EVALARC / DEVELOPMENT AUDIT

Test the grader.
Then trust the signal.

Behavioral controls for task-specific agent evaluations. Correct and faulty submissions are evaluated against the same externally enforced contract.

8/8
declared faults detected
1.000
reference correctness
docker
execution backend

Which faults depend on one case?

3 of 8 declared faults have exactly one distinct detecting case in the assessed records. Repeating a case under more seeds does not increase this margin. A surviving fault has margin zero; an environment failure remains unassessed.

Margins below are derived from the saved failing case IDs. Rendering an older audit does not rerun its controls.

Does the grader detect plausible defects?

Negative controlTargetCandidate scoreResultMarginEvidence
ack-without-workbasic0.000Detected3Inspect recorded cases
memory-onlypersistence0.783Detected3Inspect recorded cases
partial-batchtransactions0.900Detected1Inspect recorded cases
cas-always-winscompare_swap0.850Detected2Inspect recorded cases
boolean-equals-onecompare_swap0.925Detected1Inspect recorded cases
delete-noopbasic0.650Detected2Inspect recorded cases
accept-nonstring-keysvalidation0.925Detected1Inspect recorded cases
commit-on-exitcrash_recovery0.900Detected3Inspect recorded cases

Inspect the detecting cases

ack-without-work · 3 detecting cases

Target: basic. Detecting case IDs: json-types-and-overwrite, read-write-delete, seeded-state-machine.

read-write-delete · seed 17 · failed
{
  "seed": 17,
  "case_id": "read-write-delete",
  "dimension": "basic",
  "checks": {
    "basic": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 0,
  "duration_seconds": 0.608378,
  "transcript_sha256": "e9d1c1a4ce952f925f5a3d00a84348bf482007670ba13602f73d0129cc422a60",
  "error": "response mismatch at request 1; expected {\"found\": false, \"ok\": true}, got {\"ok\": true}"
}
json-types-and-overwrite · seed 17 · failed
{
  "seed": 17,
  "case_id": "json-types-and-overwrite",
  "dimension": "basic",
  "checks": {
    "basic": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.563549,
  "transcript_sha256": "ad684a2e4e62254b35236e279cda7ec03bfe6bf6a01e3c0b24a1810e6c8088d2",
  "error": "response mismatch at request 2; expected {\"found\": true, \"ok\": true, \"value\": null}, got {\"ok\": true}"
}
seeded-state-machine · seed 17 · failed
{
  "seed": 17,
  "case_id": "seeded-state-machine",
  "dimension": "basic",
  "checks": {
    "basic": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 2,
  "duration_seconds": 0.497166,
  "transcript_sha256": "ce6e264c70b36102aabe44145d2a7558766913a2725261b8f120efea98fa8659",
  "error": "response mismatch at request 3; expected {\"found\": false, \"ok\": true}, got {\"ok\": true}"
}
memory-only · 3 detecting cases

Target: persistence. Detecting case IDs: restart-preserves-acknowledged-write, restart-preserves-cas, restart-preserves-transaction.

restart-preserves-acknowledged-write · seed 17 · failed
{
  "seed": 17,
  "case_id": "restart-preserves-acknowledged-write",
  "dimension": "persistence",
  "checks": {
    "persistence": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.976095,
  "transcript_sha256": "f62cc3fb4db094e24ad23141f7cc50908d24647156c4d2ba2386a2487c078e80",
  "error": "response mismatch at request 2; expected {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}, got {\"found\": false, \"ok\": true}"
}
restart-preserves-transaction · seed 17 · failed
{
  "seed": 17,
  "case_id": "restart-preserves-transaction",
  "dimension": "persistence",
  "checks": {
    "persistence": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.996261,
  "transcript_sha256": "43cf3772fd4bd95486b1824b1d73c4e99859fb85088a7f62192d8325b4581b29",
  "error": "response mismatch at request 2; expected {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}, got {\"found\": false, \"ok\": true}"
}
restart-preserves-cas · seed 17 · failed
{
  "seed": 17,
  "case_id": "restart-preserves-cas",
  "dimension": "persistence",
  "checks": {
    "persistence": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 2,
  "duration_seconds": 0.997537,
  "transcript_sha256": "497082839111501c6aa4912312fe30686befcab05beb84ebbcb99be0d1e8aef1",
  "error": "response mismatch at request 3; expected {\"found\": true, \"ok\": true, \"value\": 321}, got {\"found\": false, \"ok\": true}"
}
partial-batch · single-case dependency

Target: transactions. Detecting case IDs: rollback-batch.

rollback-batch · seed 17 · failed
{
  "seed": 17,
  "case_id": "rollback-batch",
  "dimension": "transactions",
  "checks": {
    "transactions": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 2,
  "duration_seconds": 0.497347,
  "transcript_sha256": "f6015865dce884224cf6ccccc9fbdd367baad17842dde00fbee476f06b1aab0d",
  "error": "response mismatch at request 3; expected {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}, got {\"found\": true, \"ok\": true, \"value\": \"must roll back\"}"
}
cas-always-wins · 2 detecting cases

Target: compare_swap. Detecting case IDs: cas-match-and-mismatch, cas-type-sensitivity.

cas-match-and-mismatch · seed 17 · failed
{
  "seed": 17,
  "case_id": "cas-match-and-mismatch",
  "dimension": "compare_swap",
  "checks": {
    "compare_swap": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.557099,
  "transcript_sha256": "71199e89272547e9c1a1961a85a650b93c30e2f848912b6be28ea262907eda56",
  "error": "response mismatch at request 2; expected {\"ok\": true, \"swapped\": false}, got {\"ok\": true, \"swapped\": true}"
}
cas-type-sensitivity · seed 17 · failed
{
  "seed": 17,
  "case_id": "cas-type-sensitivity",
  "dimension": "compare_swap",
  "checks": {
    "compare_swap": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.565552,
  "transcript_sha256": "70231a539612a6973a06489411646982ea651ca401223316f7f670e6560a18bf",
  "error": "response mismatch at request 2; expected {\"ok\": true, \"swapped\": false}, got {\"ok\": true, \"swapped\": true}"
}
boolean-equals-one · single-case dependency

Target: compare_swap. Detecting case IDs: cas-type-sensitivity.

cas-type-sensitivity · seed 17 · failed
{
  "seed": 17,
  "case_id": "cas-type-sensitivity",
  "dimension": "compare_swap",
  "checks": {
    "compare_swap": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.57604,
  "transcript_sha256": "70231a539612a6973a06489411646982ea651ca401223316f7f670e6560a18bf",
  "error": "response mismatch at request 2; expected {\"ok\": true, \"swapped\": false}, got {\"ok\": true, \"swapped\": true}"
}
delete-noop · 2 detecting cases

Target: basic. Detecting case IDs: read-write-delete, seeded-state-machine.

read-write-delete · seed 17 · failed
{
  "seed": 17,
  "case_id": "read-write-delete",
  "dimension": "basic",
  "checks": {
    "basic": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 4,
  "duration_seconds": 0.557506,
  "transcript_sha256": "4122b55bc205e4703a3a9f62a5397de448a833f424e54edc914bbbf5465c70d1",
  "error": "response mismatch at request 5; expected {\"found\": false, \"ok\": true}, got {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}"
}
seeded-state-machine · seed 17 · failed
{
  "seed": 17,
  "case_id": "seeded-state-machine",
  "dimension": "basic",
  "checks": {
    "basic": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 8,
  "duration_seconds": 0.556806,
  "transcript_sha256": "0132892f9b4bac6264eb6ef2dae9da8b7bac7a6a158b56357f1319c853615de7",
  "error": "response mismatch at request 9; expected {\"found\": false, \"ok\": true}, got {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}"
}
accept-nonstring-keys · single-case dependency

Target: validation. Detecting case IDs: reject-and-continue.

reject-and-continue · seed 17 · failed
{
  "seed": 17,
  "case_id": "reject-and-continue",
  "dimension": "validation",
  "checks": {
    "validation": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 4,
  "duration_seconds": 0.519073,
  "transcript_sha256": "dc07c86f871afd6b2f208e9751bdd459e52368d54d008ca0622a8edd61a7008b",
  "error": "response mismatch at request 5; expected {\"error\": \"invalid_request\", \"ok\": false}, got {\"ok\": true}"
}
commit-on-exit · 3 detecting cases

Target: crash_recovery. Detecting case IDs: sigkill-after-acknowledgement, sigkill-after-atomic-batch, sigkill-after-delete.

sigkill-after-acknowledgement · seed 17 · failed
{
  "seed": 17,
  "case_id": "sigkill-after-acknowledgement",
  "dimension": "crash_recovery",
  "checks": {
    "crash_recovery": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.998141,
  "transcript_sha256": "f62cc3fb4db094e24ad23141f7cc50908d24647156c4d2ba2386a2487c078e80",
  "error": "response mismatch at request 2; expected {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}, got {\"found\": false, \"ok\": true}"
}
sigkill-after-atomic-batch · seed 17 · failed
{
  "seed": 17,
  "case_id": "sigkill-after-atomic-batch",
  "dimension": "crash_recovery",
  "checks": {
    "crash_recovery": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 1,
  "duration_seconds": 0.978949,
  "transcript_sha256": "43cf3772fd4bd95486b1824b1d73c4e99859fb85088a7f62192d8325b4581b29",
  "error": "response mismatch at request 2; expected {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}, got {\"found\": false, \"ok\": true}"
}
sigkill-after-delete · seed 17 · failed
{
  "seed": 17,
  "case_id": "sigkill-after-delete",
  "dimension": "crash_recovery",
  "checks": {
    "crash_recovery": false
  },
  "passed": false,
  "status": "failed",
  "checks_completed": 2,
  "duration_seconds": 1.47289,
  "transcript_sha256": "679e7ab1f7f833fb12551b29af4fdc1ad8e1051810e594a4da85fdc3420b8c42",
  "error": "response mismatch at request 3; expected {\"found\": false, \"ok\": true}, got {\"found\": true, \"ok\": true, \"value\": {\"items\": [true, null, \"\\u03bb\\n\\u4e16\\u754c\"], \"version\": 5991}}"
}

Positive control

basic1.003/3
validation1.002/2
transactions1.002/2
compare_swap1.002/2
persistence1.003/3
crash_recovery1.003/3

Reproduction record

Task: durable-kv v0.1.0
Seeds: [17]
Grader SHA-256: 52caa3136434569f96097860bc9bf85eb7ada5f7f9890c0540db81afb441871d
Cases SHA-256: 842a16a1e65700cde92ee5f5aaaa10d494a9fc9d71bd2d99fb68147b19aad63a
Image ID: sha256:78387bc3881b8273120a12ebe6c1ab22b018ccc2c9adf565ae1ac9b536e184ea
Created: 2026-09-14T01:29:51.125110+00:00

Complete JSON evidence for all controls