readTrialEvidence(evalDir: string | undefined): TrialEvidence {
|
||||
if (!evalDir || !fs.existsSync(evalDir)) return { records: [], contract: null };
|
||||
const names = fs.readdirSync(evalDir);
|
||||
const parse = (name: string) => { try { return JSON.parse(fs.readFileSync(path.join(evalDir, name), 'utf8')); } catch { return null; } };
|
||||
@@ -108,28 +113,32 @@ export function readTrialEvidence(evalDir: string | undefined): { records: any[]
|
||||
const line = fs.readFileSync(path.join(evalDir, CONTRACT_VIOLATIONS_FILE), 'utf8').split('\n').find(l => l.trim());
|
||||
if (line) contract = String(JSON.parse(line).message ?? 'contract violation');
|
||||
} catch { /* no sidecar */ }
|
||||
return { records, contract };
|
||||
return { records, contract, sessions: readSessionLedger(evalDir) };
|
||||
}
|
||||
|
||||
/** Classify one isolated trial shard. Contract evidence always fails the trial. */
|
||||
export function classifyTrialShard(
|
||||
outcome: Pick<ShardOutcome, 'status' | 'executedTests' | 'skippedTests' | 'elapsedMs' | 'runnerError'>,
|
||||
caseId: string, trial: number, plan: CaseTrialPlan,
|
||||
evidence: { records: any[]; contract: string | null },
|
||||
evidence: TrialEvidence,
|
||||
): ShardTrialRecord {
|
||||
const failedRecord = evidence.records.find(record => record.passed === false) ?? evidence.records[0];
|
||||
const base: ShardTrialRecord = {
|
||||
case: caseId, trial, kind: plan.kind, panel: plan.panel, quarantined: plan.quarantined, outcome: null,
|
||||
cost_usd: Math.round(evidence.records.reduce((sum, record) => sum + (Number(record.cost_usd) || 0), 0) * 100) / 100,
|
||||
duration_ms: outcome.elapsedMs,
|
||||
...(typeof failedRecord?.model === 'string' ? { model: failedRecord.model } : {}),
|
||||
...(typeof failedRecord?.model === 'string' ? { model: failedRecord.model } : {}), ...trialSessions(evidence.sessions ?? []),
|
||||
...trialCostKnown(evidence.records, evidence.sessions),
|
||||
};
|
||||
const failed = (failureClass: TrialFailureClass, error?: string): ShardTrialRecord => {
|
||||
const cls = evidence.contract !== null ? 'contract' : failureClass;
|
||||
const raw = evidence.contract ?? failedRecord?.error ?? error;
|
||||
return { ...base, outcome: 'failed', failure_class: cls,
|
||||
...(failedRecord?.exit_reason ? { exit_reason: String(failedRecord.exit_reason) } : {}),
|
||||
...(Number.isInteger(failedRecord?.timeout_at_turn) ? { timeout_at_turn: failedRecord.timeout_at_turn } : {}),
|
||||
...(sanitizeTrialError(raw) ? { error: sanitizeTrialError(raw) } : {}),
|
||||
...trialFailureFields({ failure_class: cls, exit_reason: failedRecord?.exit_reason, error: raw, sessions: evidence.sessions, record: failedRecord }, STALL_WINDOW_MS) };
|
||||
};
|
||||
const failed = (failureClass: TrialFailureClass, error?: string): ShardTrialRecord => ({
|
||||
...base, outcome: 'failed', failure_class: evidence.contract !== null ? 'contract' : failureClass,
|
||||
...(failedRecord?.exit_reason ? { exit_reason: String(failedRecord.exit_reason) } : {}),
|
||||
...(Number.isInteger(failedRecord?.timeout_at_turn) ? { timeout_at_turn: failedRecord.timeout_at_turn } : {}),
|
||||
...(sanitizeTrialError(evidence.contract ?? failedRecord?.error ?? error) ? { error: sanitizeTrialError(evidence.contract ?? failedRecord?.error ?? error) } : {}),
|
||||
});
|
||||
if (outcome.runnerError !== undefined) return { ...base, harness: `runner error: ${sanitizeTrialError(outcome.runnerError) ?? 'unknown'}` };
|
||||
if (outcome.status === 'never-started') return { ...base, harness: 'never started' };
|
||||
if (outcome.status === 'passed-empty') return { ...base, harness: 'hollow: executed no case' };
|
||||
@@ -873,16 +882,16 @@ async function main(): Promise<number> {
|
||||
if (options.reportDir) return runPaidReport(options.reportDir, { writeDurations: options.writeDurations });
|
||||
|
||||
if (options.caseId && options.listOnly) {
|
||||
const file = caseFile(options.caseId);
|
||||
const { file, mode, reason } = caseSelection(options.caseId);
|
||||
const plan = caseTrialPlan(options.caseId);
|
||||
const n = options.trials ?? plan.panel.n;
|
||||
console.log(`[test:paid] --case ${options.caseId}: ${n} trial(s) of ${file} (kind ${plan.kind}), list only`);
|
||||
for (let trial = 1; trial <= n; trial++) console.log(` ${trialShardKey(file, options.caseId, trial)}`);
|
||||
console.log(`[test:paid] --case ${options.caseId}: ${n} trial(s) of ${file} (kind ${plan.kind}), selects ${reason}, list only`);
|
||||
for (let trial = 1; trial <= n; trial++) console.log(` ${mode === 'file' ? `${file} (trial ${trial})` : trialShardKey(file, options.caseId, trial)}`);
|
||||
return 0;
|
||||
}
|
||||
if (options.caseId) {
|
||||
preflightAnthropicApi(process.env);
|
||||
const verdict = await runCaseDiagnosis(options.caseId, { trials: options.trials ?? undefined, jobs: options.jobs,
|
||||
const verdict = await runCaseDiagnosis(options.caseId, { trials: options.trials ?? undefined, jobs: options.jobs, runShards: runPaidShards,
|
||||
withinShardConcurrency: options.withinShardConcurrency, timeoutMs: timeoutOverride,
|
||||
evalDirBase: process.env.GSTACK_EVAL_DIR || getProjectEvalDir() });
|
||||
return verdict.status === 'PASS' ? 0 : 1;
|
||||
|
||||
@@ -265,7 +265,6 @@
|
||||
"test/model-overlays.test.ts\tTS2322\tType 'string' is not assignable to type '\"claude\" | \"fable-5\" | \"gemini\" | \"gpt\" | \"gpt-5.4\" | \"gpt-5.6-sol\" | \"gpt-6-astra\" | \"o-series\" | \"opus-4-7\" | \"opus-4-8\" | \"sonnet-5\" | undefined'.": 4,
|
||||
"test/native-auto-decide-pty.test.ts\tTS2345\tArgument of type 'string | undefined' is not assignable to parameter of type 'string'. Type 'undefined' is not assignable to type 'string'.": 1,
|
||||
"test/native-auto-decide.test.ts\tTS2345\tArgument of type '{ status: string; calls: never[]; assistantMessages: { sessionId: string; text: string; timestamp: string; }[]; }' is not assignable to parameter of type 'PlanCountTranscript'. Types of property 'status' are incompatible. Type 'string' is not assignable to type '\"error\" | \"missing\" | \"ready\"'.": 1,
|
||||
"test/office-hours-completion.test.ts\tTS18046\t'reordered.findings' is of type 'unknown'.": 1,
|
||||
"test/office-hours-review.test.ts\tTS18048\t'check.stderr' is possibly 'undefined'.": 1,
|
||||
"test/office-hours-review.test.ts\tTS18048\t'check.stdout' is possibly 'undefined'.": 1,
|
||||
"test/office-hours-review.test.ts\tTS18048\t'direct.stderr' is possibly 'undefined'.": 1,
|
||||
|
||||
@@ -4,7 +4,7 @@
|
||||
# Runs as user `ubi` (passwordless sudo) from the synced checkout.
|
||||
set -euo pipefail
|
||||
|
||||
BUN_VERSION=1.4.0
|
||||
BUN_VERSION=1.4.2
|
||||
NODE_VERSION=22.20.0
|
||||
|
||||
export DEBIAN_FRONTEND=noninteractive
|
||||
|
||||
Reference in new issue
Block a user