Files
PaperClipAI/tests/runner-e2e/runner.spec.ts
T
DottaandPaperclip 492d537340 fix(native): recover controllers across webserver replacement (#18343)
## Thinking Path

> - Paperclip manages AI agents and their work.
> - Native execution runs in a remote Runner while the webserver owns
its controller.
> - Encrypted PRP checkpoints and session snapshots already survive a
webserver restart in Postgres.
> - Ordinary warm and per-turn admission still needs local markers and
supervisor metadata to find that state.
> - Replacing the webserver with an empty native controller directory
can therefore lose a usable retained session.
> - This change adds bounded recovery metadata to the existing snapshot
transaction and reconstructs local projections from the selected durable
source.
> - The benefit is native execution continuity across webserver
replacement, with the existing ownership and agent lifecycle policies.

## Linked Issues or Issue Description

**What existing behavior does this improve?**

Native remote controller recovery and retained-session continuation
during a webserver upgrade.

**Subsystem affected**

server/ — native runtime, checkpoint persistence, recovery scheduling
and internal upgrade reporting. Product E2E qualification also changes.

**Current behavior**

Without both the local PRP marker and supervisor file, normal
continuation can return before discovering the saved Postgres
checkpoint. The snapshot already exists in Postgres, but its local reuse
compatibility envelope cannot be reconstructed.

**Proposed behavior**

Select the existing durable source, validate its identity and ownership,
restore its PRP state and supervisor projection, then use normal
authenticated admission. Missing or inconsistent evidence holds
recovery. It does not authorize fresh provider work.

**Reason and benefit**

A replacement webserver must recover active execution and later continue
its retained session without the old server's native files. This is one
component of deployment continuity. HTTP traffic switching, workspace
storage migration and Railway rollout remain separate.

**Breaking changes**

No new public endpoint, schema migration or database pool. The private
run-profile envelope is additive. Disk-independent rollback also
requires a reader for the staged controller format and compact
checkpoint part; the first production release containing these readers
is the floor. Older file-backed sessions need an owner checkpoint or the
bounded internal importer.

**Additional context**

Related foundations: #17401, #17916, #17254, #14621, #14622. The
implementation contract and open acceptance gates are in [the dated
plan](doc/plans/2026-10-30-native-controller-checkpoints.md).

## What Changed

- Persist a versioned recovery envelope, at most 4 KiB, beside the
redacted session snapshot in its existing transaction. Keep it out of
public run projections.
- Compare frozen admission separately from prepared attachment
instructions during staged recovery, preserving exact
template/identity/authority checks.
- Bound the existing lazy checkpoint pool at three connections after
measured two-slot burst backlog; keep request/renewal capacities and SQL
operation counts unchanged. The tuning is a separate commit.
- Discover and restore the exact selected retained source without local
native markers. Preserve the 32-candidate window, progress barriers and
current authority checks.
- Fence snapshot and result writes. Acquire coordinator, issue and run
locks in order and check expiry after lock waits. Combine snapshot
ownership/read work into one statement.
- Reuse existing recovery deadlines and drain/resume scheduling.
Retained-owner waits preserve the original unstarted run and actual
lease deadline. Stop, ownership and provider-progress fences precede
deferral. Add bounded native evidence reporting, retained-source import
and retention protection.
- Admit a settled warm source with an authenticated idle goal-read
suffix while retaining active-work and mixed-authority holds.
- Permit lifecycle `turn.stop` checkpoints after accepted results under
existing ownership fences. Claim the existing successor coordinator
before retiring the old warm session, and stage shared attachments only
after verified retirement. Preserve omitted isolated-home defaults and
existing fresh skill-selection order during recovery without hiding real
configuration changes.
- Carry the exact durable six-field PRP binding during active recovery
after a healthy warm handoff. Keep the Runner mismatch guard and
preserve opaque attachment IDs.
- Preserve an already verified active or idle warm process during remote
preparation; revalidate its proof/liveness and retain later
authenticated PRP attachment. Ordinary resumed leases retain suspension
checks.
- Drain retained source events under their original authority, reconcile
immutable source accounting and block unknown source tool dispatch
before target work.
- Persist compact staged handoff intent through authenticated adoption
and recover direct/listener command-loss windows without another Runner
or duplicate attachment.
- Rebuild suspended lifecycle receipts for resumed prepared tasks, and
retry missing lifecycle reads after partial PRP materialization without
overwriting contradictory evidence.
- Add normal warm/per-turn diskless continuation coverage, real Postgres
races and SQL-cost checks. Add a reproducible paired load workload.
- Give only the explicit outage fixture a 15-minute provider idle
window, with a red/green catalog regression preserving all production
timing and lifecycle policies.
- Publish the goal-less task-session pointer in the existing
post-acceptance terminal checkpoint transaction, with current ownership,
accepted-result, reset and newer-progress fences. Keep a private bound
receipt so late completion cannot duplicate or resurrect publication.
- Use a retained warm sandbox in the dedicated continuity fixture, and
accept an exact output token followed by an explanation. Keep the strict
identity oracle.
- Extend the explicit Daytona cell to a five-minute active outage and
another replacement after completion. Require retained provider
identity, a new tool receipt and a useful follow-up result. Add an owned
loopback Postgres E2E fixture for hosts that cannot start embedded
Postgres.

## Verification

**Merge recommendation — `fc68584ce6231a02e47ecd743f55b036ccea034f`:
ready for the qualified native-controller component.** Full CI, fresh
review, local application paths, controlled performance qualification
and the strict live Linux/Daytona case pass. Independent resource
cleanup is complete. Broader rollout, browser reconnection and the
unexecuted transport/platform combinations remain integration gates.
Integrated 484 passed full local recursive typecheck/build and ordinary
active→completion→cache-loss→C component qualification. Fresh review
then found a valid attached-wake failure: preparation appends attachment
instructions before the verifier compares execution to the frozen
admission. 9f4 fixes that by separately checking the pre-preparation
admission and exact prepared execution. Both hashes are captured
privately per admission; the full template, identity, configuration and
authority checks remain. Its regression fails before the fix; 36 focused
cases, 42 staged source/target cases and server typecheck pass
afterward. The review thread is resolved. Exact-head full CI passes
([run
41168228822](https://github.com/paperclipai/paperclip/actions/runs/41168228822):
48 successful jobs, one intended skip); the complete `pnpm test:run`
partitions are forced, with the five native suites in the Runner lane.
This is not a claim that an unsharded local command passed. The fresh
exact-head Greptile check succeeds with 5/5 and all six threads
resolved.

The original two-connection database comparison closes **steady PASS and
burst FAIL**. Every relative CPU/WAL/request-p99 gate passes and all
operations/ownership remain correct, but candidate arm 3 drains in
1,015.99 ms and baseline arm 4 in 1,059.67 ms against the unchanged
1,000 ms absolute gate. Those results remain failed. A separate 2→3→4→2
diagnostic reproduces drain times 1,051.57 / 956.79 / 882.02 / 1,065.72
ms: three is the smallest tested passing cap, and the repeated two-cap
failure survives run-order/cache effects.

Separately reviewable `fc68584ce6` changes the existing lazy checkpoint
pool default to three connections. Request capacity stays ten and
liveness two; idle release, shutdown fencing, transaction counts and
pool count are unchanged. Thirteen DB lifecycle/cap tests and DB
typecheck pass. Full recursive typecheck/build pass at exact fc68584,
with the matching server build stamp. The final paired native
feature-cost comparison explicitly uses the **same cap-three policy and
common current DB factory in both variants**; its baseline is
controlled, not the unmodified historical cap-two configuration.
Original gates and workload counts remain unchanged. The one strict
Linux/Daytona case started at 04:40:56 UTC only after both new profiles
passed; it passed on the first attempt with all 49 workflow checks.
Independent cleanup confirms all three owned sandboxes absent, owned
Postgres/controllers/observer gone, VM terminated, root volume absent
and ephemeral SSH key/agent removed.

**Final controlled cap-three qualification: PASS.** Baseline native
implementation `59da15066a` and candidate `fc68584ce6` use the same
current DB factory and checkpoint policy (request/liveness/checkpoint =
10/2/3). Eight 65-second ABBA arms cover 300 controllers and 10,000
historical tasks. All correctness, absolute and 10% relative gates pass;
original cap-two failures remain unchanged.

| Mode | DB CPU µs/op, baseline → candidate | WAL B/op, baseline →
candidate | DB-request p99 ms, baseline → candidate | Maximum drain |
| --- | --- | --- | --- | --- |
| Steady | 542.40 → 554.83 (+2.29%) | 1509.01 → 1558.27 (+3.26%) |
107.11 → 111.90 (+4.48%) | 26.35 ms |
| Burst | 1036.36 → 1063.11 (+2.58%) | 4153.12 → 4317.64 (+3.96%) |
103.736 → 103.841 (+0.10%) | 930.14 ms |

Every matched arm preserves operation/statement/transaction counts,
including 3,900 PRP and 3,900 snapshot transactions. All three actual
pools are included; there are zero lost owners, missed lease deadlines,
dropped operations or errors. Peak Postgres CPU is 18.31%/9.47% of two
cores and memory 614.29/684.42 MiB (steady/burst). The benchmark uses
one Node process simulating 300 controllers; it does not qualify
two-webserver overlap, HTTP latency or hosted capacity. Its observer SQL
contributes to measured database CPU. The original unmodified cap-two
baseline is separately retained as failed; this cap-three comparison is
explicitly controlled.

The exact final-source 8/512/8,192-chunk test also passes: zero recovery
SQL, transactions or changed rows in each text-only interval, zero idle
SQL, seven lifecycle transactions and twelve changed rows per sample.
Existing lease traffic remains separate. The final ordinary managed
application replay passes active recovery, completion, native-cache
removal and C continuation; its UNKNOWN local debug-daemon provenance
and partial instruction-observer limits remain recorded. The live case
uses the attested Linux producer instead.


**Live qualification: PASS, one attempt, no retries.** Campaign
`native-controller-retained-fix-fc68584ce6-ec2-20261031T044055Z` started
04:40:56 UTC on the exact final Linux webserver source; suite definition
hash `12f9bb97823c1537e43f0398e7cddf39890f77bce7b413ca4dfd944f4a30e5d5`.
The existing immutable image and attested Bookworm Runner artifact are
unchanged.

- A was SIGKILLed and its native cache discarded. B started after
300,149 ms, naturally acquired generation 2, authenticated at 316,624 ms
and resumed at 317,682 ms within the original 420,000 ms grace. The same
active run, Runner, provider PID/start fingerprint and one
submitted/started/accepted turn were preserved; one final
result/terminal and workspace finalization completed, with no
retry/recovery action.
- After active and queued completion, B was SIGKILLed and its native
cache discarded again. C received exactly one ordinary task-comment API
follow-up, completed a new tool and produced a useful reply while
preserving native/provider/Runner identities. This closes the earlier
failed post-C gate. A separate C OS-process PID field was not exported;
the physical-PID proof is specifically A→B.
- Four distinct results were accepted and published; all four
finalizations committed at attempt 1, all four accounting-v2 projections
settled, and the task-session pointer advanced to C. The coordinator
independently checked these facts and all 49 canonical checks against
the sanitized receipts.
- Canonical cleanup passed and independent checks confirmed three exact
sandbox 404s plus complete inventory/zero owned remnants, no owned PG or
controller/observer processes, EC2 terminated, root volume absent, and
the ephemeral SSH key/agent removed by 04:54:25 UTC.

Live scope is classic remote Codex warm continuity with non-native
storage dependencies available. Prepared-task completed B→C, full
per-turn application disk loss, browser reconnection, two-server load
and overall Railway/HTTP continuity were not live-qualified here.
Earlier failed campaigns and the original cap-two burst failure remain
failed historical evidence. A report-renderer cache setup failure was
preserved; the normal sanitizer subsequently succeeded with an explicit
cache path and a new output directory, without changing validators.


The repair keeps retained events under their original six-field
authority until persistence and accounting accept them. Unknown semantic
effects stay unresolved. A compact staged target survives source
bootstrap, cached-command delivery and authenticated handoff, including
an uncertain commit response. Direct and same-guest-listener recovery
require the verified original Runner and the exact endpoint. Target
snapshot publication marks readiness in memory, with retirement
coalesced into an existing command write. Headers contain only the
staged descriptor/digest; a prequeue template is written once as a
changed checkpoint part and then recovered from the existing
command/receipt. No second session snapshot transaction, streaming SQL
path, new pool or recurring idle scan is added.

| Current repair gate | Evidence and scope |
| --- | --- |
| Source-tail regression | Real Runner protocol RED reproduced
attachment rejection. Frozen core/transport pass five source-tail cases:
source usage attribution, held persistence/no premature ACK, lost commit
response plus cache loss, unknown semantic hold and accounting rejection
hold. |
| Handoff loss windows | Direct and listener modes pass
queued/prepared/activated/completed boundaries (eight cells), preserving
the actual Runner/provider process. Missing adoption proof and changed
endpoints reject recovery. |
| Compact storage | A 31 KiB attachment grant writes its separate part
once, leaves it unchanged across source snapshots and removes it at
queueing. Headers do not duplicate the grant. |
| Accounting | 21 real-Postgres usage tests and 19 source-binding tests
pass. Recovery hydrates immutable source receipts without resetting
their producer or billing classification; conflicting/incomplete usage
holds. |
| Admission and malformed evidence | Driver 16, staged evidence 42,
restart classifier 27 and source usage 19 focused cases pass. Four
malformed-file regressions pass at d6a7, whose executor is
byte-identical at 484, and server typecheck passes. Full executor file:
990 pass; 16 embedded-Postgres `initdb` setup failures before
assertions, retained separately. |
| Required checks | Exact fc68584 recursive typecheck and build pass
locally, with matching server stamp. Forced full CI succeeds (48 jobs
pass, one intended skip); all PR checks are clean. Maps/token gates pass
for their unchanged source surfaces. Fresh exact-head Greptile is 5/5,
all six threads resolved. |
| Ordinary application path | Exact fc68584 managed component passes
active recovery after natural expiry, completion, native-cache loss and
ordinary C continuation. Production bytes match before/after; temporary
fixtures restored and owned DB/services cleaned. UNKNOWN local
debug-daemon/partial instruction-observer limits remain; the separate
live case uses the attested Linux producer. |
| Producer | Locked Bookworm build from exact 5b7 Rust tree
d03e494217, equal to final fc68584 Rust
source. SHA-256
65b8f8cfdd6099e9ad24f1ca154e8a260777269cc82f6e23c6e6a180c531eaaf;
38,515,328 bytes. Actual Debian 12 execution and live admission pass.
Controller/admission/DB changes do not relabel the producer source. |
| Historical cap-two comparison | Exact 59da baseline vs 5b7 candidate,
300 controllers/10,000 history, ABBA 4×65 s per mode, all three actual
pools and real snapshots. Steady CPU/op 541.36→554.64 µs (+2.45%);
WAL/op 1512.13→1560.99 B (+3.23%); DB-request p99 102.78→105.99 ms
(+3.12%): PASS. Burst CPU/op +4.71%, WAL/op +5.29%, p99 −0.10% pass
relative gates, but absolute drains above fail. Identical operation/SQL
counts, zero lost leases/drops/errors. One Node process simulates 300
controllers; this does not qualify two-webserver overlap or HTTP
capacity. |
| Live acceptance | Strict final-source Linux classic Codex case passes
1/1 on attempt 1, all 49 checks. Active recovery within grace and useful
retained C continuation succeed; four
accepted/published/committed/settled outcomes. Independent provider,
PG/process, VM/root and SSH cleanup complete. Prepared/task and broader
rollout limits are explicit above. |

**Historical qualification snapshot through `9eb9ac6`.** The head at
this snapshot was `9eb9ac6f9e312f3ebe8040f5a7565e88ae8824f3`. Server and
Runner package sources are byte-identical to `283e05ffaa`; the two later
commits change the explicitly scoped qualification and its documentation
only. It includes master through the storage-owner fix
`45a3a92e6dc52d20923b9aa34a22fee2ec5443f9` (#18357). At that snapshot,
CI/review passed but the strict post-C native recovery gate failed; the
PR stayed draft and required another native repair. Prior `39a4343` CI
and review passed, but its full live workflow failed; that grade remains
unchanged. The first repaired `435ccd2` live attempt failed before any
injected outage. It exposed a legitimate terminal snapshot before result
acceptance; no accepted result ID or publication receipt was observed.
Its canonical report remains 0/1 `candidate_failure`, with cleanup
passed and independent provider absence confirmed. `0862062` defers
publication until result acceptance and fixes the fresh review's
coordinator/issue lock-order finding. Both failures have deterministic
real-Postgres regressions. The `0862062` live attempt then proved
accepted-result publication but failed active recovery: a resumed-lease
preparation guard demanded suspension after the exact live Runner was
already verified for adoption. `64c0b746` preserves that verified
active/idle-warm process, revalidates its proof/liveness, and keeps
ordinary suspension checks and later PRP authentication. Its live run
passed preparation but exposed a second active-warm boundary: all
durable binding fields matched except opaque turn/item attachment IDs,
which the server had regenerated. `7a203274` carries the complete
verified six-field binding and keeps the strict Runner mismatch guard.
Its cloud case then passed active adoption, completion, publication and
accounting. The queued continuation exposed accepted-result retirement
writes and an implicit-home fingerprint drift; both are fixed with
deterministic regressions in `39a4343` before this fresh full
qualification.

The prior live case exposed a real publication gap. A goal-less native
turn could persist its accepted result and terminal snapshot before
provider cleanup, while the task-session pointer was still empty. Normal
later admission then selected a fresh session. The fix publishes that
pointer and a private bound receipt in the **existing post-acceptance
terminal checkpoint transaction**, before cleanup. Earlier terminal
observations persist without a handled receipt or continuation pointer.
No result-acceptance or snapshot transaction was added. The late
completion path recognizes the receipt. This also prevents a late
callback from restoring a pointer after a user reset. Goal/control
publication keeps its existing owner.

The dedicated Daytona fixture now uses the existing warm/reusable
environment. Its sandbox, Runner and provider storage must survive. The
former generic per-turn fixture disposed the sandbox after completion
and kept provider-history backup bytes under the discarded directory.
Whole per-turn disk-loss acceptance therefore has a separate
provider-backup storage dependency. It remains distinct from the passing
native per-turn component checks with backup storage available. The
60-second lease, 300-second outage and 420-second grace are unchanged.

| Gate | Current repair evidence |
| --- | --- |
| Assembled regression | Baseline fails with an accepted result and
terminal snapshot but a NULL task pointer while real provider cleanup is
held. The fix passes normal continuation and reset cases, with no
pointer seeding. Exact merged `92b7447` passed 2/2 twice, including
after normal package TypeScript output refresh. Production files are
byte-identical at frozen `435ccd2`. |
| Persistence and concurrency | 19 helper/transaction cases and all 59
existing control-port cases pass on `0862062`. They cover bound result
IDs, reset/newer epochs, Stop, lock waits and expiry, receipt replay,
missing stubs, foreign bindings and atomic 4 KiB rollback. Existing
snapshot transaction count stays one. The new tests reproduce the exact
live missing-result-ID error and the inverse replay/status lock order
against `435ccd2`, then pass after the correction. Replayed results
still reject stale generations. |
| Completion | All 32 completion-boundary tests pass on merged
`92b7447`; the source is unchanged at `435ccd2`. Native result-shaped
session metadata is preserved. |
| Qualification fixture | A valid review report reproduced three
rejected correct replies. Exact token-boundary matching fixes them. 140
focused E2E/catalog cases and Product E2E typecheck pass. Wrong tokens,
marker prefixes/suffixes and wrong authors remain rejected;
retained-identity checks are unchanged. |
| Required checks | Full recursive typecheck and standard build passed
at `435ccd2`. Both exact commands also passed on `0862062`, with its
exact server build stamp; both also pass on `7a203274`, with its exact
server build stamp. Fresh `0862062` Greptile passed 5/5, with zero
unresolved threads. The `64c0b746` normal CI then passed all 49 jobs (48
success, one expected skip) and Greptile passed 5/5 with zero unresolved
findings. The `7a203274` normal CI also passed 48 jobs with one expected
skip, with fresh Greptile 5/5 and zero unresolved findings. Current
`39a4343` recursive typecheck and standard build both pass with
persisted exit codes and its exact build stamp; fresh Greptile is 5/5
with zero unresolved findings. Its normal CI completed successfully: 48
jobs passed and one expected job skipped. The fresh live result remains
separately failed. The `0862062` normal CI had 46 passing jobs, one
expected skip, one worker communication failure and its aggregate
failure; GitHub states the self-hosted runner lost communication. No
failed assertion or specific OOM/spot cause was established, and no
blind retry was used. |
| PostgreSQL supplement | One corrected all-pool pair at the preceding
repair passed 300-controller/10,000-history correctness and absolute
gates. SQL 76,590→75,990; transactions 9,013→8,713; UPDATE rows 11,613
each; 300 correct pointer writes each. PG CPU 31.99→30.46 s, WAL
70.36→67.87 MB, request p99 100.98→103.12 ms. This is one pair, not a
repeated improvement claim or an exact `0862062` CPU measurement. Its
two later changes defer premature publication and reorder existing
locks; they add no transaction, stream/idle writer, pool or timer. The
first collapsed-pool attempt remains invalid and retained. This is not
an exact final-head CPU/WAL measurement; later fixes change ordering,
identity selection and in-memory skill replacement without adding
snapshot transactions, pools, stream writes or idle SQL. |
| Live publication proof (`0862062`) | Baseline run
`06fcc806-f186-4190-9b57-087b61d75533` accepted result
`58f912dc-94d9-4b4d-bd3d-d5f689c83fd3`; the publication receipt names
the same result, the task pointer has parameters and `lastRunId`
matches, and the normalized native identity matches the following active
run. Accounting v2 settled. |
| Active recovery (`0862062`) | **FAIL.** SIGKILL/cache discard at
23:07:28.666 UTC; hold 300,000.859 ms; B ready at 308,896.486 ms;
natural generation-2 ownership at 312,906 ms. At about 319 s,
preparation failed `runner_harness_state_mismatch:
resumed_not_suspended`. No authenticated after-receipt passed within 420
s. Exact original Runner PID 668/start/fingerprint was independently
still alive; provider auto-stop was 15 minutes. No retained C follow-up
qualified. Canonical 0/1 `cleanup_failure`/cleanup-failed report remains
failed; subsequent independent cleanup deleted all three owned resources
with exact 404s and confirmed full intended-account inventory with zero
owned resources. |
| Verified adoption preparation (`64c0b746`) | 91 focused cases pass,
including actual preparation for four active providers and a distinct
prior/current warm proof. Unproved resumed state, dead process and stale
warm owner remain rejected. The earlier options-only fixture missed this
gate. No spawn, backup restore or PRP-authentication bypass was added.
Server typecheck passes. |
| Active warm identity (`64c0b746`) | **FAIL.** B claimed generation 2
at 312,995 ms. Preparation passed, then strict PRP recovery rejected the
generated attachment IDs at about 319 s. The saved and desired Runner,
lease, normalized session and run IDs matched; only
`turn_attachment_…`/`item_attachment_…` differed from generated
`turn-<run>`/`item-<run>`. Canonical 0/1 cleanup-failed report remains
failed; independent cleanup confirmed deletion/404s and zero owned
provider resources. |
| Six-field identity repair (`7a203274`) | Actual server-factory
regression RED4→GREEN38. A real Rust Runner test passes healthy first
run → random-ID warm handoff → held active second run → restored
controller. It verifies a new live authenticated connection, exact
identity, same Runner/provider OS identities, no duplicate launch/turn,
and no redispatch or invented receipt for an uncertain tool. The
synthetic wrong pair still fails the unchanged guard. This protocol case
does not claim business completion. Server and Runner typechecks pass. |
| Active recovery/completion (`7a203274`) | **PASS for the active
component.** Full five-minute SIGKILL/cache-loss hold. B ready at
308,936.651 ms; natural generation 2 observed 312,979 ms; authenticated
at 317,720 ms and resumed processing at 318,527 ms. Same run
`1e1ccb00-0728-4935-97da-57562d81cb42`, Runner
`246f73f7-a689-4177-a9c2-7ce711d63aed`, remote PID 669/start
23:48:27.470, provider lease/session and one submitted turn. Accepted
result `49097529-5310-4ac3-af95-5bf9b91d19f4`, publication, finalization
and accounting v2 all settled. |
| Full workflow (`7a203274`) | **FAIL, preserved.** Queued
late-requirement run `5e61f38d-c7d1-4fc1-bcbf-bce9d46ce4eb` failed
`runner did not durably suspend before checkpoint` before outage 2.
Canonical 0/1 candidate failure, cleanup passed, evidence valid. C
retention was not exercised. Independent provider absence,
controllers/PG cleanup, EC2 termination and root-volume absence all
confirmed; closed receipt SHA-256
`676a3b15eb9937b6bf995c36fb8666c88964d509e5ceb9109fb02da0aa2bc909`. |
| Retirement and stable configuration (`39a4343`) | Real Runner +
Postgres RED reproduced the exact suspension failure and nine owner
rejections. `turn.stop` is now lifecycle checkpoint maintenance under
unchanged owner/result/generation fences. GREEN25/25 proves actual
persisted suspension, exact identity and completed stop/drain/suspend;
unaccepted, foreign, stale and new-work negatives remain rejected. Two
default-home fingerprint REDs become GREEN9/9; explicit mode changes
still invalidate. No global matching exemption or historical fingerprint
rewrite. |
| Fresh full Daytona warm cell (`39a4343`) | **FAIL, preserved.**
Launched 00:15:29 UTC with automatic retries disabled. Active adoption
passed: actual offline hold 300,000.848 ms; B ready 309,922.155 ms;
natural generation 2 at 311,821 ms; authenticated 317,989 ms; resumed
330,440 ms. Same run `3e126819-e786-4b3b-b923-c975cf4f5b79`, Runner
`4079138a-8d7c-44a4-8558-07a5411d1f55`, PID 668/start and
provider/session identities. Accepted result
`08e504af-4dec-48a8-bf05-bf0a32280559` and publication passed. Queued
run `31d168c9-7ab9-4700-abb6-fafb90912237` failed the same strict
suspension error before outage 2. Its row selected a new native
identity, but no new third provider process or invocation was observed.
Recovery-time configuration still reset only `adapterConfig`; native
provider, context and workspace subtrees were unchanged. Canonical 0/1
candidate failure, evidence valid, normal cleanup passed; all campaign
provider/PG/process resources independently absent. Closed campaign
receipt SHA-256
`5e7d6281c13c812f3a946667f97f3c7d339b5ad86a3a819bd185b2fa73bc4a19`;
EC2/SSH remain temporarily held for the concrete follow-up repair.
Original 60/300/420-second limits and strict queued/second-replacement
checks remain unchanged. Direct-process path; no new live prepared-task
claim. |
| Native retirement authority (`283e05ffaa`) | Exact `39a4343` facade
regression RED3→GREEN9. The successor now uses its existing production
coordinator claim and prompt renewal before strict retirement. A real
Runner/Postgres case starts in the actual native-observed/unclaimed
phase, claims through production code, then closes and durably suspends
the predecessor in 1.657 s. All 25 real-Postgres cases pass;
unclaimed/foreign writers remain rejected. Shared attachment staging
stays after retirement; Stop and denied claims neither close nor stage.
No new claim or SQL path was added. |
| Configuration continuity (`283e05ffaa`) | Real shared/project-primary
environment selection → configuration with four skill versions
reproduces `reset: true`, changed category exactly `adapterConfig`,
after a Postgres JSONB read. Pinned restoration had reordered the
generated preference while actual runtime skills stayed identical.
Replacing existing pinned rows in place preserves the original fresh
name/key order and fingerprint. The assembled regression passes, as do
three existing pinned-version/rename/alias cases and server typecheck.
Genuine model/skill-selection changes still invalidate. No
fingerprint-format change or broad compatibility exemption. |
| Final production repair (`283e05ffaa`) | Full recursive
typecheck/build and [normal
CI](https://github.com/paperclipai/paperclip/actions/runs/41151145106)
PASS: 48 jobs plus one expected skip. Fresh review 5/5, zero unresolved
threads. Live baseline, recovered run and queued continuation all
succeeded, each accepted/published/committed once with accounting v2
settled. Their native/provider identities and configuration fingerprint
match; both recovery and continuation report no reset, no changed
categories and task-session reuse. Active outage/authentication remains
within original timing. |
| Browser-dependent continuation (`283e05ffaa`, definition 6) | **Whole
cell FAIL, preserved.** After B was killed at 00:46:38.928 and its cache
discarded, C became ready in 9,151.410 ms. `openParent()` then hit a
30-second `page.goto`/DOMContentLoaded timeout before posting the final
request. No fourth native run or C continuation was exercised. Cause of
navigation timeout is unknown. C task had no execution lock, queued
comments, handoff or recovery blocker; all three prior runs were
committed/accounted. Normal cleanup passed; independent provider
absence, controller/PG cleanup, EC2 termination and exact root-volume
absence confirmed. Sanitized report bundle SHA-256
`2a3b841d62efb2ae30ef14f96a431ac4fa7fe3a782f1144831512f78edc65022`. |
| Native API continuation qualification (`9eb9ac6f9e`, definition 7) |
**FAIL, preserved.** The real post-C API request dispatched run
`c71fbb18-f979-4fe4-93ce-9c4be46dbf21`, which failed
`native_session_interrupted` / `native_session_recovery_failed` at
01:10:00.640 UTC. It selected the same normalized native identity,
Runner and checkpoint-provider identity as the three succeeded turns. No
C result was accepted; pending accounting remains explicit. Active
recovery passed: actual offline hold 300,001.098 ms, B ready 308,335.019
ms, natural generation 2 observed 311,818 ms, authenticated 315,540 ms,
resumed 316,612 ms. Queued continuation succeeded/accepted/published. B
was then killed/cache discarded at 01:09:34.500 and C ready in 8,783.803
ms. Original classifier `transient_infrastructure`, whole-case failure
and cleanup passed are unchanged; that classifier does not establish the
root cause. All owned provider/PG/process resources, EC2, root volume
and SSH agent independently absent. Closed receipt SHA-256
`7f21ac4a22efc07b849fa7e17866083b6127b3321af12671b4fc054ce57039a9`;
sanitized bundle
`71ffef14a30a553e008320a31840a6d6c0de85c1bfdb60f783fdb53a6aae812a`.
Exact backend detail is unavailable in approved structured evidence. |
| Current required checks (`9eb9ac6f9e`) | Recursive typecheck, standard
build and exact build stamp PASS. [Normal
CI](https://github.com/paperclipai/paperclip/actions/runs/41152908417)
PASS: 48 jobs plus one expected skip. Fresh Greptile 5/5, latest edited
summary checked, zero unresolved threads. Exact8/512/8,192 stream probes
again report zero recovery SQL/transactions/rows and settled-idle SQL;
seven control transactions/twelve changed rows each, with 36 separate
legitimate renewals in the longest case. CI success does not supersede
the failed live C gate. |

The completion supplement observed secondary request/liveness/checkpoint
transaction p99 increases of 23.0%/45.7%/15.6% in its single pair.
Renewal p99 stayed below 0.865 s, with 900 successful renewals per arm
and no lost leases. The named pool caps stayed 10+2+2. An aggregate
activity sample peaked at 15 non-observer entries without backend-type
classification, so no claim is made that all were client connections. No
sampled lock episodes does not establish zero lock waits.

Twelve pure reader/selector cases also pass with the new adjacent
receipt. This is not another executed process rollback. Development
reader `5cf7aead` can parse the recovery state, but its old run-detail
redactor can expose receipt IDs/outcome to an authorized run reader. The
first production release containing recovery readers, complete
attachment-identity handling and both private-field redactions is the
rollback floor. The implicit-home fix prevents new recovery-time
fingerprint drift; it does not rewrite previously stored fingerprints
from failed candidates. The development probe is not a published
supported rollback target.

Latest actual `0862062` Runner tests independently passed 8/512/8,192
text chunks with zero recovery SQL, transactions and changed rows, plus
zero settled-idle SQL. All three observed seven setup/control
transactions and twelve changed rows. The 8,192-chunk case separately
recorded 31 legitimate lease-renewal statements, with no lost lease.
These are component SQL observations, not a production CPU/WAL claim.
The new adoption and identity branches affect recovery preparation only.
The `64c0b746`, `7a203274` and exact `39a4343` CI independently passed
all three stream sizes with the same zero-SQL/idle result; each latest
8,192-chunk case recorded 36 separate renewal statements. Exact
`39a4343` Runner 2/2 passed 133 files / 2,317 tests plus two skips; its
separate native group passed 150 tests plus eight skips.

The first API fixture revision omitted the mandatory final screenshot;
this was caught before paid launch and allocated no provider/model work.
The correction captures the actual retained Chromium page with an
explicit unqualified-browser label, without changing the standard report
validator.

Automatic approval review rejected exporting raw private API/workflow
snapshots because of sensitive payloads and insufficient
destination-specific authorization. Sanitized reports and
bound/hash-only receipts are retained; raw snapshots were not copied
through another channel. This does not change the failed grade.

Repair characterization also preserves its failed attempts. An
unconditional staging relocation broke three existing same-scope
scripted cases; the final implementation defers staging only for an
actual superseded warm scope, and all nine order/race cases pass. One
baseline transform overlapped an edit and is explicitly invalid; the
actual immutable `39a4343` loader then produced the three genuine REDs.
The independent task-directory hypothesis produced a RED but was
excluded by the live shared-workspace evidence; no directory policy
change was landed. A shared-workspace fixture authorization omission was
corrected separately from the four-skill semantic RED.

Setup failures are retained separately from semantic outcomes. These
include an old host port after the owned Postgres fixture restarted and
a sandbox-denied TSX IPC pipe. The supported external-Postgres run
passed, and the final exact recursive typecheck passes with normal
process permissions. Seven stale declaration errors after the master
merge cleared after compiling the actual merged SDK/Runner TypeScript
sources. No declarations were edited by hand.

<details>
<summary>Earlier qualification measurements and failed attempts
(historical, before this repair)</summary>

**Historical blocked gate (before the repair below).** Final native
implementation is `abed921e12d543b62f57a56435419c0dafb0bc33`; current
branch head is `af9e2d157dcc349e55374ad186c9be2d6cb96bf9`. The two later
commits change only qualification fixtures, their matcher and
documentation; `server/` and `packages/` are unchanged. It includes
master through `fd008979f5794fda3d5f54151d36acd31d0de140`.

The live Daytona measurement proves five-minute active recovery,
once-only completion and accounting, but **fails retained-session
identity after the second replacement**. The saved checkpoint and
recovery metadata exist. The task-session pointer remains empty, and
ordinary admission creates new native/provider sessions. The exact
skipped publication boundary remains unproven. A useful saved-file
result alone does not satisfy retained-session continuation. This
application-path gap blocks merge independently of CI.

A separate new-master CI failure is owned by the storage workstream (fix
in #18357). No storage guard, session-identity oracle, lease/grace limit
or provider result was relaxed.

| Gate | Observed result and exact scope |
| --- | --- |
| Distinct source recovery and compatible rollback | PASS at A
`5cf7aead09a2d9c7e2f2929d761f535ef3c1d798` → B
`5ea45f8d17e33c9a62bd79982dcf0375b07733d0` → fresh B → A. Four
controller processes, natural 60-second lease expiry, absent native
projections before replacements, three succeeded/committed unique
results, one provider thread, unchanged Runner identity, and strict idle
cleanup. 177.782 seconds. |
| Immutable installed native component | PASS at installed A
`5cf7aead09a2d9c7e2f2929d761f535ef3c1d798` → installed B
`eaef0d56552c087aaa4fa68a8655282ffbd12667` → fresh B → A. Each
controller imports its contained compiled server and DB. Three unique
committed results, stable Runner/thread, no extra provider restart on
same-contract rollback, strict cleanup. 106.672 seconds. This is
compiled component proof, separate from full CLI-native application
acceptance. |
| Normal public CLI startup | PASS for B
`eaef0d56552c087aaa4fa68a8655282ffbd12667`, Node 24.22.1, darwin-arm64:
fresh external DB, 415 migrations, health ready, root HTML 200, company
POST 201 plus GET readback. The installed tree remained unchanged after
execution. |
| Five-minute component outage | Actual controller absence 300,168 ms.
Natural generation-2 claim at 407.196 seconds after crash, same provider
attempt and Runner, original run succeeded/committed. The larger
540-second case timed out before its retained follow-ups, so it is an
active-adoption pass, not a complete four-controller pass. The complete
short source/installed cases above independently cover follow-up and
rollback. |
| Drain/resume review regression | Fixed in
`226f03e736d1ee99fea1762ab32ea3a8a9241c20`. Five facade-level cases
pass: overdue normal/restart waits, future timer cancellation, another
facade resuming the original owner, and transient retry. Against exact
prior scheduler `5ea45f8d17e33c9a62bd79982dcf0375b07733d0`, four fail
and one passes. The later head changes documentation only. |
| Final corrected-fixture Daytona | **PARTIAL: active recovery PASS;
retained identity FAIL**, exact live source `0b6afc499f`. A was
SIGKILLed and native cache discarded; actual offline hold 300,000.375
ms. B was ready at 308,326.961 ms, naturally claimed generation 2 at
312,059 ms, authenticated at 315,996 ms and resumed processing at
316,519 ms, within the original 420-second boundary. The same run,
Runner PID/start, session, lease and event prefix survived; one turn
submitted. Original execution succeeded and settled accounting. After
completion, B was killed and its native cache discarded; C was ready in
9,077.779 ms. C executed the saved project and returned correct
`queue_r`, with all four runs succeeded/accounted. However, native and
true provider-session identities changed: the unchanged retained oracle
fails. No whole-workflow pass is claimed. |
| Prior matched-producer Daytona | FAILED at exact `abed921e12`. The
matching current-source Runner passed normal admission, completed the
first task and settled its accounting. During the active revision, A was
SIGKILLed and its native cache discarded. Actual offline hold was
300,000.727 ms; B was ready at 308,304 ms and naturally claimed
generation 2 at 309,651 ms. At about 318 seconds, the remote process
check reported `process_not_running` for the original admitted Runner.
No authenticated after-receipt or retained follow-up passed. The current
failure is preserved; no automatic retry, lease/grace relaxation, debt
edit or provider replay is used. Bounded forensic read found preserved
exact-bound state with no recoverable failure, a transport-reset
diagnostic and no terminal delivery; this supports external process loss
without proving the exact cause. The sandbox had a five-minute auto-stop
policy, conflicting with the required surviving-sandbox assumption.
Owned cleanup precedes the separate corrected-fixture case. |

**Live retained-session boundary.** Prior completed run `3806…` and C
follow-up `8adf1df3-6744-40bf-b35d-100451f29d9f` both have database
session snapshots and `nativeControllerRecovery`. Their tool
fingerprint, provider configuration, workspace descriptor,
runtime-context digest, driver and work mode match. There is no recorded
force-fresh/workspace-selection change, and this is an ordinary task
rather than a conversation-generation reset. The same issue/agent
task-session row has `session_params_json: null` and `last_run_id: null`
from baseline time. Both native normalized UUID and actual provider
session/thread hashes differ after C. Full application admission
therefore did not select the retained source. No fallback to file
execution or relaxed identity comparison is accepted as proof of
continuity.

**Independent cleanup completed.** Full current-account provider
inventory confirms no resources owned by either final fixture, with
exact-ID absence checks. Owned controllers and campaign Postgres are
stopped/removed. EC2 `i-01551c09af098dde5` is terminated, its
delete-on-termination root volume returned `InvalidVolume.NotFound` at
22:14:08 UTC, and the temporary SSH key/agent were removed. The
canonical failed report and immutable oracle input are retained. Closed
qualification receipt SHA-256:
`b4b5fefce287ebfdbf2cc2238efaad62d8ad485aed0e9c0628ae95dbc18a8575`.

**Original measurement and grader repair remain separate.** The C reply
contains the requested marker followed by a newline and the correct
output. The old workflow waited for one literal space, so its normal
receipt collection stalled; that original attempt was closed and its
canonical `cleanup_failure`/interrupted result is preserved. Commit
`af9e2d157d` repairs only this reply predicate, with 144 focused cases,
E2E typecheck, discovery and diff checks passing. Ordinary string
matching stays literal. The corrected predicate accepts the saved reply,
but the unchanged retained-identity oracle still fails. Pure
active-recovery grading of the captured measurement passes 7/7; this is
not a rewritten passing campaign.

The sandbox idle-window conflict was corrected only for this explicit
fixture (15 minutes). Actual baseline and later sandbox metadata confirm
that policy; the original active box was disposed before a post-gate
policy read, so its 15-minute setting is established by the normal
create/environment configuration rather than an after-read of that box.
Production native lifecycle and the 60/300/420-second constants remain
unchanged.

The development compatibility reader A is a genuine backport commit
based on `59ff89a18fe1312b5f0bb908439bbdb1237b0252`, with the required
idle-goal reader fixes. It is not a previously published release. Its
older executor differs from B. The rollback floor must include this
reader capability. The source component cases admitted the Linux ARM64
Runner packaged at `59ff89a18fe1312b5f0bb908439bbdb1237b0252`; the
installed component case admitted the `5cf7aead` package. Their Runner
bytes are identical: SHA-256
`b1f04caa1ce04e720dec7eb62ede170eb3621948c0849b8f305ea5c5116bced8`. Each
admitted artifact remained unchanged throughout its case.

[Final public package
build](https://github.com/paperclipai/paperclip/actions/runs/41121460635)
and [reader-floor
build](https://github.com/paperclipai/paperclip/actions/runs/41122170506)
both passed all four daemon builds, package assembly and both Darwin
clean consumers. Both Linux ARM consumers failed before provider work
with the same existing local Codex permission-profile error: bubblewrap
attempts `activity.fifo/.git` and reports “Not a directory.” This
matched baseline failure remains open; it is not a native remote
continuity pass. Full post-run tree and build-stamp revalidation also
passed after the compiled native case. The installed component trees are
A `c29114c1bf71c5af00f761aeb9e3d84ed3d3075b9f5600dee3ef2543eb2b6ec5` and
B `eda575505373a3bdc630fa4919862b359791fe3a7ef637961a25b73b38efe4b0`.

**Controlled PostgreSQL qualification passed.**
Baseline/candidate/candidate/baseline workloads used 300 controllers,
10,000 historical tasks, real snapshot/encrypted PRP writes and
unchanged pools. PostgreSQL and client were each limited to two CPUs and
1536 MiB on an isolated eight-vCPU Linux host.

| Workload | PG CPU per operation | WAL per operation | Database request
p99 |
| --- | ---: | ---: | ---: |
| Steady | +1.18% | +1.81% | +1.91% |
| Aligned burst | +3.40% | +2.56% | +3.81% |
| Two logical servers | +3.11% | +2.56% | +9.68% |

All arms passed absolute/correctness gates with unchanged control
transaction counts, no dropped operations and no owner loss. Overlap p99
has little margin below the relative threshold. It uses two clients in
one Node process, not two HTTP servers. [Closed measurements and
limitations](doc/plans/2026-10-30-native-controller-qualification-evidence.json)
and exact-source reproduction patches are retained. [Earlier
contended-host
failures](doc/plans/2026-10-30-native-controller-load-evidence.json)
remain unchanged. The production DB-claim probe claimed 300 owners in
2.84 seconds and restored/committed controls and snapshots in 8.74
seconds, with zero owner loss or settled-idle SQL; its provider
references were synthetic. The measured checkpoint/snapshot/renewal
paths retain the same persistence behavior after integration; later
accounting and business writes have separate owners.

Actual Runner stream tests at `da8c7f6e1837ddd896dfe83a7524812fee813487`
passed 8/512/8,192 chunks: zero recovery SQL, transactions and changed
rows during text intervals; zero settled-idle SQL. Each retained seven
control transactions and twelve changed records. Existing renewal SQL
was measured separately. The guard/scheduler fixes add no stream path,
pool or recurring idle scan.

The live fixture used current-source producer SHA-256
`f0000c90cb83880191479da639201e659aa6ce0fc7d67c859faa172ec977849e`, a
normal locked exact-head build with pinned Rust 1.97.1 in Bookworm.
Actual network-isolated, read-only ABI execution passed (GLIBC
requirement 2.34 against runtime 2.36). The immutable base image
remained
`ghcr.io/paperclipai/paperclip-daytona-runner@sha256:36069788f53c51f228764fa36180575443dfb5d1f84a342d7c0daceb28dc61cc`,
source `a0ef3ce7`. This was an unpublished development producer,
explicitly distinct from the base image and admitted before the fresh
fixture began. The earlier host build failed ABI preflight before
admission. The preceding old-producer fixture completed its initial task
but lacked the usage-basis marker required by current master; it
remained pending/unpriced and correctly paused. It was independently
cleaned without debt or budget changes. Two preceding local Linux setup
attempts OOM-killed before provider allocation, and their cleanup
passed. These attempts remain separate from the final recovery failure.

**Failures retained and resolved separately.** The earlier `98f05fc`
probe rejected a valid settled PRP v2 goal suffix; `eaef0d5` fixed that
reader and the fresh follow-up passed. Later component cleanup failures
were traced by a scoped exception observer to the fixture adapter
rejecting directory downloads. A bounded owned-directory transfer with
exact-byte/mode preflight corrected the fixture; source and installed
strict-cleanup cases then passed. No production guard was relaxed. The
initial local integrated control-port run hit its 30-second embedded
Postgres startup hook before executing 59 cases; 18 other focused cases
passed. A bounded rerun used a fresh owned external Postgres fixture,
applied normal migrations before Vitest, and passed all 59 cases with
the original timeouts unchanged. Its container and data were removed;
the failed attempt remains retained.

The additional opt-in full Product E2E Vitest run is not green: 5,313
passed, seven failed and three skipped. An actual `abed921e12` archive
with the same dependencies, Node and IPC permissions reproduces all
seven failures with the same reasons. Five are unchanged
workflow-policy/profile-pin assertions, including Pi profile version 37
versus the test pin 35. Two database suites failed during embedded
Postgres startup before their assertions; semaphore exhaustion is
confirmed in the observer log. The observer passes its supported
external-Postgres rerun with original timeouts and source unchanged. The
controller-database client-kill/postmaster assertions remain unexecuted
after `initdb` bootstrap failure; no external override is supported. The
unit launcher stopped before its Node-test phase, which was then run
independently and passed 181/181. The original failed full run is
preserved. These results remain distinct from the passing focused checks
and normal repository CI. The first local Product E2E typecheck also
encountered stale compiled Runner declarations; a genuine package-local
TypeScript compile refreshed them and the unchanged check passed,
without a dependency install or producer rebuild.

Earlier [Daytona campaign
41108162900](https://github.com/paperclipai/paperclip/actions/runs/41108162900)
failed at sandbox creation before any agent, with the underlying
provider error unavailable. Its [sanitized
report](https://d1p6rlowie26tp.cloudfront.net/runner-e2e/campaigns/gha-41108162900-2/index.html)
remains a `cleanup_failure`. Both old creation names are absent in the
current authenticated account, but the old CI account identity was not
retained, so account equivalence is unproven. The subsequent exact
`eaef0d5` source Product E2E attempt passed image/agent setup and the
initial project, then held the controller offline for 300,003.7 ms. Its
Mac supervisor failed owned-process inspection before B readiness; no
adoption passed. Full current-company lease/label checks, provider
deletion, exact-ID 404s and complete inventory confirmed cleanup of that
attempt. Its local server and owned Postgres stopped. The prior AWS
campaigns and database qualification VM were terminated with root
volumes removed.

The old admitted test Runner lacks the newer usage-basis marker
introduced on master. The integrated component B turns retained
unpriced, incomplete accounting obligations with `costAccountingPending:
true` and no settled charge. The compatibility-reader A turn used its
legacy receipt interpretation but likewise retained pending accounting
and no settlement timestamp. This proves obligation preservation, not a
whole-run accounting migration. Runner/receiver release ordering remains
governed by the usage owner's contract.

At native implementation `abed921e12d543b62f57a56435419c0dafb0bc33`,
Greptile was 5/5 with zero unresolved threads and CI completed with 54
successful checks and five expected skips. At `0b6afc499f`, Greptile
again passed 5/5 with zero unresolved threads, but CI ended with 51
successful checks, five skips and three failures. Server 9/12 fails a
deterministic managed-skills maintenance assertion newly inherited from
master `198218bf` in merge `c80e00ad`; that test is absent from the PR
head, and the storage owner has the explicit-StorageService fix in
#18357. Runner 2/2 passed 2,283 unit cases but its runner received a
shutdown signal during the long stream probe; its cancelled step and
aggregate failure are preserved. No blind retry was used. Current
`af9e2d157d` checks are pending; the matcher-only change passes 144
focused cases and Product E2E typecheck. Earlier scheduling-sensitive CI
failures and the interrupted unsharded local test run remain recorded in
the dated plan. CI runs the repository's normal test partitions. Final
verification does not claim Railway rollout, uninterrupted HTTP service
or whole-webserver filesystem loss.

</details>

## Risks

- The task-pointer repair passes its assembled regression. The first
live repair exposed pre-acceptance ordering and the fresh review exposed
a replay lock-order issue; both are fixed and pass real Postgres tests
at `0862062`. The following live attempt exposed the
verified-running-versus-suspended preparation gate. Its narrow fix
passes 91 focused cases at `64c0b746`. The next live case exposed
synthetic attachment IDs; `7a203274` fixes that field selection and
passes real Runner authentication. The subsequent queued continuation
exposed strict retirement maintenance and a false default-home change;
`39a4343` fixes both with real Runner/Postgres and fingerprint
regressions. The `39a4343` live continuation failed again despite green
CI. `283e05ffaa` repairs the actual native-observed admission ordering
and reproduces/fixes the JSONB skill-order fingerprint drift. Both pass
targeted real-Postgres regressions and the live active plus queued
paths. The definition 6 post-C browser navigation timed out before
admission. Explicit definition 7 now verifies native continuation
through the ordinary comment API; its strict post-C native gate failed
after a real dispatch and remains historical failed evidence. The final
fc68584 repair now passes its fresh strict 49-check Linux/Daytona
campaign; the prior failed grades remain unchanged. Browser reconnection
is handed to its own workstream with the original failed evidence. No
prior failed campaign is reclassified as passing.
- Incorrect checkpoint/source binding or stale ownership could duplicate
external work. Missing evidence remains a hold; existing receipt and
finalization owners keep uncertain effects unresolved.
- Recovery must retain maintenance writes after accepted results without
allowing a stale business turn to write. Real lock-wait and generation
tests cover that boundary.
- The current two-connection paired burst profile fails its absolute
drain gate in both baseline and candidate, despite passing relative load
gates and all correctness checks. A causal capacity diagnostic selected
three slots; the final controlled cap-three profiles pass unchanged
thresholds. Full application replacement and two-server overlap capacity
remain separate. There is no new streaming writer, pool or recurring
idle scan.
- The remote sandbox and Runner must survive the recovery window.
Provider auto-stop is a separate lifetime bound from reconnect grace;
the explicit qualification fixture uses a longer idle window.
- Workspace and provider backup availability remain external
dependencies. This PR does not certify whole-webserver filesystem loss,
uninterrupted HTTP service or a Railway rollout.

## Model Used

Implementation and initial verification used OpenAI GPT-6 through Codex;
the initial hosted snapshot was not exposed. The follow-up qualification
used three OpenAI `gpt-6.1-sol` workers coordinated through Codex, with
repository tools, code execution, CI and scoped test infrastructure.
Context-window size is not exposed.

## Checklist

- [x] I have included a thinking path that traces from project context
to this change
- [x] I have specified the model used (with version and capability
details)
- [x] I have checked ROADMAP.md and confirmed this PR does not duplicate
planned core work
- [x] I have searched GitHub for duplicate or related PRs and linked
them above
- [x] I have either (a) linked existing issues with `Fixes: #` / `Closes
#` / `Refs #` OR (b) described the issue in-PR following the relevant
issue template
- [x] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
and contains no internal Paperclip ticket id or instance-derived details
- [x] I have run tests locally and they pass (focused and exact-source
application/cost checks pass; full forced CI passes; earlier local
embedded-PG setup failures remain recorded)
- [x] I have added or updated tests where applicable
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] All Paperclip CI gates are green (exact fc68584 full CI and
auxiliary checks pass; forced complete test partitions)
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
(exact fc68584 check and edited summary pass; all six threads resolved)
- [x] I will address all Greptile and reviewer comments before
requesting merge

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-10-31 07:33:52 -05:00

3251 lines
159 KiB
TypeScript

import { kimiCandidateAdmissionForExecution } from "./server-restart-request.js";
import { runPlanRevisionAcceptance, isPendingPlanConfirmation, normalizePlanMarkdown, expectPlanStageVisible } from "./plan-revision-flow.js";
import { createWorkerTiming } from "./worker-timing.js";
import { captureTaskChat, type TaskChatEvidence } from "./task-chat.js";
import { FEEDBACK_PERSISTENCE_SUITE, gradeFeedbackPersistence } from "./feedback-persistence.js";
import { EvalAttemptTrace, evalBraintrustConfig, EVAL_OBSERVABILITY_SECRET_NAMES } from "./braintrust.js";
import { captureFollowupStopFailureApiState } from "./followup-stop-evidence.js";
import { controllerOutageForExecution, validControllerOutageReceipt, type ControllerOutage, type ControllerOutageReceipt } from "./controller-outage.js";
import { runFollowupStopResumeFlow } from "./followup-stop-resume-flow.js";
import { runFinalizationRecoveryFlow } from "./finalization-recovery-flow.js";
import type { FinalizationCrashBinding } from "./finalization-crash-boundary.js";
import { runCaseHandoffsFlow } from "./case-handoffs-flow.js";
import { runCaseReceiptFlow } from "./case-receipt-flow.js";
import { runTaskWorkspacesFlow } from "./task-workspaces-flow.js";
import { runRepositoryVerificationFlow } from "./repository-verification-flow.js";
import { runAutomaticRepositoryFlow } from "./automatic-repository-flow.js";
import { runNativeActiveStopFlow as runCopilotActiveStopFlow } from "./copilot-active-stop-flow.js";
import { nativeActiveStopTasks as copilotActiveStopTasks } from "./copilot-active-stop-tasks.js";
import type { RestartRunnerIdentity } from "./process-tree-owner.js";
import { runNativeActiveStopFlow } from "./native-active-stop-flow.js";
import { runKimiLifecycleFlow } from "./kimi-lifecycle.js";
import { runPiControlsFlow } from "./pi-controls-flow.js";
import { runCursorNativeFlow } from "./cursor-native-flow.js";
import { createRemoteNativeBootstrap, createRemoteFixtureClient } from "./remote-native-bootstrap.js";
import { runCopilotProtectionFlow } from "./copilot-protection-flow.js";
import { runPiNativeFlow } from "./pi-native-flow.js";
import { seedPiFile, collectPiFileEvidence } from "./pi-file-evidence.js";
import { runPlanTaskFlow } from "./plan-task-flow.js";
import { assertNativeCompletionSelection, NATIVE_COMPLETION_PREFLIGHT_ENV, verifyNativeCompletionPreflight } from "./native-completion-admission.js";
import { assertNativeInstructionSelection, verifyNativeInstructionPreflight, NATIVE_INSTRUCTION_PREFLIGHT_ENV, NATIVE_INSTRUCTION_SUITE, NATIVE_INSTRUCTION_DEFAULT_SHA256 } from "./native-instruction-consolidation.js";
import { captureNativeDefault, gradeNativeDefault, nativeCompletionWorkspaceDigest } from "./native-completion-defaults.js";
import { gradeNativeCompletion, gradeNativeCompletionFinalAnswer } from "./native-completion-scoring.js";
import { assertNativeBlockerReply } from "./native-blocker-visible.js";
import { warmManagedFileEvidence } from "./warm-managed-files.js";
import { summarizeStartupReuse } from "../../packages/adapter-utils/src/startup-reuse.js";
import { gitFinalizationEvidence, gitStreamingEvidence, setupGitStreamingWorkspace } from "./daytona-git-streaming.js";
import { runsCompletionUpdateProbe, completionQualityControls, completionQualityStatus, judgeCompletionQuality, reserveCompletionQuality, type CompletionQualityRecord } from "./completion-quality.js";
import { runNativeProviderLossFlow } from "./native-provider-loss-flow.js";
import { mayAllocateRemoteResources, runCleanupWithObservers, verifyCleanupAssertions, type CleanupAssertion } from "./cleanup-verification.js";
import { completionDelivery, type CompletionObservation } from "./completion-updates.js";
import { runInstructionPersistenceFlow } from "./instruction-persistence.js";
import { gradeApiResponsePaging, readResponseProof, responseEvidenceDescription } from "./api-response-reading.js";
import { runBlockerFlow } from "./blocker-flow.js";
import { largeJournalEvidence } from "./journal-evidence.js";
import { observeBrowserBootstrap } from "./browser-bootstrap-diagnostics.js";
import { runAccountingFlow } from "./accounting-flow.js";
import type { Issue } from "../../packages/shared/src/types/issue.js";
import { lifecycleLiveCase, gradeLifecycleRepair } from "./lifecycle-live-cases.js";
import { runContinuationFlow } from "./continuation-flow.js";
import { runEverydayFlow } from "./everyday-flow.js";
import { gradeTaskTitle } from "./task-titles.js";
import { runContextIntegrityFlow } from "./context-integrity-flow.js";
import { runRepositorySkillsFlow } from "./repository-skills.js";
import { captureStockHarness, gradeStockHarness, gradeStockHire } from "./stock-harness.js";
import { verifyStockHarnessPreflight, STOCK_PREFLIGHT_ENV } from "./stock-harness-admission.js";
import { createTaskThroughUi, submitTaskReply } from "./user-actions.js";
import { armResponseTiming } from "./response-timing.js";
import { projectFirstTaskResult, runFirstTaskFlow, setupFirstTaskFixtures } from "./first-task-flow.js";
import { runChatFlow } from "./chat-flow.js";
import { restartChatServer } from "./chat-restart.js";
import { matchesRunCount, minimumRunCount } from "./run-count.js";
import { createHash, randomBytes } from "node:crypto";
import { mkdir, readFile, rename, writeFile } from "node:fs/promises";
import path from "node:path";
import { expect, test, type Page } from "@playwright/test";
import { RunnerApi, pollUntil } from "./api.js";
import { buildRuntimeUsage, summarizeExecutionBilling } from "./billing.js";
import { collectAttemptFeedback, emptyFeedbackSummary } from "./feedback.js";
import { createKimiLifecycleTiming, finishBudgetedEvalSetup, waitForEvalAgentReady } from "./budgeted-setup.js";
import { establishPublicMcpSession, runPublicMcpFlow } from "./public-mcp-flow.js";
import { assistantUsage } from "./public-mcp-model.js";
import { runnerExecutionById } from "./catalog.js";
import { classifyFailure } from "./failure-classifier.js";
import { runnerE2EServerControlPaths } from "./harness-env.js";
import { setupConnectionReview } from "./connection-reviews.js";
import { setupLiveFixtures, selectFixtureRuntimeLeases, fixtureRuntimeCostCoverage, waitForFixtureDaytonaProvider, type LiveFixtureValues } from "./live-fixtures.js";
import { evaluateMatcher, persistedFinalRunMessage, type MatcherResult } from "./matchers.js";
import { readRegisteredArtifacts } from "./registered-artifact.js";
import {
acceptedPlanSessionResetFailures,
collectRunEvents,
hasTerminalMalformedPlanConfirmation,
isControlPlaneGovernedResponseWait,
isNonExecutingReviewFenceRun,
isOpenRouterDeepSeekHelloTerminalVariance,
numberedPlanStepCount,
providerSessionContinuityFailures,
reasoningProjectionFailures,
} from "./run-observations.js";
import { resolveRunnerE2ESource } from "./source.js";
import { isValidNativePrpEnvelope } from "./native-event-envelope.js";
import {
isPublicRunnerScreenshotRoute,
PUBLIC_RUNNER_SCREENSHOT_MARKER,
} from "./screenshot-policy.js";
import {
assertSecretFree,
findSecretLeakInJsonValues,
normalizedSecrets,
sanitizeJson,
browserDiagnosticUrl,
} from "./redaction.js";
import {
CREDENTIAL_NAMES,
type CredentialName,
type FailureClass,
type RunnerE2EResult,
} from "./types.js";
import { assertRunnerE2EPrerequisites } from "./prerequisites.js";
interface IssueRecord {
id: string;
identifier?: string | null;
companyId: string;
title: string;
status: string;
statusVersion?: number;
workMode?: string;
assigneeAgentId?: string | null;
projectId?: string | null;
projectWorkspaceId?: string | null;
executionWorkspaceId?: string | null;
executionRunId?: string | null;
checkoutRunId?: string | null;
scheduledRetry?: unknown;
monitorNextCheckAt?: string | null;
}
interface CommentRecord {
id: string;
body?: string | null;
authorType?: string | null;
authorAgentId?: string | null;
authorUserId?: string | null;
createdByRunId?: string | null;
createdAt?: string;
}
interface RunRecord {
id: string;
companyId: string;
agentId: string;
status: string;
runtimeMode?: string;
nativeIssueId?: string | null;
continuationAttempt?: number;
retryOfRunId?: string | null;
runnerInstanceId?: string | null;
nativeSessionId?: string | null;
processPid?: number | null;
processStartedAt?: string | null;
contextSnapshot?: Record<string, unknown> | null;
runnerProfileJson?: Record<string, unknown> | null;
usageJson?: Record<string, unknown> | null;
resultJson?: Record<string, unknown> | null;
sessionIdBefore?: string | null;
sessionIdAfter?: string | null;
error?: string | null;
errorCode?: string | null;
startedAt?: string | null;
finishedAt?: string | null;
}
interface EnvironmentLeaseRecord {
id: string;
status?: string;
cleanupStatus?: string | null;
leasePolicy?: string;
providerLeaseId?: string | null;
executionWorkspaceId?: string | null;
metadata?: Record<string, unknown> | null;
issueId?: string | null;
heartbeatRunId?: string | null;
provider?: string | null;
acquiredAt?: string | null;
releasedAt?: string | null;
updatedAt?: string | null;
}
interface InteractionRecord {
id: string;
status: string;
kind?: string;
sourceRunId?: string | null;
continuationPolicy?: string;
payload?: {
version?: number;
acceptLabel?: string;
rejectLabel?: string;
target?: {
type?: string;
key?: string;
revisionId?: string;
revisionNumber?: number;
};
questions?: Array<{
id: string;
prompt: string;
selectionMode: "single" | "multi";
required?: boolean;
options: Array<{ id: string; label: string }>;
}>;
};
result?: {
version?: number;
answers?: Array<{
questionId?: string;
optionIds?: string[];
otherText?: string;
}>;
} | null;
}
interface IssueDocumentRecord {
id: string;
key: string;
body?: string | null;
latestRevisionId?: string | null;
latestRevisionNumber?: number;
}
interface RunEventRecord {
seq?: number;
eventType?: string;
payload?: Record<string, unknown> | null;
sourceInstanceId?: string | null;
sourceEventId?: string | null;
sourceSeq?: number | null;
protocolSchemaVersion?: number | null;
}
const TERMINAL_RUN_STATUSES = new Set([
"succeeded",
"interrupted",
"failed",
"cancelled",
"timed_out",
]);
const DEFINITIVE_FAILURE_RUN_STATUSES = new Set([
"failed",
"cancelled",
"timed_out",
]);
function definitiveRunFailure(runs: readonly RunRecord[]) {
const failed = runs.find((run) =>
DEFINITIVE_FAILURE_RUN_STATUSES.has(run.status),
);
if (!failed) return undefined;
return `heartbeat run ${failed.id} ended ${failed.status}${failed.errorCode ? ` (${failed.errorCode})` : ""}${failed.error ? `: ${failed.error}` : ""}`;
}
function chronologicalRunTime(run: RunRecord): number {
const value = run.startedAt ?? run.finishedAt;
if (!value) return 0;
const parsed = Date.parse(value);
return Number.isNaN(parsed) ? 0 : parsed;
}
function sortRunsChronologically(runs: readonly RunRecord[]): RunRecord[] {
return [...runs].sort(
(left, right) =>
chronologicalRunTime(left) - chronologicalRunTime(right) ||
left.id.localeCompare(right.id),
);
}
async function restartIsolatedPaperclipServer(input: {
api: RunnerApi;
requestId: string;
kimiCandidateAdmission?: import("../../packages/paperclip-runner/src/drivers/acpx/candidate-admission.js").AcpxCandidateAdmission;
deadlineAt: number;
preserveRunner?: RestartRunnerIdentity;
discardControllerCache?: boolean;
crashAfterResult?: FinalizationCrashBinding;
controllerDowntimeMs?: number;
controllerOutage?: ControllerOutage;
recoveryBinding?: import("./controller-recovery-observer.js").RecoveryBinding;
onControllerOutage?: (receipt: ControllerOutageReceipt) => Promise<void>;
}): Promise<void> {
const {
controlDirectory,
restartRequestPath: requestPath,
restartAcknowledgementPath: acknowledgementPath,
} = runnerE2EServerControlPaths(temporaryRoot!);
await mkdir(controlDirectory, { recursive: true });
const temporaryRequestPath = `${requestPath}.${process.pid}.${input.requestId}.tmp`;
await writeFile(
temporaryRequestPath,
JSON.stringify({ requestId: input.requestId, ...(input.kimiCandidateAdmission ? { kimiCandidateAdmission: input.kimiCandidateAdmission } : {}), ...(input.controllerDowntimeMs !== undefined && !input.controllerOutage ? { controllerDowntimeMs: input.controllerDowntimeMs } : {}), discardControllerCache: input.discardControllerCache === true,
...(input.controllerOutage ? { controllerOutage: input.controllerOutage } : {}),
...(input.recoveryBinding ? { recoveryBinding: input.recoveryBinding } : {}),
...(input.preserveRunner ? { preserveRunner: input.preserveRunner } : {}),
...(input.crashAfterResult ? { crashAfterResult: input.crashAfterResult } : {}) }),
{ encoding: "utf8", mode: 0o600 },
);
await rename(temporaryRequestPath, requestPath);
const acknowledgement = await pollUntil({
label: `isolated server restart ${input.requestId}`,
timeoutFailureClass: "transient_infrastructure",
deadlineAt: input.deadlineAt,
intervalMs: 250,
load: async () => {
try {
return JSON.parse(
await readFile(acknowledgementPath, "utf8"),
) as Record<string, unknown>;
} catch {
return {};
}
},
accept: (acknowledgement) =>
acknowledgement.requestId === input.requestId &&
acknowledgement.status === "ready" &&
(!input.discardControllerCache || acknowledgement.controllerCacheDiscarded === true) &&
(!input.controllerOutage || validControllerOutageReceipt(acknowledgement.controllerOutage, input.controllerOutage, input.requestId)),
reject: (acknowledgement) =>
acknowledgement.requestId === input.requestId &&
acknowledgement.status === "failed"
? String(acknowledgement.message ?? "replacement server failed")
: undefined,
});
await pollUntil({
label: `replacement server health ${input.requestId}`,
timeoutFailureClass: "transient_infrastructure",
deadlineAt: input.deadlineAt,
intervalMs: 250,
load: () => input.api.get<Record<string, unknown>>("/api/health"),
accept: (health) => Boolean(health),
});
if (input.controllerOutage) await input.onControllerOutage?.(acknowledgement.controllerOutage as ControllerOutageReceipt);
}
const executionIds = (() => {
const encoded = process.env.PAPERCLIP_RUNNER_E2E_EXECUTION_IDS;
if (encoded) {
const parsed = JSON.parse(encoded) as unknown;
if (
!Array.isArray(parsed) ||
parsed.length === 0 ||
parsed.some((value) => typeof value !== "string")
) {
throw new Error(
"PAPERCLIP_RUNNER_E2E_EXECUTION_IDS must be a non-empty JSON string array",
);
}
return parsed;
}
const single = process.env.PAPERCLIP_RUNNER_E2E_EXECUTION_ID;
if (!single)
throw new Error("PAPERCLIP_RUNNER_E2E_EXECUTION_IDS is required");
return [single];
})();
const executions = executionIds.map(runnerExecutionById);
// Direct Playwright invocation must enforce the same admission boundary as
// launch.ts before reading any provider credential from the environment.
assertRunnerE2EPrerequisites(executions);
const attempt = Number(process.env.PAPERCLIP_RUNNER_E2E_ATTEMPT ?? "1");
const temporaryRoot = process.env.PAPERCLIP_RUNNER_E2E_TEMP_ROOT;
const privateRoot = process.env.PAPERCLIP_RUNNER_E2E_PRIVATE_DIR;
const workspacePath = process.env.PAPERCLIP_RUNNER_E2E_WORKSPACE;
if (!temporaryRoot || !privateRoot || !workspacePath)
throw new Error("Runner E2E temporary/private/workspace paths are required");
function record(value: unknown): Record<string, unknown> {
return value && typeof value === "object" && !Array.isArray(value)
? (value as Record<string, unknown>)
: {};
}
function credentialValues(): Partial<Record<CredentialName, string>> {
return Object.fromEntries(
CREDENTIAL_NAMES.flatMap((name) => {
const value = process.env[name]?.trim();
return value ? [[name, value]] : [];
}),
);
}
async function writeSanitizedJson(
directory: string,
name: string,
value: unknown,
secrets: readonly string[],
signal?: AbortSignal,
) {
signal?.throwIfAborted();
const raw = `${JSON.stringify(value)}\n`;
// Public API payloads may contain harmless provider-shaped fixtures or
// opaque generated tokens. Reject exact campaign credentials before writing,
// then sanitize all known shapes and assert the published JSON is clean.
assertSecretFree(raw, secrets, name, { includeShapes: false });
const safe = `${JSON.stringify(sanitizeJson(value, secrets), null, 2)}\n`;
const leak = findSecretLeakInJsonValues(JSON.parse(safe), secrets);
if (leak) throw new Error(`Secret leak in ${name}: ${leak}`);
signal?.throwIfAborted();
await mkdir(directory, { recursive: true });
await writeFile(path.join(directory, name), safe, { encoding: "utf8", signal });
}
async function submitTaskRevision(page: Page, body: string): Promise<number> {
const revise = page
.getByRole("button", { name: "Continue work", exact: true })
.last();
await expect(revise).toBeVisible({ timeout: 30_000 });
await revise.click();
const reason = page.getByPlaceholder("Add a short note").last();
await expect(reason).toBeVisible({ timeout: 30_000 });
await reason.fill(body);
const submittedAtMs = Date.now();
await page
.getByRole("button", { name: "Continue work", exact: true })
.last()
.click();
return submittedAtMs;
}
function matchingRuns(runs: RunRecord[], issue: IssueRecord) {
const explicit = new Set(
[issue.executionRunId, issue.checkoutRunId].filter(Boolean),
);
return runs.filter((run) => {
if (isNonExecutingReviewFenceRun(run)) return false;
const context = record(run.contextSnapshot);
return (
context.issueId === issue.id ||
context.taskId === issue.id ||
explicit.has(run.id)
);
});
}
function isPendingQuestion(interaction: InteractionRecord) {
return (
interaction.kind === "ask_user_questions" &&
interaction.status === "pending" &&
Boolean(interaction.payload?.questions?.length)
);
}
function isPendingWarmConfirmation(interaction: InteractionRecord) {
return (
interaction.kind === "request_confirmation" &&
interaction.status === "pending" &&
interaction.continuationPolicy === "wake_assignee" &&
interaction.payload?.target?.type === "custom"
);
}
function nativeRunEventIntegrityFailures(
run: RunRecord,
events: readonly RunEventRecord[],
): string[] {
const failures: string[] = [];
let lastOuterSeq = 0;
const lastSourceSeq = new Map<string, number>();
const sourceEventIds = new Set<string>();
const runnerEventTypes: string[] = [];
const runTerminalSources: unknown[] = [];
const resultAcceptedSources: unknown[] = [];
for (const event of events) {
if (typeof event.seq !== "number" || event.seq <= lastOuterSeq) {
failures.push(
`run ${run.id} event sequence is not strictly monotonic at ${String(event.seq)}`,
);
} else {
lastOuterSeq = event.seq;
}
const envelope = record(event.payload?.prpEvent);
if (Object.keys(envelope).length === 0) continue;
if (!isValidNativePrpEnvelope(envelope, event.protocolSchemaVersion)) {
failures.push(
`run ${run.id} exposed a malformed PRP envelope (schema=${String(envelope.schema)}, version=${String(envelope.schemaVersion)})`,
);
}
if (envelope.runId !== run.id) {
failures.push(
`run ${run.id} exposed an event bound to ${String(envelope.runId)}`,
);
}
if (envelope.eventType !== event.eventType) {
failures.push(
`run ${run.id} event discriminator changed from ${String(event.eventType)} to ${String(envelope.eventType)}`,
);
}
if (
envelope.sourceInstanceId !== event.sourceInstanceId ||
envelope.sourceEventId !== event.sourceEventId ||
envelope.sourceSeq !== event.sourceSeq
) {
failures.push(
`run ${run.id} event source identity changed during persistence/redaction`,
);
}
if (typeof event.sourceEventId === "string") {
if (sourceEventIds.has(event.sourceEventId)) {
failures.push(
`run ${run.id} duplicated source event ${event.sourceEventId}`,
);
}
sourceEventIds.add(event.sourceEventId);
}
if (
typeof event.sourceInstanceId === "string" &&
typeof event.sourceSeq === "number"
) {
const previous = lastSourceSeq.get(event.sourceInstanceId) ?? 0;
if (event.sourceSeq <= previous) {
failures.push(
`run ${run.id} source ${event.sourceInstanceId} sequence regressed from ${previous} to ${event.sourceSeq}`,
);
}
lastSourceSeq.set(event.sourceInstanceId, event.sourceSeq);
}
if (
envelope.sourceKind === "runner" &&
typeof event.eventType === "string"
) {
runnerEventTypes.push(event.eventType);
}
if (event.eventType === "run.terminal") {
runTerminalSources.push(envelope.sourceKind);
}
if (event.eventType === "run.result.accepted") {
resultAcceptedSources.push(envelope.sourceKind);
}
}
const runnerResultCount = runnerEventTypes.filter(
(value) => value === "run.result.proposed",
).length;
if (
runnerResultCount !== 1 &&
!(runnerResultCount === 0 && isControlPlaneGovernedResponseWait(events))
) {
failures.push(
`run ${run.id} must persist exactly one runner semantic result`,
);
}
if (resultAcceptedSources.length !== 1) {
failures.push(
`run ${run.id} must persist exactly one accepted semantic result`,
);
} else if (resultAcceptedSources[0] !== "control_plane") {
failures.push(
`run ${run.id} accepted semantic result must be control-plane authoritative`,
);
}
if (runTerminalSources.length !== 1) {
failures.push(`run ${run.id} must persist exactly one terminal event`);
} else if (runTerminalSources[0] !== "control_plane") {
failures.push(
`run ${run.id} terminal event must be control-plane authoritative`,
);
}
return failures;
}
for (const execution of executions) {
const controllerOutage = controllerOutageForExecution(execution, process.env);
const privateDir = path.join(privateRoot, "cases", execution.task.id);
const resultPath = path.join(privateDir, "result.json");
const snapshotsDir = path.join(privateDir, "snapshots");
const deadlineMs = execution.task.attemptTimeoutMs[execution.environment.id];
test(`${execution.id} completes through the browser and public APIs`, async ({
page,
request,
}, testInfo) => {
test.skip(execution.suite.id === "first-task" && execution.profile.id === "runner-acpx-claude"
&& (process.platform !== "darwin" || process.arch !== "arm64"),
"Ordinary native Claude onboarding is qualified on Apple Silicon only in this delivery.");
const kimiLifecycleTiming = execution.task.flow === "kimi_native_lifecycle"
? createKimiLifecycleTiming({ startedAt: Date.now(), deadlineMs }) : undefined;
test.setTimeout(
kimiLifecycleTiming?.playwrightTimeoutMs ?? (execution.task.flow === "warm_three_turn"
? deadlineMs
: deadlineMs + 90_000),
);
// Remote allocation uncertainty survives a worker crash before publication.
if (mayAllocateRemoteResources(execution.environment.id)) {
await writeFile(path.join(privateRoot, "resource-admission-started"), "started\n", { mode: 0o600 });
}
const startedAtMs = Date.now();
const startedAt = new Date(startedAtMs).toISOString();
const workerTiming = await createWorkerTiming(path.join(snapshotsDir, "worker-timing.json"), { executionId: execution.id, attempt }, startedAtMs);
const nonce = `${randomBytes(6).toString("hex")}-${attempt}`;
const marker = execution.task.buildVisibleMarker(nonce);
const title = execution.task.buildTitle(nonce);
let prompt = execution.task.buildPrompt(nonce);
let apiResponseSourceId: string | undefined;
let titleCreation: { issue: unknown; submitted: unknown } | undefined;
let lifecycleBlockerId: string | null = null;
if (execution.suite.id === "native-completion") {
assertNativeCompletionSelection([execution]);
verifyNativeCompletionPreflight(process.env[NATIVE_COMPLETION_PREFLIGHT_ENV]);
}
if (execution.suite.id === NATIVE_INSTRUCTION_SUITE) {
assertNativeInstructionSelection([execution]);
verifyNativeInstructionPreflight(process.env[NATIVE_INSTRUCTION_PREFLIGHT_ENV]);
}
const credentials = credentialValues();
const secrets = normalizedSecrets([...Object.values(credentials), ...EVAL_OBSERVABILITY_SECRET_NAMES.map(name => process.env[name]), process.env.PAPERCLIP_E2E_GATEWAY_TOKEN]);
const traceConfig = evalBraintrustConfig(process.env);
const trace = traceConfig ? new EvalAttemptTrace(traceConfig, {
campaignId: process.env.PAPERCLIP_E2E_ACCOUNTING_CAMPAIGN_ID ?? "",
attemptId: process.env.PAPERCLIP_E2E_ACCOUNTING_ATTEMPT_ID ?? "",
executionId: execution.id, provider: execution.profile.provider, model: execution.profile.model, environment: execution.environment.id,
}, {
suiteId: execution.suite.id, suiteDefinitionHash: execution.suiteDefinitionHash,
profileId: execution.profile.id, caseId: execution.task.id, sourceRevision: resolveRunnerE2ESource().sha,
}) : undefined;
const saveTrace = async () => { if (trace) await writeSanitizedJson(snapshotsDir, "braintrust-trace.json", trace.snapshot(), secrets); };
const onCompanyCreated = async (id: string) => {
const admission = kimiCandidateAdmissionForExecution(execution, id);
if (admission) {
await restartIsolatedPaperclipServer({ api, requestId: `kimi-startup-${nonce}`,
kimiCandidateAdmission: admission, deadlineAt: startedAtMs + deadlineMs });
if (execution.environment.id === "daytona") {
await waitForFixtureDaytonaProvider(api, { companyId: id, deadlineAt: startedAtMs + deadlineMs });
}
}
if (trace) { try { await trace.configureCompany(api, id); } finally { await saveTrace(); } }
};
const api = new RunnerApi(request);
const companyRunFlow = execution.suite.id === "task-titles" || ["plan_task_guidance", "blocker_guidance", "continuation_accounting", "continuation", "context_integrity", "repository_skills", "agent_chat", "everyday_workflow", "first_task", "instruction_persistence", "pi_native", "pi_controls", "copilot_protection", "cursor_native", "native_active_stop", "native_provider_loss", "kimi_native_lifecycle", "public_mcp", "case_handoff"].includes(execution.task.flow);
const publicMcpUsage = execution.task.flow === "public_mcp" ? assistantUsage(execution.profile.provider === "claude" ? "anthropic" : "openai", execution.profile.model) : undefined;
let publicMcpUserId = "";
const consoleDiagnostics: Array<Record<string, unknown>> = [];
const networkDiagnostics: Array<Record<string, unknown>> = [];
const pageLifecycleDiagnostics: Array<Record<string, unknown>> = [];
const browserBootstrap = observeBrowserBootstrap(page);
let nativeInitial: { issueIds: string[]; agentIds: string[]; workspaceDigest: string | null } | undefined;
const nativeWorkspaceDigest = () => execution.task.id === "native-blocked-report"
? nativeCompletionWorkspaceDigest(workspacePath!) : Promise.resolve(null);
let fixtures: LiveFixtureValues | undefined;
let reviewProvider: Awaited<ReturnType<typeof setupConnectionReview>> | undefined;
let issue: IssueRecord | undefined;
let taskChat: TaskChatEvidence | undefined;
let workflowCase: { id: string; identifier?: string } | undefined;
let selectedRuns: RunRecord[] = [];
let feedback = emptyFeedbackSummary();
let runtimeLeases: EnvironmentLeaseRecord[] = [];
let matcherResults: MatcherResult[] = [];
let downloadedResponseProof: Awaited<ReturnType<typeof readResponseProof>> | undefined;
let piFileSeed: Awaited<ReturnType<typeof seedPiFile>> | undefined;
const completionQuality: CompletionQualityRecord[] = [];
const completionEvidence = async (name: string, data: unknown) => {
await writeSanitizedJson(snapshotsDir, name, data, secrets);
if (!["completion-updates", "confirmation-replies"].includes(execution.suite.id) || !name.endsWith("completion-update.json")) return;
const probe = data as { observation?: CompletionObservation };
if (!probe.observation || !completionDelivery(probe.observation).checks.every(c => c.passed)) return;
if (!credentials.OPENAI_API_KEY) {
await writeSanitizedJson(snapshotsDir, "completion-quality-unqualified.json", { reason: "Judge credential unavailable in this provider-scoped job; run the separate judge against retained evidence." }, secrets);
return;
}
const samples = [{ name, purpose: "product" as const, expectedPass: true, observation: probe.observation },
...(execution.profile.id === "runner-codex" && execution.task.id === "handoff-completion-idle" ? completionQualityControls(probe.observation).map(c => ({ ...c, purpose: "calibration" as const, name: `${name}-${c.name}` })) : [])];
for (const sample of samples) {
const pending = { ...reserveCompletionQuality(sample.observation, 0.50, secrets), name: sample.name, purpose: sample.purpose, expectedPass: sample.expectedPass };
completionQuality.push(pending);
// Persist reservation and exact input before the one paid request. A crash
// leaves usage unknown, never zero, and the next campaign is a new attempt.
await writeSanitizedJson(snapshotsDir, `${sample.name}.quality-input.json`, sample, secrets);
await writeSanitizedJson(snapshotsDir, "completion-quality-ledger.json", completionQuality, secrets);
const result = await judgeCompletionQuality(sample.observation, pending, credentials.OPENAI_API_KEY ?? "", undefined, { approvedFixture: true, secrets });
completionQuality[completionQuality.length - 1] = { ...result, name: sample.name, purpose: sample.purpose, expectedPass: sample.expectedPass };
await writeSanitizedJson(snapshotsDir, "completion-quality-ledger.json", completionQuality, secrets);
}
};
let firstTaskEvidence: RunnerE2EResult["firstTask"];
let turnTimings: NonNullable<RunnerE2EResult["turnTimings"]> | undefined;
let responseTiming: ReturnType<typeof armResponseTiming> | undefined;
let startupReuse: RunnerE2EResult["startupReuse"];
const turnSubmissionTimesMs: number[] = [];
const screenshots: NonNullable<RunnerE2EResult["screenshots"]> = [];
let primaryError: unknown;
let failureClassOverride: FailureClass | undefined;
let cleanup: RunnerE2EResult["cleanup"] = "not_started";
const cleanupAssertions: CleanupAssertion[] = [];
const beforeEnvironmentTeardownAssertions: CleanupAssertion[] = [];
const registerCleanupAssertion = (assertion: CleanupAssertion) => {
if (cleanupAssertions.length >= 8) throw new Error("Cleanup assertion bound exceeded");
cleanupAssertions.push(assertion);
};
const registerBeforeEnvironmentTeardownAssertion = (assertion: CleanupAssertion) => {
if (beforeEnvironmentTeardownAssertions.length >= 8) throw new Error("Remote cleanup assertion bound exceeded");
beforeEnvironmentTeardownAssertions.push(assertion);
};
const capturePrivateScreenshot = async (id: string, file: string) => {
const screenshotPath = path.join(privateDir, file);
await page.screenshot({ path: screenshotPath, fullPage: true });
await testInfo.attach(id, {
path: screenshotPath,
contentType: "image/png",
});
};
const isReviewedFixtureScreenshotRoute = () =>
isPublicRunnerScreenshotRoute(page.url(), {
chatAgentId: execution.task.flow === "agent_chat" ? fixtures?.agent.id : undefined,
issuePrefix: fixtures?.company.issuePrefix,
issueId: issue?.id,
issueIdentifier: issue?.identifier,
caseId: execution.task.flow === "case_handoff" ? workflowCase?.id : undefined,
caseIdentifier: execution.task.flow === "case_handoff" ? workflowCase?.identifier : undefined,
});
const captureScreenshot = async (
id: string,
label: string,
file: string,
) => {
if (!isReviewedFixtureScreenshotRoute()) {
throw new Error(
`Public runner screenshot blocked on non-task route ${page.url()}`,
);
}
await capturePrivateScreenshot(id, file);
screenshots.push({
id,
label,
file,
publication: PUBLIC_RUNNER_SCREENSHOT_MARKER,
sha256: createHash("sha256")
.update(await readFile(path.join(privateDir, file)))
.digest("hex"),
});
};
const captureRuntimeLeases = async () => {
if (!fixtures || execution.environment.id !== "daytona") return;
const listed = await api.get<EnvironmentLeaseRecord[]>(
`/api/environments/${fixtures.environment.id}/leases`,
);
runtimeLeases = selectFixtureRuntimeLeases({ leases: listed,
runIds: selectedRuns.map(run => run.id), issueId: issue?.id,
setupProbeLeaseId: fixtures.setupProbe?.leaseId, setupLeaseIds: fixtures.setupLeases?.map(lease => lease.leaseId) });
};
const cancelActiveRunsForCleanup = async () => {
if (!issue && !(companyRunFlow && fixtures)) return;
const cleanupIssueId = issue?.id;
const runs = await api.get<RunRecord[]>(
companyRunFlow && fixtures ? `/api/companies/${fixtures.company.id}/heartbeat-runs?limit=100` : `/api/issues/${cleanupIssueId}/runs`,
);
const activeRunIds = [
...new Set(
runs
.filter((run) => !TERMINAL_RUN_STATUSES.has(run.status))
.map((run) => run.id)
.filter(
(runId): runId is string =>
typeof runId === "string" && runId.length > 0,
),
),
];
for (const runId of activeRunIds) {
await api.post(`/api/heartbeat-runs/${runId}/cancel`);
}
if (activeRunIds.length === 0) return;
const activeIds = new Set(activeRunIds);
await pollUntil({
label: `cleanup cancellation for issue ${cleanupIssueId}`,
deadlineAt: Date.now() + 45_000,
load: () => api.get<RunRecord[]>(companyRunFlow && fixtures ? `/api/companies/${fixtures.company.id}/heartbeat-runs?limit=100` : `/api/issues/${cleanupIssueId}/runs`),
accept: (currentRuns) =>
currentRuns
.filter((run) => activeIds.has(run.id))
.every((run) => TERMINAL_RUN_STATUSES.has(run.status)),
intervalMs: 500,
});
};
const captureFailureApiState = async () => {
if (!fixtures) return;
if (execution.task.flow === "followup_stop_resume") {
const captured = await captureFollowupStopFailureApiState({ api, companyId: fixtures.company.id,
agentId: fixtures.agent.id, issueId: issue?.id,
evidence: (name, value) => writeSanitizedJson(snapshotsDir, name, value, secrets) });
if (captured) { issue = captured.issue as IssueRecord; selectedRuns = captured.runs as RunRecord[]; }
return;
}
if (companyRunFlow && !issue) {
const companyRuns = await api.get<RunRecord[]>(`/api/companies/${fixtures.company.id}/heartbeat-runs?limit=100`);
selectedRuns = await Promise.all(companyRuns.map(run => api.get<RunRecord>(`/api/heartbeat-runs/${run.id}`)));
// Settings are already restored on failure, so chat resolution may be
// gated. Direct task access still permits evidence and usage capture.
const sourceId = selectedRuns.map(run => record(run.contextSnapshot).issueId).find(id => typeof id === "string");
if (typeof sourceId === "string") issue = await api.get<IssueRecord>(`/api/issues/${sourceId}`);
}
if (!issue) return;
const capture = async <T>(operation: () => Promise<T>) =>
operation().catch((error) => ({
evidenceCaptureError:
error instanceof Error ? error.message : String(error),
}));
const [currentIssue, listedRuns, comments, interactions] =
await Promise.all([
capture(() => api.get<IssueRecord>(`/api/issues/${issue!.id}`)),
capture(() =>
api.get<RunRecord[]>(
companyRunFlow
? `/api/companies/${fixtures!.company.id}/heartbeat-runs?limit=100`
: `/api/companies/${fixtures!.company.id}/heartbeat-runs?agentId=${fixtures!.agent.id}&limit=20`,
),
),
capture(() =>
api.get<CommentRecord[]>(
`/api/issues/${issue!.id}/comments?order=asc`,
),
),
capture(() =>
api.get<InteractionRecord[]>(
`/api/issues/${issue!.id}/interactions`,
),
),
]);
const taskRuns = Array.isArray(listedRuns)
? companyRunFlow ? listedRuns : matchingRuns(listedRuns, "id" in currentIssue ? currentIssue : issue)
: [];
const detailedRuns = await Promise.all(
taskRuns.map((candidate) =>
capture(() =>
api.get<RunRecord>(`/api/heartbeat-runs/${candidate.id}`),
).then((value) => ("id" in value ? value : candidate)),
),
);
if (detailedRuns.length > 0) selectedRuns = detailedRuns;
const runEvidence = await Promise.all(
detailedRuns.map(async (candidate) => ({
runId: candidate.id,
workspaceOperations: await capture(() =>
api.get<unknown>(`/api/heartbeat-runs/${candidate.id}/workspace-operations`),
),
log: await capture(() =>
api.get<unknown>(
`/api/heartbeat-runs/${candidate.id}/log?limitBytes=1048576`,
),
),
events: await capture(() =>
collectRunEvents<RunEventRecord>((afterSeq, limit) =>
api.get(`/api/heartbeat-runs/${candidate.id}/events?afterSeq=${afterSeq}&limit=${limit}`),
),
),
})),
);
await writeSanitizedJson(
snapshotsDir,
"api-state.json",
{
capturePhase: "failure",
issue: currentIssue,
runs: detailedRuns,
comments,
interactions,
runEvidence,
},
secrets,
);
};
page.on("console", (message) => {
if (["error", "warning"].includes(message.type())) {
consoleDiagnostics.push({
type: message.type(),
text: message.text(),
location: { ...message.location(), url: browserDiagnosticUrl(message.location().url) },
});
}
});
page.on("pageerror", (error) => {
pageLifecycleDiagnostics.push({
type: "pageerror",
message: error.message,
stack: error.stack ?? null,
url: browserDiagnosticUrl(page.url()),
});
});
page.on("framenavigated", (frame) => {
if (frame === page.mainFrame())
pageLifecycleDiagnostics.push({
type: "navigation",
url: browserDiagnosticUrl(frame.url()),
at: new Date().toISOString(),
});
});
page.on("requestfailed", (requestEvent) => {
networkDiagnostics.push({
method: requestEvent.method(),
url: browserDiagnosticUrl(requestEvent.url()),
failure: requestEvent.failure()?.errorText ?? null,
});
});
page.on("response", (response) => {
if (response.status() >= 400) {
networkDiagnostics.push({
method: response.request().method(),
url: browserDiagnosticUrl(response.url()),
status: response.status(),
statusText: response.statusText(),
});
}
});
try {
trace?.phase("setup");
await saveTrace();
if (execution.task.flow === "public_mcp") publicMcpUserId = await establishPublicMcpSession(api, page, secrets);
if (execution.suite.id === "stock-harness") {
const receipt = verifyStockHarnessPreflight(process.env[STOCK_PREFLIGHT_ENV]);
await writeSanitizedJson(snapshotsDir, "stock-harness-preflight.json", receipt, secrets);
}
const nativeRolloutEnabled = execution.profile.provider !== "codex";
const experimental = await api.patch<{
enableNativeRunner: boolean;
}>("/api/instance/settings/experimental", {
enableNativeRunner: nativeRolloutEnabled,
...(["warm_three_turn", "everyday_workflow"].includes(execution.task.flow)
? { enableIsolatedWorkspaces: true }
: {}),
...(execution.profile.generation === "native" &&
execution.environment.id === "daytona"
? { enableRunnerPreviewIngress: true }
: {}),
});
expect(experimental.enableNativeRunner).toBe(nativeRolloutEnabled);
if (execution.suite.id === "extended-harnesses" && execution.profile.qualificationCandidate === "pi" && execution.task.id === "file-edit-validate") {
piFileSeed = await seedPiFile(workspacePath, nonce);
await writeSanitizedJson(snapshotsDir, "pi-file-seed.json", piFileSeed, secrets);
}
if (execution.suite.id === "daytona-git-streaming") {
await setupGitStreamingWorkspace(workspacePath);
}
fixtures = execution.task.flow === "first_task"
? await setupFirstTaskFixtures({ page, api, execution, nonce, credentials, daytonaImage: process.env.PAPERCLIP_E2E_DAYTONA_IMAGE, deadlineAt: startedAtMs + deadlineMs, observe: value => { fixtures = value; }, onCompanyCreated })
: await setupLiveFixtures({
api,
execution,
executionNonce: nonce,
workspacePath,
credentials,
daytonaImage: process.env.PAPERCLIP_E2E_DAYTONA_IMAGE,
cleanupDeadlineAt: kimiLifecycleTiming?.cleanupDeadlineAt,
onCompanyCreated,
});
await finishBudgetedEvalSetup({ executionId: execution.id, environmentId: execution.environment.id, agentId: fixtures.agent.id,
companyId: fixtures.company.id, deadlineAt: startedAtMs + deadlineMs, api,
observe: observations => writeSanitizedJson(snapshotsDir, "budgeted-setup.json", { observations }, secrets) });
const kimiWorkDeadlineAt = kimiLifecycleTiming ? await (async () => {
await waitForEvalAgentReady({ environmentId: execution.environment.id, agentId: fixtures.agent.id, companyId: fixtures.company.id,
deadlineAt: kimiLifecycleTiming.setupDeadlineAt, api,
observe: observations => writeSanitizedJson(snapshotsDir, "kimi-lifecycle-setup.json", { observations }, secrets) });
const readyAt = Date.now(), workDeadlineAt = kimiLifecycleTiming.beginWork(readyAt);
await writeSanitizedJson(snapshotsDir, "kimi-lifecycle-timing.json", { setupDeadlineAt: kimiLifecycleTiming.setupDeadlineAt,
readyAt, workDeadlineAt, setupBudgetMs: deadlineMs, workBudgetMs: deadlineMs, playwrightTimeoutMs: kimiLifecycleTiming.playwrightTimeoutMs }, secrets);
return workDeadlineAt;
})() : undefined;
await workerTiming.enter("execution");
trace?.phase("execution");
if (["native-completion", NATIVE_INSTRUCTION_SUITE].includes(execution.suite.id)) {
const receipt = await captureNativeDefault({ api, agentId: fixtures.agent.id, companyId: fixtures.company.id });
const grade = gradeNativeDefault(receipt, execution.suite.id === NATIVE_INSTRUCTION_SUITE ? NATIVE_INSTRUCTION_DEFAULT_SHA256 : undefined);
await writeSanitizedJson(snapshotsDir, "native-default-before-execution.json", { receipt, grade }, secrets);
if (!grade.passed) throw new Error("Native production-default hire admission failed before execution");
const [issues, agents, workspaceDigest] = await Promise.all([
api.get<Array<{ id: string }>>(`/api/companies/${fixtures.company.id}/issues?limit=100`),
api.get<Array<{ id: string }>>(`/api/companies/${fixtures.company.id}/agents`), nativeWorkspaceDigest(),
]);
nativeInitial = { issueIds: issues.map(value => value.id), agentIds: agents.map(value => value.id), workspaceDigest };
}
const remoteBootstrap = execution.environment.id === "daytona"
&& ["pi_native", "pi_controls", "copilot_protection", "cursor_native", "native_active_stop", "native_provider_loss", "kimi_native_lifecycle"].includes(execution.task.flow)
? createRemoteNativeBootstrap({
api, daytona: await createRemoteFixtureClient(credentials.DAYTONA_API_KEY ?? ""),
companyId: fixtures.company.id, environmentId: fixtures.environment.id, agentId: fixtures.agent.id,
image: process.env.PAPERCLIP_E2E_DAYTONA_IMAGE ?? "",
nodeSha256: process.env.PAPERCLIP_E2E_DAYTONA_NODE_SHA256 ?? "",
runnerdSha256: process.env.PAPERCLIP_E2E_DAYTONA_RUNNERD_SHA256 ?? "",
deadlineAt: kimiWorkDeadlineAt ?? startedAtMs + deadlineMs,
requireCopilotTaskContext: execution.profile.qualificationCandidate === "copilot",
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
}) : undefined;
if (execution.suite.id === "stock-harness") {
const hire = await captureStockHarness({ api, companyId: fixtures.company.id,
agentId: fixtures.agent.id, generation: execution.profile.generation, runIds: [] });
const checks = gradeStockHire(hire);
await writeSanitizedJson(snapshotsDir, "stock-harness-hire.json", {
capturePhase: "before-provider", ...hire, checks,
}, secrets);
const failed = checks.filter(check => !check.passed);
if (failed.length) throw new Error(`Stock hire matcher failures: ${failed.map(check => check.id).join(", ")}`);
}
if (execution.suite.id === "api-response-reading") {
const source = await api.post<{ id: string }>(`/api/companies/${fixtures.company.id}/issues`, {
title: `Synthetic diagnostic evidence ${nonce}`,
description: responseEvidenceDescription(nonce), status: "backlog",
});
apiResponseSourceId = source.id;
prompt = prompt.replaceAll("{{API_RESPONSE_SOURCE_ID}}", source.id);
}
if (execution.suite.id === "lifecycle-baseline" && lifecycleLiveCase(execution.task.id)?.family === "blocker") {
const prerequisite = await api.post<{ id: string }>(`/api/companies/${fixtures.company.id}/issues`, {
title: `Supply dataset ${nonce}`,
description: "Fixture operator prerequisite: dataset has not been supplied.",
status: "backlog",
});
lifecycleBlockerId = prerequisite.id;
prompt = prompt.replaceAll("{{LIFECYCLE_BLOCKER_ID}}", prerequisite.id);
}
await writeSanitizedJson(
snapshotsDir,
"fixtures.json",
{
executionId: execution.id,
lifecycleBlockerId,
companyId: fixtures.company.id,
environmentId: fixtures.environment.id,
agentId: fixtures.agent.id,
...(fixtures.setupProbe ? { setupProbe: { durationMs: fixtures.setupProbe.durationMs,
...(fixtures.setupProbe.defaultRunner ? { defaultRunner: fixtures.setupProbe.defaultRunner } : {}) } } : {}),
secretIds: Object.fromEntries(
Object.entries(fixtures.secretRefs).map(([name, ref]) => [
name,
ref?.secretId,
]),
),
persistedAgent: await api.get<unknown>(
`/api/agents/${fixtures.agent.id}`,
),
persistedEnvironment: await api.get<unknown>(
`/api/environments/${fixtures.environment.id}`,
),
},
secrets,
);
if (execution.task.flow === "case_handoff") {
const receiptSmoke = execution.task.id === "receipt-smoke";
const runFlow = receiptSmoke ? runCaseReceiptFlow : runCaseHandoffsFlow;
const journey = await runFlow({ page, api, fixtures, execution, nonce, workspacePath,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks, currentCase) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
workflowCase = { id: currentCase.id, identifier: currentCase.identifier };
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const,
path: `${receiptSmoke ? "caseReceipt" : "caseHandoffs"}.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
await writeSanitizedJson(snapshotsDir, "api-state.json", {
capturePhase: "completed",
issue: { id: issue.id, identifier: issue.identifier, status: issue.status,
assigneeAgentId: issue.assigneeAgentId, statusVersion: issue.statusVersion },
workflowCase,
runs: selectedRuns.map(candidate => ({ id: candidate.id, companyId: candidate.companyId,
agentId: candidate.agentId, status: candidate.status, runtimeMode: candidate.runtimeMode,
nativeIssueId: candidate.nativeIssueId, startedAt: candidate.startedAt,
finishedAt: candidate.finishedAt, errorCode: candidate.errorCode })),
matcherResults,
}, secrets);
await captureScreenshot("final-state", "Final visible task state", "final-state.png");
} else if (execution.task.flow === "followup_stop_resume") {
const journey = await runFollowupStopResumeFlow({ page, api, fixtures, execution, nonce, workspacePath: workspacePath!, temporaryRoot,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `followupStopResume.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
} else if (execution.task.flow === "finalization_recovery") {
const journey = await runFinalizationRecoveryFlow({ page, api, fixtures, execution, nonce, temporaryRoot,
deadlineAt: startedAtMs + deadlineMs - 30_000,
restart: binding => restartIsolatedPaperclipServer({ api, requestId: `finalization-${nonce}`, deadlineAt: startedAtMs + deadlineMs, crashAfterResult: binding }),
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `finalizationRecovery.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
} else if (execution.task.flow === "repository_verification") {
const journey = await runRepositoryVerificationFlow({ page, api, fixtures, execution, nonce, workspacePath: workspacePath!, temporaryRoot,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `repositoryVerification.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
} else if (execution.task.flow === "automatic_repository") {
const journey = await runAutomaticRepositoryFlow({ page, api, fixtures, execution, nonce, temporaryRoot,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `automaticRepository.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
} else if (execution.task.flow === "task_workspaces") {
const journey = await runTaskWorkspacesFlow({ page, api, fixtures, execution, nonce, temporaryRoot,
deadlineAt: startedAtMs + deadlineMs - 30_000,
restart: () => restartIsolatedPaperclipServer({ api, requestId: `task-workspaces-${nonce}`, deadlineAt: startedAtMs + deadlineMs }),
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `taskWorkspaces.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue as IssueRecord; selectedRuns = journey.runs as RunRecord[];
} else if (execution.task.flow === "public_mcp") {
const journey = await runPublicMcpFlow({
page, api, fixtures, execution, nonce, secrets, userId: publicMcpUserId,
credential: credentials[execution.profile.credential]!, usage: publicMcpUsage!, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, runs) => { issue = currentIssue; selectedRuns = runs; },
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = journey.issue; selectedRuns = journey.runs;
matcherResults = journey.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `publicMcp.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "plan_task_guidance") {
const planning = await runPlanTaskFlow({ page, api, fixtures, execution, nonce, workspacePath,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `planning.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
}, capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = planning.issue as IssueRecord; selectedRuns = planning.runs as RunRecord[];
} else if (execution.task.flow === "blocker_guidance") {
const blocker = await runBlockerFlow({ page, api, fixtures, execution, nonce, workspacePath,
deadlineAt: startedAtMs + deadlineMs - 30_000,
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue; selectedRuns = currentRuns;
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `blocker.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
},
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = blocker.issue as IssueRecord; selectedRuns = blocker.runs as RunRecord[];
} else if (execution.task.flow === "continuation_accounting") {
const accounting = await runAccountingFlow({
page, api, fixtures, execution, nonce, deadlineAt: startedAtMs + deadlineMs - 60_000,
restart: () => restartIsolatedPaperclipServer({ api, requestId: `accounting-${nonce}`, deadlineAt: startedAtMs + deadlineMs }),
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue; selectedRuns = currentRuns;
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `accounting.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
},
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = accounting.issue as IssueRecord; selectedRuns = accounting.runs as RunRecord[];
} else if (execution.task.flow === "pi_controls") {
const story = await runPiControlsFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `piControls.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "kimi_native_lifecycle") {
const story = await runKimiLifecycleFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: kimiWorkDeadlineAt!,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `kimiNativeLifecycle.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "native_provider_loss") {
const story = await runNativeProviderLossFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `nativeProviderLoss.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "native_active_stop") {
const story = await (execution.profile.qualificationCandidate === "copilot" ? runCopilotActiveStopFlow : runNativeActiveStopFlow)({
page, api, fixtures, execution: execution.profile.qualificationCandidate === "copilot" && execution.task.id === "pending-permission-stop" ? { ...execution, task: copilotActiveStopTasks[0]! } : execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `nativeActiveStop.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "copilot_protection") {
const story = await runCopilotProtectionFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot, evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `copilotProtection.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "cursor_native") {
const story = await runCursorNativeFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion, registerBeforeEnvironmentTeardownAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `cursorNative.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "pi_native") {
const story = await runPiNativeFlow({
page, api, fixtures, execution, nonce, workspacePath, deadlineAt: startedAtMs + deadlineMs,
restart: preserveRunner => restartIsolatedPaperclipServer({ api, requestId: `pi-native-${nonce}`, deadlineAt: startedAtMs + deadlineMs, preserveRunner }),
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
remoteBootstrap, registerCleanupAssertion: remoteBootstrap ? registerBeforeEnvironmentTeardownAssertion : registerCleanupAssertion,
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `piNative.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "instruction_persistence") {
const story = await runInstructionPersistenceFlow({
page, api, fixtures, execution, nonce, secrets, deadlineAt: startedAtMs + deadlineMs,
restart: () => restartIsolatedPaperclipServer({ api, requestId: `instructions-${nonce}`, deadlineAt: startedAtMs + deadlineMs }),
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `instructions.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "continuation") {
const continuation = await runContinuationFlow({
page, api, fixtures, execution, nonce, secrets, workspacePath, deadlineAt: startedAtMs + deadlineMs - 60_000,
restart: () => restartIsolatedPaperclipServer({ api, requestId: `continuation-${nonce}`, deadlineAt: startedAtMs + deadlineMs }),
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue; selectedRuns = currentRuns;
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `continuation.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
},
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = continuation.issue as IssueRecord; selectedRuns = continuation.runs as RunRecord[];
} else if (execution.task.flow === "everyday_workflow") {
const { readUpgradeArtifacts, upgradeArtifactKey } = await import("./upgrade-artifacts.js");
const upgradeManifest = controllerOutage ? await readUpgradeArtifacts(process.env[upgradeArtifactKey]) : null;
let restartIndex = 0;
const story = await runEverydayFlow({
page, api, fixtures, execution, nonce, workspacePath, privateDir,
deadlineAt: startedAtMs + deadlineMs,
restartCount: upgradeManifest ? upgradeManifest.sequence.length -
(execution.suite.definitionMetadata?.retainedControllerContinuation === true ? 2 : 1) : 1,
restart: async (recoveryBinding, controllerDowntimeMs) => {
let receipt: ControllerOutageReceipt | undefined;
restartIndex += 1;
await restartIsolatedPaperclipServer({ api, requestId: `story-${nonce}-${restartIndex}`, deadlineAt: startedAtMs + deadlineMs,
discardControllerCache: controllerOutage !== undefined,
controllerOutage: controllerOutage && !recoveryBinding ? { ...controllerOutage, holdMs: 1000 } : controllerOutage,
controllerDowntimeMs,
...(controllerOutage ? { recoveryBinding } : {}),
onControllerOutage: async value => {
receipt = value;
await writeSanitizedJson(snapshotsDir, `controller-outage-${restartIndex}.json`, value, secrets);
} });
return receipt;
},
observe: (storyIssue, storyRuns) => { issue = storyIssue; selectedRuns = storyRuns; },
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
});
issue = story.issue; selectedRuns = story.runs;
matcherResults = story.evidence.checks.map(check => ({ matcher: { kind: "json_path" as const, path: check.id, expected: true }, passed: check.passed, detail: check.detail }));
} else if (execution.task.flow === "repository_skills") {
const story = await runRepositorySkillsFlow({ page, api, fixtures, execution, nonce, credentials, secrets,
deadlineAt: startedAtMs + deadlineMs, capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
observe: (currentIssue, currentRuns) => { issue = currentIssue as IssueRecord; selectedRuns = currentRuns as RunRecord[]; },
});
issue = story.issue as IssueRecord; selectedRuns = story.runs as RunRecord[];
matcherResults = story.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `repositorySkills.${check.id}`, expected: true }, passed: check.passed, detail: check.id }));
} else if (execution.task.flow === "context_integrity") {
const contextIntegrity = await runContextIntegrityFlow({
page, api, fixtures, execution, nonce, deadlineAt: startedAtMs + deadlineMs,
capture: captureScreenshot,
evidence: (name, data) => writeSanitizedJson(snapshotsDir, name, data, secrets),
observe: (currentIssue, currentRuns, checks) => {
issue = currentIssue as unknown as IssueRecord;
selectedRuns = currentRuns as unknown as RunRecord[];
matcherResults = checks.map(check => ({ matcher: { kind: "json_path" as const, path: `contextIntegrity.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
},
});
issue = contextIntegrity.issue as unknown as IssueRecord;
selectedRuns = contextIntegrity.runs as unknown as RunRecord[];
} else if (execution.task.flow === "agent_chat") {
const chat = await runChatFlow({
page, api, fixtures, execution, nonce, workspacePath,
deadlineAt: startedAtMs + deadlineMs,
restart: () => restartChatServer(page, () => restartIsolatedPaperclipServer({ api, requestId: `chat-${nonce}`, deadlineAt: startedAtMs + deadlineMs })),
observe: (chatIssue, chatRuns) => { issue = chatIssue; selectedRuns = chatRuns; },
capture: captureScreenshot,
evidence: completionEvidence,
check: (id, passed, detail) => matcherResults.push({ matcher: { kind: "json_path", path: `chat.${id}`, expected: true }, passed, detail }),
});
issue = chat.issue; selectedRuns = chat.runs;
if (matcherResults.length === 0) matcherResults = [{ matcher: { kind: "issue_status", expected: "in_review" }, passed: true, detail: "Chat workflow and durable handoff/session assertions passed" }];
} else if (execution.task.flow === "first_task") {
const firstTask = await runFirstTaskFlow({
page, api, fixtures, execution, nonce, deadlineAt: startedAtMs + deadlineMs, secrets,
observe: (currentIssue, runs, evidence) => {
issue = currentIssue; selectedRuns = runs; firstTaskEvidence = evidence;
matcherResults = evidence.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `firstTask.checks.${check.id}`, expected: true }, passed: check.passed, detail: check.detail }));
},
createOrdinary: async (taskTitle, taskPrompt) => {
const createdTask = await createTaskThroughUi({ page, issuePrefix: fixtures!.company.issuePrefix!, agentName: fixtures!.agent.name, title: taskTitle, prompt: taskPrompt, workMode: "standard" });
const created = await pollUntil({ label: "ordinary UI-created task", deadlineAt: startedAtMs + deadlineMs,
load: async () => (await api.get<IssueRecord[]>(`/api/companies/${fixtures!.company.id}/issues?limit=100`)).find(row => row.id === createdTask.issueId), accept: row => Boolean(row) });
if (!created) throw new Error("Missing ordinary task");
return created;
},
capture: captureScreenshot,
evidence: completionEvidence,
});
issue = firstTask.issue as IssueRecord; selectedRuns = firstTask.runs as RunRecord[];
} else {
const issuePrefix = fixtures.company.issuePrefix;
if (!issuePrefix)
throw new Error(
"Created fixture company did not return an issue prefix",
);
if (execution.task.flow === "governed_tool_review") {
reviewProvider = await setupConnectionReview({ page, api, prefix: issuePrefix, companyId: fixtures.company.id, agentId: fixtures.agent.id, marker });
}
// Capture the actual creation response: a generated title can change before
// a later list query, and the initial provisional state is part of the oracle.
const titleCreationResponse = execution.suite.id === "task-titles"
? page.waitForResponse(response => response.request().method() === "POST"
&& new URL(response.url()).pathname === `/api/companies/${fixtures!.company.id}/issues`,
{ timeout: Math.max(1, startedAtMs + deadlineMs - Date.now()) })
.catch((cause: unknown) => new Error("Could not capture task creation response", { cause }))
: null;
const createdTask = await createTaskThroughUi({
page,
issuePrefix,
agentName: fixtures.agent.name,
title,
prompt,
workMode: execution.task.workMode,
projectName: fixtures.project?.name,
requireExplicitTitle: execution.suite.id === "task-titles" && Boolean(title),
beforeDispatch: execution.suite.id === "codie-latency-diagnostic"
? () => { responseTiming = armResponseTiming({ page }); } : undefined,
});
turnSubmissionTimesMs.push(createdTask.submittedAtMs);
const deadlineAt = startedAtMs + deadlineMs;
if (titleCreationResponse) {
const response = await titleCreationResponse;
if (response instanceof Error) throw response;
expect(response.status()).toBe(201);
issue = await response.json() as IssueRecord;
titleCreation = { issue, submitted: response.request().postDataJSON() };
await writeSanitizedJson(snapshotsDir, "task-title-creation.json", titleCreation, secrets);
} else issue = await pollUntil({
label: `UI-created issue ${title}`,
deadlineAt,
load: async () => {
const issues = await api.get<IssueRecord[]>(
`/api/companies/${fixtures!.company.id}/issues?limit=50`,
);
return issues.find((candidate) => candidate.id === createdTask.issueId);
},
accept: (candidate): candidate is IssueRecord => Boolean(candidate),
});
if (!issue) throw new Error(`Issue ${title} disappeared after creation`);
if (
issue.companyId !== fixtures.company.id ||
issue.assigneeAgentId !== fixtures.agent.id
) {
throw new Error(
"UI-created issue does not belong to the fixture company and agent",
);
}
if (issue.workMode !== execution.task.workMode) {
throw new Error(
`UI-created issue work mode was ${String(issue.workMode)}; expected ${execution.task.workMode}`,
);
}
if (fixtures.project) {
if (
issue.projectId !== fixtures.project.id ||
issue.projectWorkspaceId !== fixtures.project.primaryWorkspace?.id
) {
throw new Error(
`Warm task did not retain its project/workspace scope: ${JSON.stringify({ projectId: issue.projectId, projectWorkspaceId: issue.projectWorkspaceId })}`,
);
}
}
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
);
const loadTaskState = async () => {
const observeSettlement = responseTiming?.beginSettlement();
const [currentIssue, runs, comments, interactions] = await Promise.all([
api.get<IssueRecord>(`/api/issues/${issue!.id}`),
api.get<RunRecord[]>(
companyRunFlow
? `/api/companies/${fixtures!.company.id}/heartbeat-runs?limit=100`
: `/api/companies/${fixtures!.company.id}/heartbeat-runs?agentId=${fixtures!.agent.id}&limit=20`,
),
api.get<CommentRecord[]>(
`/api/issues/${issue!.id}/comments?order=asc`,
),
api.get<InteractionRecord[]>(`/api/issues/${issue!.id}/interactions`),
]);
observeSettlement?.({ present: currentIssue.status === execution.task.expectedTerminalState.issue
&& matchingRuns(runs, currentIssue).length >= minimumRunCount(execution.task)
&& matchingRuns(runs, currentIssue).every(run => TERMINAL_RUN_STATUSES.has(run.status)) });
return {
currentIssue,
taskRuns: matchingRuns(runs, currentIssue),
comments,
interactions,
};
};
const rejectPlanConfirmationPoll = (input: {
taskRuns: RunRecord[];
interactions: InteractionRecord[];
minimumRunCount: number;
}) => {
const runFailure = definitiveRunFailure(input.taskRuns);
if (runFailure) return runFailure;
if (
!hasTerminalMalformedPlanConfirmation({
runs: input.taskRuns,
interactions: input.interactions,
minimumRunCount: input.minimumRunCount,
})
) {
return undefined;
}
failureClassOverride = "provider_variance";
return "succeeded heartbeat run created a pending request_confirmation without a revision-bound Plan target";
};
let planLifecycleEvidence: Record<string, unknown> | null = null;
let questionLifecycleEvidence: Record<string, unknown> | null = null;
let warmLifecycleEvidence: Record<string, unknown> | null = null;
let expectedQuestionResolution: {
interactionId: string;
optionId: string;
} | null = null;
if (execution.task.flow === "governed_tool_review") {
await expect(page.getByRole("button", { name: "Approve & run", exact: true })).toBeVisible({ timeout: Math.max(1, deadlineAt - Date.now()) });
expect(reviewProvider!.invocationCount()).toBe(0);
await pollUntil({ label: "governed waiting turn", deadlineAt, load: loadTaskState, accept: state => state.taskRuns.length === 1 && state.taskRuns.every(run => TERMINAL_RUN_STATUSES.has(run.status)) });
await captureScreenshot("tool-review-pending", "Connection review awaiting a human", "tool-review-pending.png");
await page.getByRole("button", { name: "Dismiss Approve tool action" }).click();
await page.getByRole("button", { name: "Review request", exact: true }).click();
if (execution.task.toolReviewDecision === "restart") {
await restartIsolatedPaperclipServer({ api, requestId: `tool-review-${nonce}`, deadlineAt });
await page.reload();
}
if (execution.task.toolReviewDecision === "always") {
await page.getByRole("button", { name: "Approval options", exact: true }).click();
await page.getByRole("menuitem", { name: "Always allow", exact: true }).click();
} else {
await page.getByRole("button", { name: execution.task.toolReviewDecision === "decline" ? "Decline" : "Approve & run", exact: true }).click();
}
} else if (execution.task.flow === "plan_revision_acceptance") {
const planMarkers = execution.task.buildPlanMarkers?.(nonce);
const revisionRequest = execution.task.buildRevisionRequest?.(nonce);
if (!planMarkers || !revisionRequest) throw new Error(`Plan fixture ${execution.task.id} is missing lifecycle factories`);
planLifecycleEvidence = await runPlanRevisionAcceptance({ page, api, origin: api.baseURL,
issue, issuePrefix, native: execution.profile.expectedRuntimeMode === "native", planMarkers, revisionRequest,
deadlineAt, load: loadTaskState, reject: rejectPlanConfirmationPoll, capture: captureScreenshot });
} else if (execution.task.flow === "question_resume_completion") {
const expectedAnswer = execution.task.buildQuestionAnswer?.(nonce);
if (!expectedAnswer) {
throw new Error(
`Question fixture ${execution.task.id} is missing its answer factory`,
);
}
const pendingState = await pollUntil({
label: `pending user question for issue ${issue.id}`,
deadlineAt,
load: loadTaskState,
accept: ({ taskRuns, interactions }) =>
taskRuns.length >= 1 &&
taskRuns.every((run) => TERMINAL_RUN_STATUSES.has(run.status)) &&
interactions.some(isPendingQuestion),
reject: ({ taskRuns }) => definitiveRunFailure(taskRuns),
});
const questionInteractions = pendingState.interactions.filter(
(interaction) => interaction.kind === "ask_user_questions",
);
if (
questionInteractions.length !== 1 ||
!isPendingQuestion(questionInteractions[0]!)
) {
throw new Error(
`Expected exactly one pending question interaction; observed ${JSON.stringify(questionInteractions)}`,
);
}
const questionInteraction = questionInteractions[0]!;
const questions = questionInteraction.payload?.questions ?? [];
if (questionInteraction.payload?.version !== 1) {
throw new Error("Question interaction omitted payload version 1");
}
if (questions.length !== 1) {
throw new Error(
`Expected exactly one question; observed ${questions.length}`,
);
}
const question = questions[0]!;
if (
question.id !== "verification-word" ||
question.prompt !== "Choose the verification word." ||
question.selectionMode !== "single" ||
question.required !== true ||
JSON.stringify(question.options) !==
JSON.stringify([
{ id: "cobalt", label: "Cobalt" },
{ id: "amber", label: "Amber" },
])
) {
throw new Error(
`Question interaction did not preserve the exact verification contract: ${JSON.stringify(question)}`,
);
}
const selectedOption = question.options.find(
(option) => option.label === expectedAnswer.optionLabel,
);
if (!selectedOption) {
throw new Error(
`Question interaction omitted ${expectedAnswer.optionLabel}`,
);
}
expectedQuestionResolution = {
interactionId: questionInteraction.id,
optionId: selectedOption.id,
};
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
{ waitUntil: "domcontentloaded" },
);
await expect(
page
.getByRole("radio", {
name: expectedAnswer.optionLabel,
exact: true,
})
.last(),
).toBeVisible({ timeout: 30_000 });
await captureScreenshot(
"question-pending",
"Structured question awaiting an answer",
"question-pending.png",
);
if (execution.task.restartServerBeforeQuestionAnswer) {
const restartRequestId = `question-wait-${nonce}`;
await restartIsolatedPaperclipServer({
api,
requestId: restartRequestId,
deadlineAt,
});
const documentSentinel = `__paperclip_runner_restart_${nonce.replaceAll("-", "_")}`;
await page.evaluate(
(key) => Reflect.set(window, key, true),
documentSentinel,
);
try {
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
// The replacement Vite server can commit and render a fresh
// document while its navigation lifecycle remains unsettled.
// The sentinel and explicit assertions below prove the new
// document and durable state even when Playwright times out.
{ waitUntil: "commit" },
);
} catch (error) {
if (!(error instanceof Error) || error.name !== "TimeoutError") {
throw error;
}
}
await expect(
page
.getByRole("radio", {
name: expectedAnswer.optionLabel,
exact: true,
})
.last(),
).toBeVisible({ timeout: 30_000 });
expect(
await page.evaluate(
(key) => Reflect.get(window, key) === true,
documentSentinel,
),
).toBe(false);
const reloadedInteractions = await api.get<InteractionRecord[]>(
`/api/issues/${issue.id}/interactions`,
);
const reloadedQuestions = reloadedInteractions.filter(
(interaction) => interaction.kind === "ask_user_questions",
);
if (
reloadedQuestions.length !== 1 ||
reloadedQuestions[0]?.id !== questionInteraction.id ||
!isPendingQuestion(reloadedQuestions[0])
) {
throw new Error(
`Server restart did not preserve the exact pending interaction ${questionInteraction.id}: ${JSON.stringify(reloadedQuestions)}`,
);
}
await captureScreenshot(
"question-pending-after-server-restart",
"Structured question preserved across server restart",
"question-pending-after-server-restart.png",
);
}
await page
.getByRole("radio", {
name: expectedAnswer.optionLabel,
exact: true,
})
.last()
.check();
// The one-question form is on its last page, so selecting the radio
// records the answer and the existing form button submits it.
await page
.getByRole("button", { name: "Submit answers", exact: true })
.last()
.click();
questionLifecycleEvidence = {
interaction: questionInteraction,
answer: expectedAnswer.optionLabel,
expectedMarker: expectedAnswer.expectedMarker,
serverRestartedBeforeAnswer:
execution.task.restartServerBeforeQuestionAnswer ?? false,
};
} else if (execution.task.flow === "plan_approval_completion") {
const planMarkers = execution.task.buildPlanMarkers?.(nonce);
if (!planMarkers) {
throw new Error(
`Plan fixture ${execution.task.id} is missing its marker factory`,
);
}
const pendingState = await pollUntil({
label: `pending plan approval for issue ${issue.id}`,
deadlineAt,
load: loadTaskState,
accept: ({ taskRuns, interactions }) =>
taskRuns.length >= 1 &&
taskRuns.every((run) => TERMINAL_RUN_STATUSES.has(run.status)) &&
interactions.some(isPendingPlanConfirmation),
reject: ({ taskRuns, interactions }) =>
rejectPlanConfirmationPoll({
taskRuns,
interactions,
minimumRunCount: 1,
}),
});
const interaction = pendingState.interactions.find(
isPendingPlanConfirmation,
)!;
const plan = await api.get<IssueDocumentRecord>(
`/api/issues/${issue.id}/documents/plan`,
);
if (!normalizePlanMarkdown(plan.body).includes(planMarkers.draft)) {
throw new Error(`Plan document did not contain ${planMarkers.draft}`);
}
if (numberedPlanStepCount(plan.body) !== 2) {
throw new Error("Plan must contain exactly two numbered steps");
}
if (interaction.payload?.target?.revisionId !== plan.latestRevisionId) {
throw new Error(
"Plan confirmation did not target the latest Plan revision",
);
}
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
{ waitUntil: "domcontentloaded" },
);
await expectPlanStageVisible(page, {
native: execution.profile.expectedRuntimeMode === "native",
revision: plan.latestRevisionNumber ?? null,
});
await captureScreenshot(
"plan-pending",
"Plan awaiting approval",
"plan-pending.png",
);
await page
.getByRole("button", {
name: interaction.payload?.acceptLabel ?? "Approve",
exact: true,
})
.last()
.click();
planLifecycleEvidence = { interaction, plan };
} else if (execution.task.flow === "warm_three_turn") {
const followups = execution.task.buildFollowupMessages?.(nonce);
if (!followups || !fixtures.project?.primaryWorkspace?.id) {
throw new Error(
`Warm fixture ${execution.task.id} is missing its project or follow-up messages`,
);
}
const workspaceFile = path.join(
workspacePath,
`daytona-warm-${nonce}.txt`,
);
const turnEvidence: Array<Record<string, unknown>> = [];
for (const completedTurn of [1, 2] as const) {
const turnDeadlineAt = Math.min(
deadlineAt,
turnSubmissionTimesMs[completedTurn - 1]! +
(execution.task.turnTimeoutMs ?? 10 * 60_000),
);
const waitingState = await pollUntil({
label: `warm ${execution.environment.id} turn ${completedTurn} review state for issue ${issue.id}`,
deadlineAt: turnDeadlineAt,
load: loadTaskState,
accept: ({ currentIssue, taskRuns, interactions }) => {
const latestRun = sortRunsChronologically(taskRuns).at(-1);
const pendingConfirmations = interactions.filter(
isPendingWarmConfirmation,
);
return (
currentIssue.status === "in_review" &&
taskRuns.length === completedTurn &&
taskRuns.every((run) => run.status === "succeeded") &&
pendingConfirmations.length === 1 &&
pendingConfirmations[0]?.sourceRunId === latestRun?.id
);
},
reject: ({ taskRuns }) =>
definitiveRunFailure(taskRuns) ??
(taskRuns.length > completedTurn
? `warm turn ${completedTurn} dispatched duplicate runs`
: undefined),
});
if (execution.suite.id === "daytona-git-streaming") {
await writeSanitizedJson(snapshotsDir, `git-copyback-turn-${completedTurn}.json`, await gitStreamingEvidence(workspacePath, completedTurn), secrets);
const finalization = await gitFinalizationEvidence(api, issue.id, waitingState.taskRuns.map(run => run.id));
await writeSanitizedJson(snapshotsDir, `git-finalization-turn-${completedTurn}.json`, finalization, secrets);
expect(finalization.passed, finalization.failures.join("; ")).toBe(true);
}
if (execution.suite.id === "daytona-warm-continuity") {
const evidence = await warmManagedFileEvidence(api, fixtures.agent.id, sortRunsChronologically(waitingState.taskRuns).at(-1)!.id, completedTurn, nonce, execution.profile.generation === "native");
await writeSanitizedJson(snapshotsDir, `managed-warm-turn-${completedTurn}.json`, evidence, secrets);
expect(evidence.passed, JSON.stringify(evidence.checks)).toBe(true);
}
const expectedPrefix = `${Array.from(
{ length: completedTurn },
(_, index) => `T${index + 1}-${nonce}`,
).join("\n")}\n`;
const hostContent = await readFile(workspaceFile, "utf8");
if (hostContent !== expectedPrefix) {
throw new Error(
`Host workspace was not finalized after warm turn ${completedTurn}: expected ${JSON.stringify(expectedPrefix)}, observed ${JSON.stringify(hostContent)}`,
);
}
if (
waitingState.currentIssue.projectId !== fixtures.project.id ||
waitingState.currentIssue.projectWorkspaceId !==
fixtures.project.primaryWorkspace.id ||
!waitingState.currentIssue.executionWorkspaceId
) {
throw new Error(
`Warm turn ${completedTurn} lost its project execution-workspace scope`,
);
}
const chronologicalRuns = sortRunsChronologically(
waitingState.taskRuns,
);
const completedRunIds = new Set(
chronologicalRuns.map((candidate) => candidate.id),
);
let completedLeases: EnvironmentLeaseRecord[] = [];
if (execution.environment.id === "daytona") {
const retainedTurnLeases = await pollUntil({
label: `retained Daytona leases after warm turn ${completedTurn}`,
deadlineAt: Math.min(turnDeadlineAt, Date.now() + 30_000),
intervalMs: 500,
load: () =>
api.get<EnvironmentLeaseRecord[]>(
`/api/environments/${fixtures!.environment.id}/leases`,
),
accept: (leases) => {
const runOrder = new Map(
chronologicalRuns.map((candidate, index) => [
candidate.id,
index,
]),
);
const completed = leases
.filter(
(lease) =>
lease.heartbeatRunId &&
completedRunIds.has(lease.heartbeatRunId),
)
.sort(
(left, right) =>
(runOrder.get(left.heartbeatRunId ?? "") ?? 0) -
(runOrder.get(right.heartbeatRunId ?? "") ?? 0),
);
return (
completed.length === completedTurn &&
completed
.slice(0, -1)
.every(
(lease) =>
lease.status === "expired" &&
lease.cleanupStatus === "success",
) &&
completed.at(-1)?.status === "retained" &&
completed.every(
(lease) =>
lease.leasePolicy === "reuse_by_environment" &&
typeof lease.providerLeaseId === "string" &&
record(lease.metadata).sandboxState === "started",
) &&
completed
.slice(1)
.every(
(lease) =>
record(lease.metadata).resumedFromState === "started",
)
);
},
});
const turnRunOrder = new Map(
chronologicalRuns.map((candidate, index) => [candidate.id, index]),
);
completedLeases = retainedTurnLeases
.filter(
(lease) =>
lease.heartbeatRunId &&
completedRunIds.has(lease.heartbeatRunId),
)
.sort(
(left, right) =>
(turnRunOrder.get(left.heartbeatRunId ?? "") ?? 0) -
(turnRunOrder.get(right.heartbeatRunId ?? "") ?? 0),
);
if (
new Set(completedLeases.map((lease) => lease.providerLeaseId))
.size !== 1
) {
throw new Error(
`Warm turn ${completedTurn} replaced its Daytona sandbox`,
);
}
}
const journalEvidence = execution.suite.id === "daytona-journal-continuity" && completedTurn === 1
? await largeJournalEvidence({
stateRoot: path.join(process.env.PAPERCLIP_HOME!, "instances", process.env.PAPERCLIP_INSTANCE_ID!, "runtime", "paperclip-runner", "durable-sessions"),
runId: chronologicalRuns.at(-1)!.id,
minimumBytes: 2 * 1024 * 1024,
minimumCompletedStimulusCalls: 240,
})
: undefined;
if (journalEvidence) {
await writeSanitizedJson(snapshotsDir, "large-journal-boundary.json", journalEvidence, secrets);
}
turnEvidence.push({
...(journalEvidence ? { journalEvidence } : {}),
turn: completedTurn,
issue: waitingState.currentIssue,
run: chronologicalRuns.at(-1),
hostContent,
leases: completedLeases,
});
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
{ waitUntil: "domcontentloaded" },
);
// Navigation commits before React has loaded the task. Bind the
// screenshot to this turn's visible reply and exact pending review,
// rather than capturing a spinner after DOMContentLoaded.
const reviewUiTimeout = () => {
const remaining = turnDeadlineAt - Date.now();
if (remaining <= 0) throw new Error(`Warm turn ${completedTurn} UI deadline elapsed`);
return Math.min(30_000, remaining);
};
await expect(page.getByTestId("issue-detail-header").getByRole("button", {
name: "Change status (current: In Review)", exact: true,
})).toBeVisible({ timeout: reviewUiTimeout() });
const turnReply = page.getByTestId("task-chat-thread")
.getByTestId("task-chat-agent-bubble")
.filter({ hasText: `PAPERCLIP_E2E_WARM_T${completedTurn}_${nonce}` }).last();
await expect(turnReply).toBeVisible({ timeout: reviewUiTimeout() });
const pendingReview = waitingState.interactions.find(isPendingWarmConfirmation)!;
const reviewCard = page.locator(`[id=${JSON.stringify(`interaction-${pendingReview.id}`)}]`);
await expect(reviewCard).toBeVisible({ timeout: reviewUiTimeout() });
await expect(reviewCard.getByRole("button", { name: "Continue work", exact: true }))
.toBeEnabled({ timeout: reviewUiTimeout() });
await expect(page.getByTestId("task-chat-history-loading"))
.toHaveCount(0, { timeout: reviewUiTimeout() });
await captureScreenshot(
`warm-turn-${completedTurn}`,
`Warm ${execution.environment.id} turn ${completedTurn} awaiting review`,
`warm-turn-${completedTurn}.png`,
);
turnSubmissionTimesMs.push(
await submitTaskRevision(page, followups[completedTurn - 1]),
);
}
warmLifecycleEvidence = { turns: turnEvidence };
}
const taskMatchers = execution.task.buildMatchers(nonce, execution);
const terminalDeadlineAt =
execution.task.flow === "warm_three_turn"
? Math.min(
deadlineAt,
turnSubmissionTimesMs.at(-1)! +
(execution.task.turnTimeoutMs ?? 10 * 60_000),
)
: deadlineAt;
let terminal = await pollUntil({
label: `issue ${issue.id} and heartbeat run terminal state`,
deadlineAt: terminalDeadlineAt,
load: loadTaskState,
accept: ({ currentIssue, taskRuns }) =>
currentIssue.status === execution.task.expectedTerminalState.issue &&
taskRuns.length >= minimumRunCount(execution.task) &&
taskRuns.every((run) => TERMINAL_RUN_STATUSES.has(run.status)),
reject: ({ taskRuns }) => definitiveRunFailure(taskRuns),
});
// Finalization commits the issue/run decision before the derived agent
// comment is guaranteed to be visible through the comments endpoint.
// Give that projection a short consistency window so a successful
// terminal response is not misclassified as an empty provider reply.
// If no comment arrives, retain the original terminal observation and
// let the ordinary message matcher report the product failure.
if (taskMatchers.some((matcher) => matcher.kind.startsWith("message_"))) {
terminal = await pollUntil({
label: `final agent comment for issue ${issue.id}`,
deadlineAt: Math.min(deadlineAt, Date.now() + 30_000),
load: async () => {
const state = await loadTaskState();
const finalRun = sortRunsChronologically(state.taskRuns).at(-1);
if (!finalRun) return state;
// The compact run list omits the native presentation receipt.
// Poll its public detail together with comments so an earlier
// deliverable-preparation comment cannot settle this wait.
const detailed = await api.get<RunRecord>(`/api/heartbeat-runs/${finalRun.id}`);
return { ...state, taskRuns: state.taskRuns.map(run => run.id === detailed.id ? detailed : run) };
},
accept: ({ taskRuns, comments }) => {
if (!matchesRunCount(execution.task, taskRuns.length)) {
return false;
}
const finalRun = sortRunsChronologically(taskRuns).at(-1);
return Boolean(finalRun && persistedFinalRunMessage(
comments.filter(comment => comment.authorAgentId === fixtures!.agent.id), finalRun,
).trim());
},
reject: ({ taskRuns }) => definitiveRunFailure(taskRuns),
}).catch(() => terminal);
}
issue = terminal.currentIssue;
selectedRuns = terminal.taskRuns;
if (lifecycleBlockerId && execution.profile.expectedRuntimeMode === "legacy") {
const blockedIssue = await api.get<Pick<Issue, "blockedBy">>(`/api/issues/${issue.id}`);
expect(blockedIssue.blockedBy?.map((blocker) => blocker.id)).toEqual([lifecycleBlockerId]);
expect((await api.get<{ status: string }>(`/api/issues/${lifecycleBlockerId}`)).status).toBe("backlog");
}
if (reviewProvider) {
expect(reviewProvider.invocationCount()).toBe(execution.task.toolReviewDecision === "decline" ? 0 : execution.task.toolReviewDecision === "always" ? 2 : 1);
const pending = await api.get<{ actionRequests: unknown[] }>(`/api/companies/${fixtures.company.id}/tools/action-requests?status=pending`);
expect(pending.actionRequests).toHaveLength(0);
await writeSanitizedJson(snapshotsDir, "connection-review.json", { source: "local MCP fixture", connectionId: reviewProvider.connectionId, providerCalls: reviewProvider.invocationCount(), decision: execution.task.toolReviewDecision, issueId: issue.id }, secrets);
}
if (!matchesRunCount(execution.task, selectedRuns.length)) {
const runLogs = await Promise.all(
selectedRuns.map(async (candidate) => ({
runId: candidate.id,
log: await api
.get<unknown>(
`/api/heartbeat-runs/${candidate.id}/log?limitBytes=1048576`,
)
.catch((error) => ({
evidenceCaptureError:
error instanceof Error ? error.message : String(error),
})),
})),
);
await writeSanitizedJson(
snapshotsDir,
"api-state.json",
{ issue, runs: selectedRuns, runLogs, ...terminal },
secrets,
);
throw new Error(
`Expected ${minimumRunCount(execution.task)}..${execution.task.expectedRunCount} task heartbeat run(s); observed ${selectedRuns.length}`,
);
}
// The company run-list endpoint intentionally returns only a compact,
// allowlisted context summary. Hydrate each selected run through the
// public detail endpoint before asserting environment/lease metadata.
selectedRuns = await Promise.all(
selectedRuns.map((candidate) =>
api.get<RunRecord>(`/api/heartbeat-runs/${candidate.id}`),
),
);
selectedRuns = sortRunsChronologically(selectedRuns);
if (execution.suite.id === "daytona-git-streaming") {
await writeSanitizedJson(snapshotsDir, "git-copyback-turn-3.json", await gitStreamingEvidence(workspacePath, 3), secrets);
const finalization = await gitFinalizationEvidence(api, issue.id, selectedRuns.map(run => run.id));
await writeSanitizedJson(snapshotsDir, "git-finalization-turn-3.json", finalization, secrets);
expect(finalization.passed, finalization.failures.join("; ")).toBe(true);
}
const lifecycleProbe = execution.suite.id === "lifecycle-baseline" ? lifecycleLiveCase(execution.task.id) : undefined;
if (lifecycleProbe?.family === "repair") {
const grade = gradeLifecycleRepair({ runs: selectedRuns, comments: terminal.comments,
agentId: fixtures.agent.id, narrative: lifecycleProbe.narrative });
await writeSanitizedJson(snapshotsDir, "lifecycle-repair.json", { grade, runs: selectedRuns, comments: terminal.comments }, secrets);
expect(grade.passed, grade.detail).toBe(true);
expect(terminal.currentIssue.scheduledRetry).toBeNull();
expect(terminal.currentIssue.monitorNextCheckAt).toBeNull();
expect(terminal.interactions.filter(i => i.status === "pending")).toHaveLength(0);
}
if (execution.task.flow === "warm_three_turn") {
turnTimings = selectedRuns.map((candidate, index) => {
const submittedAtMs = turnSubmissionTimesMs[index]!;
const runStartedAtMs = candidate.startedAt
? Date.parse(candidate.startedAt)
: Number.NaN;
const runFinishedAtMs = candidate.finishedAt
? Date.parse(candidate.finishedAt)
: Number.NaN;
const acquisition = record(
record(candidate.contextSnapshot).paperclipEnvironment,
).sandboxLeaseAcquisition;
const acquisitionOutcome = record(acquisition).outcome;
return {
turn: index + 1,
submittedAt: new Date(submittedAtMs).toISOString(),
runStartedAt: candidate.startedAt ?? null,
runFinishedAt: candidate.finishedAt ?? null,
schedulerLatencyMs: Number.isFinite(runStartedAtMs)
? Math.max(0, runStartedAtMs - submittedAtMs)
: null,
runDurationMs:
Number.isFinite(runStartedAtMs) &&
Number.isFinite(runFinishedAtMs)
? Math.max(0, runFinishedAtMs - runStartedAtMs)
: null,
responseLatencyMs: Number.isFinite(runFinishedAtMs)
? Math.max(0, runFinishedAtMs - submittedAtMs)
: null,
runId: candidate.id,
leaseAcquisitionOutcome:
acquisitionOutcome === "created" ||
acquisitionOutcome === "resumed" ||
acquisitionOutcome === "replacement"
? acquisitionOutcome
: "unknown",
};
});
}
if (execution.suite.id === "daytona-warm-continuity") {
const evidence = await warmManagedFileEvidence(api, fixtures.agent.id, selectedRuns.at(-1)!.id, 3, nonce, execution.profile.generation === "native");
await writeSanitizedJson(snapshotsDir, "managed-warm-turn-3.json", evidence, secrets);
expect(evidence.passed, JSON.stringify(evidence.checks)).toBe(true);
}
const finalRun = selectedRuns.at(-1)!;
const run =
selectedRuns.find(
(candidate) => candidate.id === issue!.executionRunId,
) ?? selectedRuns[0];
const [
persistedAgentValue,
persistedEnvironmentValue,
runLogs,
runEventsByRun,
] = await Promise.all([
api.get<unknown>(`/api/agents/${fixtures.agent.id}`),
api.get<unknown>(`/api/environments/${fixtures.environment.id}`),
Promise.all(
selectedRuns.map(async (candidate) => ({
runId: candidate.id,
log: await api
.get<unknown>(
`/api/heartbeat-runs/${candidate.id}/log?limitBytes=1048576`,
)
.catch((error) => ({
evidenceCaptureError:
error instanceof Error ? error.message : String(error),
})),
})),
),
Promise.all(
selectedRuns.map(async (candidate) => {
try {
return {
runId: candidate.id,
events: await collectRunEvents<RunEventRecord>((afterSeq, limit) =>
api.get(`/api/heartbeat-runs/${candidate.id}/events?afterSeq=${afterSeq}&limit=${limit}`),
),
error: null,
};
} catch (error) {
return {
runId: candidate.id,
events: [] as RunEventRecord[],
error: error instanceof Error ? error.message : String(error),
};
}
}),
),
]);
const runLog =
runLogs.find((candidate) => candidate.runId === run.id)?.log ?? {};
const persistedAgent = record(persistedAgentValue);
const persistedEnvironment = record(persistedEnvironmentValue);
const agentComments = terminal.comments.filter(
(comment) =>
comment.authorAgentId === fixtures!.agent.id ||
selectedRuns.some(
(candidate) => comment.createdByRunId === candidate.id,
),
);
// Message matchers intentionally use persisted agent comments only.
// A semantic finish summary can differ from the actual user-facing text;
// accepting it here would let backend metadata mask a truncated UI
// response. The browser assertion below remains the visible source of
// truth after the comment projection has settled.
const message = agentComments
.map((comment) => comment.body ?? "")
.join("\n");
const finalRunMessage = persistedFinalRunMessage(agentComments, finalRun);
const pendingInteractions = terminal.interactions.filter(
(interaction) => interaction.status === "pending",
);
const invariantFailures: string[] = reasoningProjectionFailures(
runEventsByRun.flatMap((entry) => entry.events),
);
if (pendingInteractions.length > 0)
invariantFailures.push(
`expected no unresolved interaction; observed ${pendingInteractions.length}`,
);
if (expectedQuestionResolution) {
const questionInteractions = terminal.interactions.filter(
(interaction) => interaction.kind === "ask_user_questions",
);
if (
questionInteractions.length !== 1 ||
questionInteractions[0]?.id !==
expectedQuestionResolution.interactionId
) {
invariantFailures.push(
`expected exactly one stable question interaction ${expectedQuestionResolution.interactionId}; observed ${JSON.stringify(questionInteractions)}`,
);
}
const resolvedQuestion = questionInteractions.find(
(interaction) =>
interaction.id === expectedQuestionResolution.interactionId,
);
const answers = resolvedQuestion?.result?.answers ?? [];
if (
resolvedQuestion?.status !== "answered" ||
resolvedQuestion.result?.version !== 1 ||
answers.length !== 1 ||
answers[0]?.questionId !== "verification-word" ||
JSON.stringify(answers[0]?.optionIds) !==
JSON.stringify([expectedQuestionResolution.optionId]) ||
answers[0]?.otherText !== undefined
) {
invariantFailures.push(
`expected interaction ${expectedQuestionResolution.interactionId} to resolve with exactly ${expectedQuestionResolution.optionId}; observed ${JSON.stringify(resolvedQuestion)}`,
);
}
const continuationRun = selectedRuns.at(-1);
const continuationContext = record(continuationRun?.contextSnapshot);
if (
!continuationRun ||
continuationContext.interactionId !==
expectedQuestionResolution.interactionId
) {
invariantFailures.push(
`expected the continuation run to consume interaction ${expectedQuestionResolution.interactionId}; observed ${String(continuationContext.interactionId)}`,
);
}
if (execution.profile.generation === "native" && continuationRun) {
const runnerProfile = record(continuationRun.runnerProfileJson);
const nativeExecutionInput = record(
runnerProfile.nativeExecutionInput,
);
const interactionResponses = Array.isArray(
nativeExecutionInput.interactionResponses,
)
? nativeExecutionInput.interactionResponses.map(record)
: [];
const matchingResponses = interactionResponses.filter(
(candidate) =>
candidate.interactionId ===
expectedQuestionResolution.interactionId,
);
const responseEnvelope = matchingResponses[0];
const response = record(responseEnvelope?.response);
const responseResult = record(response.result);
const responseAnswers = Array.isArray(responseResult.answers)
? responseResult.answers.map(record)
: [];
if (
matchingResponses.length !== 1 ||
responseEnvelope?.kind !== "ask_user_questions" ||
response.status !== "answered" ||
responseResult.version !== 1 ||
responseAnswers.length !== 1 ||
responseAnswers[0]?.questionId !== "verification-word" ||
JSON.stringify(responseAnswers[0]?.optionIds) !==
JSON.stringify([expectedQuestionResolution.optionId])
) {
invariantFailures.push(
`expected native continuation input to carry exactly one answered interaction ${expectedQuestionResolution.interactionId}; observed ${JSON.stringify(matchingResponses)}`,
);
}
}
questionLifecycleEvidence = {
...(questionLifecycleEvidence ?? {}),
resolvedInteraction: resolvedQuestion ?? null,
continuationRunId: continuationRun?.id ?? null,
};
}
for (const candidate of selectedRuns) {
if (
(candidate.continuationAttempt ?? 0) !== 0 ||
candidate.retryOfRunId
)
invariantFailures.push(
`expected run ${candidate.id} without a recovery continuation`,
);
}
for (const captured of runEventsByRun) {
if (captured.error) {
invariantFailures.push(
`run ${captured.runId} events query failed: ${captured.error}`,
);
}
}
if (execution.suite.id === "api-response-reading") {
const paging = gradeApiResponsePaging(runEventsByRun.flatMap(captured => captured.events), apiResponseSourceId ?? "");
await writeSanitizedJson(snapshotsDir, "api-response-pagination.json", paging, secrets);
if (!paging.passed) invariantFailures.push(`Bounded response paging was not proven: ${paging.failure}`);
}
if (piFileSeed) {
const proof = await collectPiFileEvidence({ page, api, nonce, companyId: fixtures.company.id,
issueId: issue.id, agentId: fixtures.agent.id, run: finalRun,
events: runEventsByRun.find(entry => entry.runId === finalRun.id)?.events ?? [],
seed: piFileSeed, workspace: workspacePath,
environment: execution.environment.id === "daytona" ? "daytona" : "local", environmentId: fixtures.environment.id,
evidence: data => writeSanitizedJson(snapshotsDir, "pi-file-observation.json", data, secrets) });
await writeSanitizedJson(snapshotsDir, "pi-file-evidence.json", proof, secrets);
}
if (execution.suite.id === "task-titles") {
const titleEvidence = gradeTaskTitle({
initial: titleCreation?.issue, submitted: titleCreation?.submitted,
final: issue, prompt, explicitTitle: title, agentId: fixtures.agent.id,
runId: selectedRuns[0]!.id,
events: runEventsByRun.find(captured => captured.runId === selectedRuns[0]!.id)?.events ?? [],
activity: await api.get<unknown[]>(`/api/issues/${issue.id}/activity`),
});
await writeSanitizedJson(snapshotsDir, "task-title.json", titleEvidence, secrets);
invariantFailures.push(...titleEvidence.checks.filter(check => !check.passed).map(check => `${check.id}: ${check.detail}`));
}
const context = record(run.contextSnapshot);
const environmentContext = record(context.paperclipEnvironment);
const workspaceContext = record(context.paperclipWorkspace);
const environmentDriver =
environmentContext.driver ?? persistedEnvironment.driver;
const observedEnvironment =
environmentDriver === "sandbox" ? "daytona" : environmentDriver;
const observedRuntimeMode =
run.runtimeMode ??
(persistedAgent.adapterType === "paperclip_runner"
? "native"
: "legacy");
const matcherObservation = {
message,
issueStatus: issue.status,
runStatus: selectedRuns.every(
(candidate) =>
candidate.status === execution.task.expectedTerminalState.run,
)
? execution.task.expectedTerminalState.run
: selectedRuns.map((candidate) => candidate.status).join(","),
runtimeMode: observedRuntimeMode,
environment:
typeof observedEnvironment === "string"
? observedEnvironment
: undefined,
json: {
issue,
run,
comments: terminal.comments,
interactions: terminal.interactions,
},
};
const fileObservations = Object.fromEntries(
await Promise.all(
taskMatchers
.filter(
(matcher) =>
matcher.kind === "file_exists" ||
matcher.kind === "file_exact" ||
matcher.kind === "file_contains",
)
.map(async (matcher) => [
matcher.path,
await readFile(
path.isAbsolute(matcher.path)
? matcher.path
: path.join(workspacePath, matcher.path),
"utf8",
).catch(() => undefined),
]),
),
);
if (execution.suite.id === "api-response-reading") {
downloadedResponseProof = await readResponseProof(api, issue.id, run.id);
fileObservations["api-response-proof.txt"] = downloadedResponseProof.content;
await writeSanitizedJson(snapshotsDir, "downloaded-response-proof.json", downloadedResponseProof, secrets);
}
const registeredArtifacts = await readRegisteredArtifacts(api, issue.id, run.id,
taskMatchers.filter(matcher => matcher.kind === "artifact_exact").map(matcher => matcher.name));
if (registeredArtifacts.length > 0) {
await writeSanitizedJson(snapshotsDir, "downloaded-registered-artifacts.json", registeredArtifacts, secrets);
}
matcherResults.push(...await Promise.all(
taskMatchers.map((matcher) =>
evaluateMatcher(matcher, {
...matcherObservation,
files: fileObservations,
artifacts: registeredArtifacts,
// Multi-run tasks intentionally retain earlier waiting/revision
// replies. Exact completion text belongs to the chronological
// final run, while occurrence checks still span every agent
// comment so duplicate terminal markers cannot be hidden.
message:
matcher.kind === "message_exact" ? finalRunMessage : message,
}),
),
));
const failedMatchers = matcherResults.filter((result) => !result.passed);
const exactMessageMatcher = taskMatchers.find(
(matcher) => matcher.kind === "message_exact",
);
if (
execution.profile.id === "runner-opencode" &&
execution.task.id === "structured-question-restart-resume" &&
exactMessageMatcher?.kind === "message_exact" &&
finalRunMessage === exactMessageMatcher.expected.replace(/-\d+$/, "") &&
record(finalRun.resultJson).summary === exactMessageMatcher.expected
) {
// OpenCode can occasionally copy the complete marker into the
// accepted semantic result while dropping only the synthetic attempt
// suffix from its visible answer. Keep exact matching strict, but let
// the campaign retry this narrowly proven provider variance once in a
// fresh harness. A repeated near miss remains a failed cell.
failureClassOverride = "provider_variance";
}
const retryInteractiveTerminalProviderVariance =
(execution.suite.id === "openrouter-model-breadth" &&
(execution.task.id === "question-resume-complete" ||
execution.task.id === "plan-approve-complete")) ||
(execution.suite.id === "core-compatibility" &&
execution.profile.generation === "native" &&
execution.task.id === "plan-revise-accept");
if (
retryInteractiveTerminalProviderVariance &&
exactMessageMatcher?.kind === "message_exact" &&
selectedRuns.every((candidate) => candidate.status === "succeeded") &&
issue.status === "done" &&
finalRunMessage !== exactMessageMatcher.expected &&
record(finalRun.resultJson).summary === exactMessageMatcher.expected
) {
// An interactive breadth model can occasionally satisfy the durable
// terminal contract while paraphrasing the visible final response.
// Preserve the exact persisted-message and DOM assertions, but
// classify this narrowly proven provider variance for one
// fresh-harness retry.
failureClassOverride = "provider_variance";
}
const observedEnvironmentId =
environmentContext.id ??
(execution.environment.id === "local"
? persistedAgent.defaultEnvironmentId
: undefined);
if (observedEnvironmentId !== fixtures.environment.id) {
invariantFailures.push(
`Expected environment ${fixtures.environment.id}; observed ${String(observedEnvironmentId)}`,
);
}
if (
execution.environment.id === "daytona" &&
typeof environmentContext.leaseId !== "string"
) {
invariantFailures.push(
"expected a Daytona sandbox lease on the run context",
);
}
if (execution.task.flow === "warm_three_turn") {
const runEnvironmentContexts = selectedRuns.map((candidate) =>
record(record(candidate.contextSnapshot).paperclipEnvironment),
);
const leaseIds = runEnvironmentContexts.map((entry) => entry.leaseId);
const acquisitionOutcomes = runEnvironmentContexts.map(
(entry) => record(entry.sandboxLeaseAcquisition).outcome,
);
const projectWorkspaceIds = selectedRuns.map(
(candidate) =>
record(record(candidate.contextSnapshot).paperclipWorkspace)
.workspaceId,
);
const executionWorkspaceIds = selectedRuns.map(
(candidate) => record(candidate.contextSnapshot).executionWorkspaceId,
);
if (execution.environment.id === "daytona") {
if (
leaseIds.some(
(leaseId) => typeof leaseId !== "string" || leaseId.length === 0,
)
) {
invariantFailures.push(
`expected a persisted Daytona lease row for every warm turn; observed ${JSON.stringify(leaseIds)}`,
);
}
if (
JSON.stringify(acquisitionOutcomes) !==
JSON.stringify(["created", "resumed", "resumed"])
) {
invariantFailures.push(
`expected warm lease outcomes created,resumed,resumed; observed ${JSON.stringify(acquisitionOutcomes)}`,
);
}
}
if (
!fixtures.project?.primaryWorkspace?.id ||
projectWorkspaceIds.some(
(workspaceId) =>
workspaceId !== fixtures!.project!.primaryWorkspace!.id,
) ||
new Set(executionWorkspaceIds).size !== 1 ||
executionWorkspaceIds[0] !== issue.executionWorkspaceId ||
issue.projectId !== fixtures.project.id ||
issue.projectWorkspaceId !== fixtures.project.primaryWorkspace.id ||
!issue.executionWorkspaceId
) {
invariantFailures.push(
`warm task did not preserve project, project-workspace, and execution-workspace identity: ${JSON.stringify({ projectWorkspaceIds, executionWorkspaceIds, projectId: issue.projectId, projectWorkspaceId: issue.projectWorkspaceId, executionWorkspaceId: issue.executionWorkspaceId })}`,
);
}
if (execution.profile.generation === "native") {
const stableIdentityFields: Array<{
label: string;
values: unknown[];
}> = [
{
label: "native session",
values: selectedRuns.map(
(candidate) => candidate.nativeSessionId,
),
},
{
label: "runner instance",
values: selectedRuns.map(
(candidate) => candidate.runnerInstanceId,
),
},
{
label: "provider session",
values: selectedRuns.map((candidate) => candidate.sessionIdAfter),
},
{
label: "runner pid",
values: selectedRuns.map((candidate) => candidate.processPid),
},
{
label: "runner process fingerprint",
values: selectedRuns.map(
(candidate) => candidate.processStartedAt,
),
},
];
for (const { label, values } of stableIdentityFields) {
if (
values.some(
(value) =>
value === null ||
value === undefined ||
String(value).length === 0,
) ||
new Set(values).size !== 1
) {
invariantFailures.push(
`expected one stable ${label} across native warm turns; observed ${JSON.stringify(values)}`,
);
}
}
}
if (
turnTimings?.length !== 3 ||
turnTimings.some(
(timing) =>
timing.runStartedAt === null ||
timing.runFinishedAt === null ||
timing.schedulerLatencyMs === null ||
timing.runDurationMs === null ||
timing.responseLatencyMs === null ||
timing.responseLatencyMs >
(execution.task.turnTimeoutMs ?? 10 * 60_000),
)
) {
invariantFailures.push(
`warm turn timing data was incomplete or exceeded its structural deadline: ${JSON.stringify(turnTimings)}`,
);
}
let warmLeases: EnvironmentLeaseRecord[] = [];
let providerLeaseIds: Array<string | null | undefined> = [];
let resumedFromStates: unknown[] = [];
if (execution.environment.id === "daytona") {
const retainedLeases = await pollUntil({
label: `terminal warm Daytona lease history for issue ${issue.id}`,
deadlineAt: Math.min(deadlineAt, Date.now() + 30_000),
intervalMs: 500,
load: () =>
api.get<EnvironmentLeaseRecord[]>(
`/api/environments/${fixtures!.environment.id}/leases`,
),
accept: (leases) => {
const warmLeases = leases.filter(
(lease) =>
lease.issueId === issue!.id &&
selectedRuns.some(
(candidate) => candidate.id === lease.heartbeatRunId,
),
);
return (
warmLeases.length === 3 &&
warmLeases.every((lease) => {
const runIndex = selectedRuns.findIndex(
(candidate) => candidate.id === lease.heartbeatRunId,
);
return (
lease.status ===
(runIndex === selectedRuns.length - 1
? "retained"
: "expired") &&
lease.cleanupStatus === "success" &&
lease.leasePolicy === "reuse_by_environment" &&
typeof lease.providerLeaseId === "string" &&
record(lease.metadata).sandboxState === "started"
);
})
);
},
});
const selectedRunOrder = new Map(
selectedRuns.map((candidate, index) => [candidate.id, index]),
);
warmLeases = retainedLeases
.filter(
(lease) =>
lease.issueId === issue!.id &&
selectedRunOrder.has(lease.heartbeatRunId ?? ""),
)
.sort(
(left, right) =>
(selectedRunOrder.get(left.heartbeatRunId ?? "") ?? 0) -
(selectedRunOrder.get(right.heartbeatRunId ?? "") ?? 0),
);
providerLeaseIds = warmLeases.map(
(lease) => lease.providerLeaseId,
);
resumedFromStates = warmLeases
.slice(1)
.map((lease) => record(lease.metadata).resumedFromState);
if (
new Set(providerLeaseIds).size !== 1 ||
typeof providerLeaseIds[0] !== "string" ||
JSON.stringify(resumedFromStates) !==
JSON.stringify(["started", "started"])
) {
invariantFailures.push(
`expected one continuously-started Daytona sandbox; observed ${JSON.stringify({ providerLeaseIds, resumedFromStates })}`,
);
}
}
warmLifecycleEvidence = {
...(warmLifecycleEvidence ?? {}),
leaseIds,
acquisitionOutcomes,
projectWorkspaceIds,
executionWorkspaceIds,
providerLeaseIds,
resumedFromStates,
retainedLeases: warmLeases,
turnTimings,
};
}
if (execution.environment.id === "local") {
const runLogContent = String(record(runLog).content ?? "");
// The log endpoint returns NDJSON, so quotes inside each `chunk` are
// escaped. Accept both that wire representation and a decoded chunk.
const fallbackWorkspace =
/Using fallback workspace \\\"([^"\\]+)\\\"/.exec(
runLogContent,
)?.[1] ??
/Using fallback workspace "([^"]+)"/.exec(runLogContent)?.[1];
const cwd = String(workspaceContext.cwd ?? fallbackWorkspace ?? "");
if (!cwd.startsWith(`${temporaryRoot}/`)) {
invariantFailures.push(
`local run workspace escaped the isolated root: ${cwd}`,
);
}
}
const runEvents =
runEventsByRun.find((candidate) => candidate.runId === run.id)
?.events ?? [];
if (execution.profile.generation === "native") {
for (const candidate of selectedRuns) {
const candidateEvents =
runEventsByRun.find((captured) => captured.runId === candidate.id)
?.events ?? [];
const runnerInstanceObserved =
Boolean(candidate.runnerInstanceId) ||
candidateEvents.some(
(event) =>
typeof event.sourceInstanceId === "string" &&
event.sourceInstanceId.length > 0 &&
!event.sourceInstanceId.endsWith(":control"),
);
if (!runnerInstanceObserved) {
invariantFailures.push(
`expected native run ${candidate.id} events from a runner instance`,
);
}
invariantFailures.push(
...nativeRunEventIntegrityFailures(candidate, candidateEvents),
);
}
if (
(execution.profile.provider === "codex" ||
execution.profile.provider === "opencode") &&
selectedRuns.length > 1
) {
invariantFailures.push(
...providerSessionContinuityFailures(
execution.profile.provider,
selectedRuns,
),
);
} else if (
execution.profile.provider === "acpx" &&
selectedRuns.length > 1
) {
for (let index = 1; index < selectedRuns.length; index += 1) {
const previousSessionId = selectedRuns[index - 1]?.sessionIdAfter;
const current = selectedRuns[index]!;
const currentSessionId = current.sessionIdAfter;
if (!previousSessionId || !currentSessionId) {
invariantFailures.push(
"expected ACPX to record provider session identity across all heartbeat runs",
);
continue;
}
const acceptedPlanResetFailures = acceptedPlanSessionResetFailures(
"acpx",
previousSessionId,
current,
);
if (acceptedPlanResetFailures) {
invariantFailures.push(...acceptedPlanResetFailures);
continue;
}
if (previousSessionId === currentSessionId) continue;
const currentEvents =
runEventsByRun.find((captured) => captured.runId === current.id)
?.events ?? [];
const recordedContinuityBreak = currentEvents.some((event) => {
const payload = record(event.payload);
return (
event.eventType === "run.performance.span" &&
payload.span === "provider.session.continuity_break" &&
payload.previousProviderSessionId === previousSessionId &&
payload.replacementProviderSessionId === currentSessionId
);
});
if (!recordedContinuityBreak) {
invariantFailures.push(
`ACPX provider session changed from ${previousSessionId} to ${currentSessionId} without a matching continuity event`,
);
}
}
}
const spanPayloads = runEvents
.filter((event) => event.eventType === "run.performance.span")
.map((event) => record(event.payload));
const selectedTransport = spanPayloads.find(
(payload) => payload.span === "runner.transport.selected",
);
const expectedTransport =
execution.environment.id === "daytona"
? "provider_ingress"
: "local_loopback";
if (selectedTransport?.mode !== expectedTransport) {
invariantFailures.push(
`Expected native runner transport ${expectedTransport}; observed ${String(selectedTransport?.mode)}`,
);
}
if (execution.environment.id === "daytona") {
const authenticated = spanPayloads.some(
(event) =>
event.span === "runner.prp.authenticate" &&
event.outcome === "ok",
);
if (!authenticated) {
invariantFailures.push(
"expected authenticated native Daytona runner preview ingress",
);
}
}
} else {
for (const candidate of selectedRuns) {
const candidateEvents =
runEventsByRun.find((captured) => captured.runId === candidate.id)
?.events ?? [];
if (
candidate.runtimeMode !== "legacy" ||
candidate.runnerInstanceId ||
candidateEvents.some(
(event) =>
Object.keys(record(event.payload?.prpEvent)).length > 0,
)
) {
invariantFailures.push(
`legacy run ${candidate.id} crossed into native runner persistence`,
);
}
}
}
startupReuse = runEventsByRun.map(captured => ({ runId: captured.runId,
...summarizeStartupReuse(captured.error ? [] : captured.events) }));
await writeSanitizedJson(snapshotsDir, "startup-reuse.json", startupReuse, secrets);
await writeSanitizedJson(
snapshotsDir,
"api-state.json",
{
issue,
run,
runs: selectedRuns,
comments: terminal.comments,
interactions: terminal.interactions,
planLifecycleEvidence,
questionLifecycleEvidence,
warmLifecycleEvidence,
matcherResults,
invariantFailures,
runEvents,
runEventsByRun,
runLogs,
},
secrets,
);
if (
exactMessageMatcher?.kind === "message_exact" &&
isOpenRouterDeepSeekHelloTerminalVariance({
suiteId: execution.suite.id,
profileId: execution.profile.id,
taskId: execution.task.id,
expectedMarker: exactMessageMatcher.expected,
finalRunMessage,
allAgentMessages: message,
semanticSummary: record(finalRun.resultJson).summary,
issueStatus: issue.status,
runStatuses: selectedRuns.map((candidate) => candidate.status),
matcherResults,
invariantFailures,
})
) {
// DeepSeek can occasionally complete the semantic finish correctly but
// expose its pre-tool acknowledgement as the visible final answer. Keep
// exact persisted-message, global occurrence, and DOM checks strict,
// while allowing one fresh-harness retry only for the zero-marker form.
failureClassOverride = "provider_variance";
}
// The backend polling above can observe a terminal transition before a
// websocket invalidation reaches the already-open task page. Reload the
// canonical task route so the screenshot and UI assertions prove the
// persisted final state, not a stale client cache.
await page.goto(
`/${encodeURIComponent(issuePrefix)}/issues/${encodeURIComponent(issue.identifier ?? issue.id)}`,
{ waitUntil: "domcontentloaded" },
);
const visibleAgentReplies = page
.getByTestId("task-chat-thread")
.getByTestId("task-chat-agent-bubble");
if (execution.suite.id === "task-titles") {
await expect(page.getByRole("heading", { name: issue.title.trim(), exact: true })).toBeVisible();
}
if (execution.suite.id === "api-response-reading") {
// File delivery renders a card instead of an exact summary bubble.
// Prove the visible link points to the same independently checked bytes.
const proofLink = page.getByRole("link", { name: "Open api-response-proof.txt", exact: true }).first();
await expect(proofLink).toBeVisible({ timeout: 30_000 });
await expect(proofLink).toHaveAttribute("href", `/api/attachments/${downloadedResponseProof!.attachmentId}/content`);
} else if (execution.suite.id === "task-titles") {
await expect(visibleAgentReplies.filter({ hasText: marker }).last()).toBeVisible({ timeout: 30_000 });
} else if (["native-completion", NATIVE_INSTRUCTION_SUITE].includes(execution.suite.id) && execution.task.id === "native-blocked-report") {
await assertNativeBlockerReply(visibleAgentReplies, marker);
} else if (execution.task.flow === "warm_three_turn") {
// Prove the persisted user-facing response is visible, independently
// of the byte-for-byte workspace checks and lease continuity checks.
expect(finalRunMessage.trim()).not.toBe("");
await expect(visibleAgentReplies.filter({ hasText: finalRunMessage }).last())
.toBeVisible({ timeout: 30_000 });
} else {
const escapedMarker = marker.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
const terminalAgentReplies = visibleAgentReplies.filter({
hasText: new RegExp(`^\\s*${escapedMarker}\\s*$`),
});
await expect(terminalAgentReplies).toHaveCount(1, { timeout: 30_000 });
await expect(terminalAgentReplies.first()).toBeVisible();
// A string-valued toHaveText assertion compares the complete rendered
// text while normalizing ordinary DOM whitespace. This keeps Markdown
// layout differences harmless without allowing prefixed, suffixed, or
// substituted provider prose to masquerade as the requested response.
await expect(terminalAgentReplies.first()).toHaveText(marker, {
useInnerText: true,
});
}
// The fixture owns the expected disposition; blocked workflows must prove
// their Blocked UI rather than inheriting the completion-only Done check.
const expectedStatus = execution.task.expectedTerminalState.issue;
const expectedStatusLabel = expectedStatus.replace(/_/g, " ").replace(/\b\w/g, (letter) => letter.toUpperCase());
await expect(
page.getByTestId("issue-detail-header").getByRole("button", {
// Blocked includes the live blocker-attention explanation in its
// accessible name. The exact persisted status is asserted separately.
name: `Change status (current: ${expectedStatusLabel}${expectedStatus === "blocked" ? "" : ")"}`,
exact: expectedStatus !== "blocked",
}),
).toBeVisible({ timeout: 30_000 });
if (execution.task.flow === "warm_three_turn") {
const continuedReceipts = page
.getByTestId("task-chat-interaction-receipt")
.filter({ hasText: "Selected “Continue work”" });
await expect(continuedReceipts).toHaveCount(2, { timeout: 30_000 });
await expect(
page.getByText("Declined request", { exact: true }),
).toHaveCount(0);
}
await captureScreenshot(
"final-state",
"Final visible task state",
"final-state.png",
);
if (failedMatchers.length > 0) {
throw new Error(
`Matcher failure: ${failedMatchers.map((result) => result.detail).join("; ")}; run error: ${run.errorCode ?? "none"} ${run.error ?? "none"}`,
);
}
if (invariantFailures.length > 0) {
throw new Error(
`Runtime invariant failure: ${invariantFailures.join("; ")}`,
);
}
}
if (["native-completion", NATIVE_INSTRUCTION_SUITE].includes(execution.suite.id)) {
if (!issue || !nativeInitial) throw new Error("Missing native qualification issue or pre-execution receipt");
const [currentIssue, companyRuns, issues, agents, comments, documents, interactions, workspaceDigest] = await Promise.all([
api.get<IssueRecord>(`/api/issues/${issue.id}`),
api.get<RunRecord[]>(`/api/companies/${fixtures.company.id}/heartbeat-runs?limit=100`),
api.get<Array<{ id: string }>>(`/api/companies/${fixtures.company.id}/issues?limit=100`),
api.get<Array<{ id: string }>>(`/api/companies/${fixtures.company.id}/agents`),
api.get<CommentRecord[]>(`/api/issues/${issue.id}/comments`),
api.get<IssueDocumentRecord[]>(`/api/issues/${issue.id}/documents`),
api.get<InteractionRecord[]>(`/api/issues/${issue.id}/interactions`), nativeWorkspaceDigest(),
]);
const detailedRuns = await Promise.all(companyRuns.map(candidate => api.get<RunRecord>(`/api/heartbeat-runs/${candidate.id}`)));
selectedRuns = detailedRuns; issue = currentIssue;
const events = detailedRuns.length === 1 ? await collectRunEvents<RunEventRecord>((afterSeq, limit) => api.get(`/api/heartbeat-runs/${detailedRuns[0]!.id}/events?afterSeq=${afterSeq}&limit=${limit}`)) : [];
const observation = { caseId: execution.task.id as "assigned-skill-explicit-invocation" | "native-blocked-report",
companyId: fixtures.company.id, agentId: fixtures.agent.id, issue: currentIssue as unknown as Record<string, unknown>,
runs: detailedRuns as unknown as Record<string, unknown>[], comments: comments as unknown as Record<string, unknown>[], events: events as unknown as Record<string, unknown>[],
initial: nativeInitial, state: { issueIds: issues.map(value => value.id), agentIds: agents.map(value => value.id), documentCount: documents.length, interactionCount: interactions.length },
workspaceChanged: workspaceDigest !== nativeInitial.workspaceDigest, marker,
documentLinkContext: { appOrigin: new URL(page.url()).origin, issuePrefix: fixtures.company.issuePrefix ?? "",
issueIdentifier: currentIssue.identifier ?? "", documents } };
const grade = execution.suite.id === NATIVE_INSTRUCTION_SUITE
? gradeNativeCompletionFinalAnswer(observation) : gradeNativeCompletion(observation);
const integrity = detailedRuns.flatMap(candidate => nativeRunEventIntegrityFailures(candidate, events));
await writeSanitizedJson(snapshotsDir, "native-completion.json", { observation, grade, integrity }, secrets);
matcherResults.push(...grade.checks.map(check => ({ matcher: { kind: "json_path" as const, path: `nativeCompletion.${check.id}`, expected: true }, passed: check.passed, detail: check.detail })));
if (!grade.passed || integrity.length) throw new Error(`Native completion matcher failure: ${[...grade.checks.filter(check => !check.passed).map(check => check.id), ...integrity].join("; ")}`);
if (execution.suite.id === NATIVE_INSTRUCTION_SUITE && execution.task.id === "assigned-skill-explicit-invocation") {
const document = documents[0]!;
const href = `/${encodeURIComponent(fixtures.company.issuePrefix!)}/issues/${encodeURIComponent(currentIssue.identifier!)}#document-${encodeURIComponent(document.key)}`;
let opened = false;
try {
const link = page.getByTestId("task-chat-agent-bubble").locator(`a[href=${JSON.stringify(href)}]`).last();
await expect(link).toBeVisible({ timeout: 30_000 });
await link.click();
await expect(page).toHaveURL(new URL(href, observation.documentLinkContext.appOrigin).href);
const target = page.locator([
`[id=${JSON.stringify(`document-${document.key}`)}]:visible`,
`[id=${JSON.stringify(`side-panel-content-document:${document.key}`)}]:visible`,
].join(", "));
await expect(target).toHaveCount(1);
await expect(target).toBeVisible();
await expect(target).toContainText(marker);
opened = true;
await captureScreenshot("document-final-link", "Final reply link opens the saved document", "document-final-link.png");
} finally {
matcherResults.push({ matcher: { kind: "json_path", path: "nativeCompletion.visible-document-navigation", expected: true }, passed: opened,
detail: "The rendered final reply link opens this task's saved document and shows its original content marker." });
await writeSanitizedJson(snapshotsDir, "native-document-navigation.json", { href, documentKey: document.key, revisionId: document.latestRevisionId, opened }, secrets);
}
}
}
if (runsCompletionUpdateProbe(execution) && credentials.OPENAI_API_KEY) {
const qualification = completionQualityStatus(completionQuality);
if (qualification === "unqualified") {
failureClassOverride = "permanent_infrastructure";
throw new Error("Completion accuracy is unqualified: missing, invalid, or miscalibrated judge evidence");
}
expect(qualification, "completion reply accuracy").toBe("passed");
}
} catch (error) {
primaryError = error;
if (execution.task.flow === "first_task" && !firstTaskEvidence && classifyFailure(error) === "candidate_failure") {
failureClassOverride = "permanent_infrastructure";
}
try {
await captureFailureApiState();
} catch (captureError) {
const failureClass = classifyFailure(captureError);
if (failureClass === "secret_leak") {
failureClassOverride = "secret_leak";
primaryError = new AggregateError(
[primaryError, captureError],
`Failure evidence contained unsafe data: ${captureError instanceof Error ? captureError.message : String(captureError)}`,
);
} else {
networkDiagnostics.push({
evidenceCaptureError:
captureError instanceof Error
? captureError.message
: String(captureError),
});
}
}
if (!page.isClosed()) {
if (isReviewedFixtureScreenshotRoute()) {
await captureScreenshot(
"failure",
"Task state at failure",
"failure.png",
).catch(() => undefined);
} else {
await capturePrivateScreenshot("failure", "failure.png").catch(
() => undefined,
);
}
}
} finally {
await workerTiming.enter("cleanup");
trace?.phase("cleanup", Boolean(primaryError));
await responseTiming?.stop();
await reviewProvider?.close();
try {
await writeSanitizedJson(
snapshotsDir,
"browser-diagnostics.json",
{
console: consoleDiagnostics,
network: networkDiagnostics,
lifecycle: pageLifecycleDiagnostics,
bootstrap: await browserBootstrap.snapshot(),
},
secrets,
);
} catch (error) {
primaryError = new AggregateError(
[primaryError, error].filter(Boolean),
`Browser diagnostics contained unsafe data: ${error instanceof Error ? error.message : String(error)}`,
);
failureClassOverride = "secret_leak";
}
browserBootstrap.dispose();
if (fixtures) {
await captureRuntimeLeases().catch((error) => {
networkDiagnostics.push({
runtimeBillingCaptureError:
error instanceof Error ? error.message : String(error),
});
});
try {
const verification = await runCleanupWithObservers({
retireRuns: async () => {
await cancelActiveRunsForCleanup();
if (companyRunFlow) {
const companyRuns = await api.get<RunRecord[]>(`/api/companies/${fixtures!.company.id}/heartbeat-runs?limit=100`);
selectedRuns = await Promise.all(companyRuns.map(run => api.get<RunRecord>(`/api/heartbeat-runs/${run.id}`)));
await writeSanitizedJson(snapshotsDir, execution.task.flow === "first_task" ? "first-task-final-run-ledger.json" : "chat-final-run-ledger.json", selectedRuns, secrets);
}
if (execution.suite.id === "stock-harness") {
try {
const stock = await captureStockHarness({ api, companyId: fixtures!.company.id,
agentId: fixtures!.agent.id, generation: execution.profile.generation,
runIds: selectedRuns.map(run => run.id) });
const checks = gradeStockHarness(stock);
matcherResults.push(...checks.map(check => ({ matcher: { kind: "json_path" as const,
path: `stockHarness.${check.id}`, expected: true }, passed: check.passed, detail: check.detail })));
await writeSanitizedJson(snapshotsDir, "stock-harness.json", { ...stock, checks }, secrets);
const failures = checks.filter(check => !check.passed);
if (failures.length && !primaryError) primaryError = new Error(`Stock harness matcher failures: ${failures.map(check => check.id).join(", ")}`);
} catch (error) {
await writeSanitizedJson(snapshotsDir, "stock-harness-evidence-error.json", {
error: error instanceof Error ? error.message : String(error),
}, secrets);
if (!primaryError) primaryError = error;
}
}
},
assertions: beforeEnvironmentTeardownAssertions,
teardown: async () => {
try {
taskChat = await captureTaskChat({ get: route => api.get(route, { timeout: 5_000 }), companyId: fixtures!.company.id,
issueIds: [...new Set([issue?.id, ...selectedRuns.flatMap(run => [run.nativeIssueId,
run.contextSnapshot?.issueId as string | undefined])].filter((id): id is string => Boolean(id)))], secrets });
} catch {
// Supplemental evidence must never prevent owned fixture cleanup.
taskChat = { schema: "paperclip.runner-e2e.task-chat/v1", capturedAt: new Date().toISOString(),
status: "unavailable", threads: [], missing: ["capture"] };
}
await workerTiming.enter("feedback");
trace?.phase("feedback");
try {
feedback = await collectAttemptFeedback({ api, companyId: fixtures!.company.id,
onCheckpoint: summary => { feedback = summary; },
persist: async (evidence, signal) => {
await writeSanitizedJson(snapshotsDir, "agent-feedback.pending.json", evidence, secrets, signal);
signal.throwIfAborted();
await rename(path.join(snapshotsDir, "agent-feedback.pending.json"), path.join(snapshotsDir, "agent-feedback.json"));
signal.throwIfAborted();
},
});
} catch {
feedback.collection = feedback.bodiesCollected ? "partial" : "unavailable";
feedback.invocationCollection = "unavailable";
networkDiagnostics.push({ feedbackCaptureError: "feedback_evidence_write_failed" });
} finally {
await workerTiming.enter("cleanup");
trace?.phase("cleanup", !["complete", "not_started"].includes(feedback.collection));
await fixtures!.teardown();
}
},
});
matcherResults.push(...verification.checks.map(check => ({
matcher: { kind: "json_path" as const, path: `remoteCleanup.${check.id}`, expected: true }, passed: check.passed, detail: check.detail,
})));
if (verification.errors.length) throw new AggregateError(verification.errors, "Cleanup or remote retirement verification failed");
cleanup = "passed";
} catch (error) {
cleanup = "failed";
const priorFailureClass = primaryError
? classifyFailure(primaryError)
: undefined;
const cleanupFailureClass = classifyFailure(error);
failureClassOverride =
priorFailureClass === "secret_leak"
? priorFailureClass
: cleanupFailureClass === "cleanup_failure"
? cleanupFailureClass
: (priorFailureClass ?? cleanupFailureClass);
primaryError = new AggregateError(
[primaryError, error].filter(Boolean),
primaryError
? `${primaryError instanceof Error ? primaryError.message : String(primaryError)}; Cleanup also failed: ${error instanceof Error ? error.message : String(error)}`
: `Cleanup failed after test execution: ${error instanceof Error ? error.message : String(error)}`,
);
}
if (runtimeLeases.length > 0) {
runtimeLeases = await Promise.all(
runtimeLeases.map((lease) =>
api
.get<EnvironmentLeaseRecord>(
`/api/environment-leases/${lease.id}`,
)
.catch(() => lease),
),
);
}
} else {
cleanup =
primaryError &&
/(?:cleanup|teardown) failed/i.test(
primaryError instanceof Error
? primaryError.message
: String(primaryError),
)
? "failed"
: "passed";
}
// Local fixture teardown does not retire the API server. These assertions
// independently prove the recorded provider processes retired, while the
// workspace still exists, before final grading. The outer supervisor
// remains responsible for complete server/database cleanup.
if (cleanupAssertions.length > 0) {
const verification = await verifyCleanupAssertions(cleanupAssertions);
matcherResults.push(...verification.checks.map(check => ({
matcher: { kind: "json_path" as const, path: `cleanup.${check.id}`, expected: true },
passed: check.passed, detail: check.detail,
})));
if (verification.errors.length > 0) {
cleanup = "failed";
primaryError = new AggregateError([primaryError, ...verification.errors].filter(Boolean), "Cleanup verification failed after provider settlement");
if (failureClassOverride !== "secret_leak") failureClassOverride = "cleanup_failure";
}
}
await workerTiming.enter("finalization");
if (execution.suite.id === FEEDBACK_PERSISTENCE_SUITE) {
let evidence: unknown;
try { evidence = JSON.parse(await readFile(path.join(snapshotsDir, "agent-feedback.json"), "utf8")); } catch { evidence = undefined; }
const checks = gradeFeedbackPersistence({ evidence, nonce, companyId: fixtures?.company.id,
agentId: fixtures?.agent.id, taskId: issue?.id, runIds: selectedRuns.map(run => run.id) });
matcherResults.push(...checks.map(check => ({ matcher: { kind: "json_path" as const,
path: `feedbackPersistence.${check.id}`, expected: true }, passed: check.passed, detail: check.detail })));
if (checks.some(check => !check.passed)) primaryError ??= new Error("Feedback persistence matcher failed");
}
const traceReceipt = await trace?.export({ failed: Boolean(primaryError), runIds: selectedRuns.map(run => run.id),
cleanup, complaints: feedback.complaints, suggestions: feedback.suggestions });
await saveTrace();
const finishedAtMs = Date.now();
const runtimeUsage = fixtureRuntimeCostCoverage(buildRuntimeUsage({
environmentId: execution.environment.id,
runs: selectedRuns,
leases: runtimeLeases,
fallbackFinishedAt: new Date(finishedAtMs),
}), fixtures ?? {});
const resultWithoutBilling: RunnerE2EResult = {
schema: "paperclip.runner-e2e.result/v2",
...(process.env.PAPERCLIP_E2E_ACCOUNTING_CAMPAIGN_ID && process.env.PAPERCLIP_E2E_ACCOUNTING_ATTEMPT_ID ? {
accountingIdentity: {
campaignId: process.env.PAPERCLIP_E2E_ACCOUNTING_CAMPAIGN_ID,
attemptId: process.env.PAPERCLIP_E2E_ACCOUNTING_ATTEMPT_ID,
},
} : {}),
executionId: execution.id,
suiteId: execution.suite.id,
suiteDefinitionHash: execution.suiteDefinitionHash,
source: resolveRunnerE2ESource(firstTaskEvidence?.source ? { ...firstTaskEvidence.source, workflowRunUrl: null } : undefined),
...(execution.profile.ranking
? { rankingSnapshot: execution.profile.ranking }
: {}),
attempt,
status: primaryError ? "failed" : "passed",
...(primaryError
? {
failureClass:
failureClassOverride ?? classifyFailure(primaryError),
error:
primaryError instanceof Error
? primaryError.message
: String(primaryError),
}
: {}),
profileId: execution.profile.id,
environmentId: execution.environment.id,
caseId: execution.task.id,
provider: execution.profile.provider,
model: firstTaskEvidence ? firstTaskEvidence.observedModels[0] ?? firstTaskEvidence.configuredModel ?? "provider-default (unreported)" : execution.profile.model,
...(firstTaskEvidence ? { firstTask: firstTaskEvidence } : {}),
...(taskChat ? { taskChat } : {}),
...(completionQuality.length ? { completionQuality } : {}),
...(publicMcpUsage ? { publicMcp: publicMcpUsage } : {}),
runtimeMode: execution.profile.expectedRuntimeMode,
targetUrl: testInfo.project.use.baseURL,
issueId: issue?.id,
issueIdentifier: issue?.identifier ?? null,
runIds: selectedRuns.map((run) => run.id),
...(turnTimings ? { turnTimings } : {}),
...(responseTiming ? { latencyProbes: responseTiming.results() } : {}),
...(startupReuse ? { startupReuse } : {}),
startedAt,
finishedAt: new Date(finishedAtMs).toISOString(),
durationMs: finishedAtMs - startedAtMs,
workerTiming: await workerTiming.finish(finishedAtMs),
usage:
selectedRuns.length === 1
? (selectedRuns[0]?.usageJson ?? null)
: {
runs: selectedRuns.map((candidate) => ({
runId: candidate.id,
usage: candidate.usageJson ?? null,
})),
},
runtimeUsage,
feedback,
...(traceReceipt ? { tracing: traceReceipt } : {}),
matcherResults,
screenshots,
cleanup,
};
const result: RunnerE2EResult = {
...resultWithoutBilling,
billing: summarizeExecutionBilling(resultWithoutBilling),
};
if (firstTaskEvidence) {
await writeSanitizedJson(snapshotsDir, "first-task-result-private.json", result, secrets);
}
await mkdir(privateDir, { recursive: true });
await writeFile(
resultPath,
`${JSON.stringify(sanitizeJson(projectFirstTaskResult(result), secrets), null, 2)}\n`,
"utf8",
);
await testInfo.attach("runner-e2e-result", {
path: resultPath,
contentType: "application/json",
});
}
if (primaryError) throw primaryError;
});
}