8월 7일까지 워킹트리에만 남아 있던 미커밋 작업을 커밋한다. 여러 사본 폴더(worktree·clone)에 흩어져 있던 중간 스냅샷을 정리하기 전에 원본을 git 이력으로 고정하는 것이 목적이다. - contracts/routes/services: measurement, outcome_trajectory, rupture_repair, deliberate_practice, calibration_transfer, supervision_research, multimodal_alliance, continuous_improvement 계열 신규 모듈과 테스트 - infra/db/init: 07~16 마이그레이션(측정 기반~calibration transfer 실행) - apps/web: 세션 리뷰 카드·관리 화면·E2E 스펙 추가 - docs/ops: G0~G8 라이브 통합·배포·롤백 증거 문서와 evidence JSON/PNG - scripts: smoke·ledger·릴리스 에이전트·NAS 프리뷰 운영 스크립트 engine.public 로그 .bak과 apps/web/test-results 산출물은 커밋에서 제외했다.
298 lines
8.2 KiB
JSON
298 lines
8.2 KiB
JSON
{
|
|
"schema_version": "alliance-calibration-comparison.v1",
|
|
"generated_at": "2026-08-06 13:08:07.515265+00:00",
|
|
"baseline": {
|
|
"run": {
|
|
"run_id": "oas-g0-human-gold-v1",
|
|
"provider": "human_gold",
|
|
"model": "oas-g0-directional-reference",
|
|
"prompt_bundle_version": "not-applicable",
|
|
"generated_at": "2026-08-06 13:08:07.514242+00:00"
|
|
},
|
|
"expected_count": 9,
|
|
"ready_count": 9,
|
|
"missing_or_failed_count": 0,
|
|
"error_count": 0,
|
|
"recovered_after_retry_count": 0,
|
|
"direction_accuracy": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0,
|
|
"dimension_means": {
|
|
"bond": 0.35000000000000003,
|
|
"goal": 0.2,
|
|
"task": 0.39999999999999997
|
|
},
|
|
"rows": [
|
|
{
|
|
"case_id": "oas-g0-001",
|
|
"perspective": "independent_observer",
|
|
"dimension": "goal",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-002",
|
|
"perspective": "independent_observer",
|
|
"dimension": "task",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-003",
|
|
"perspective": "independent_observer",
|
|
"dimension": "bond",
|
|
"expected_direction": "drop",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-005",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-006",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "task",
|
|
"expected_direction": "rise",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.8,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-007",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "high",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.8,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "independent_observer",
|
|
"dimension": "goal",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "independent_observer",
|
|
"dimension": "task",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 1.0,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
}
|
|
]
|
|
},
|
|
"candidate": {
|
|
"run": {
|
|
"run_id": "g1-candidate-prompt-1.2.0-2026-08-06",
|
|
"provider": "claude_cli",
|
|
"model": "claude-opus-4-8",
|
|
"prompt_bundle_version": "1.2.0",
|
|
"generated_at": "2026-08-06 12:48:10.139793+00:00"
|
|
},
|
|
"expected_count": 9,
|
|
"ready_count": 9,
|
|
"missing_or_failed_count": 0,
|
|
"error_count": 0,
|
|
"recovered_after_retry_count": 2,
|
|
"direction_accuracy": 1.0,
|
|
"evidence_precision": 0.7272727272727273,
|
|
"evidence_recall": 0.8888888888888888,
|
|
"dimension_means": {
|
|
"bond": 0.3,
|
|
"goal": 0.175,
|
|
"task": 0.43333333333333335
|
|
},
|
|
"rows": [
|
|
{
|
|
"case_id": "oas-g0-001",
|
|
"perspective": "independent_observer",
|
|
"dimension": "goal",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.15,
|
|
"confidence": 0.85,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-002",
|
|
"perspective": "independent_observer",
|
|
"dimension": "task",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.1,
|
|
"confidence": 0.6,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-003",
|
|
"perspective": "independent_observer",
|
|
"dimension": "bond",
|
|
"expected_direction": "drop",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.3,
|
|
"confidence": 0.2,
|
|
"evidence_precision": 0.5,
|
|
"evidence_recall": 0.5
|
|
},
|
|
{
|
|
"case_id": "oas-g0-005",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.1,
|
|
"confidence": 0.75,
|
|
"evidence_precision": 0.5,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-006",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "task",
|
|
"expected_direction": "rise",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.8,
|
|
"confidence": 0.75,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-007",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.1,
|
|
"confidence": 0.75,
|
|
"evidence_precision": 0.3333333333333333,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "client_agent_report",
|
|
"dimension": "bond",
|
|
"expected_direction": "high",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.7,
|
|
"confidence": 0.45,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "independent_observer",
|
|
"dimension": "goal",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.2,
|
|
"confidence": 0.45,
|
|
"evidence_precision": 1.0,
|
|
"evidence_recall": 1.0
|
|
},
|
|
{
|
|
"case_id": "oas-g0-008",
|
|
"perspective": "independent_observer",
|
|
"dimension": "task",
|
|
"expected_direction": "low",
|
|
"status": "ready",
|
|
"error_code": null,
|
|
"direction_match": true,
|
|
"value": 0.4,
|
|
"confidence": 0.4,
|
|
"evidence_precision": 0.3333333333333333,
|
|
"evidence_recall": 0.5
|
|
}
|
|
]
|
|
},
|
|
"drift": {
|
|
"comparable_predictions": 9,
|
|
"mean_absolute_score_delta": 0.08333333333333334,
|
|
"dimension_mean_delta": {
|
|
"bond": -0.05000000000000003,
|
|
"goal": -0.02500000000000001,
|
|
"task": 0.03333333333333333
|
|
},
|
|
"direction_accuracy_delta": 0.0,
|
|
"evidence_recall_delta": -0.11111111111111116,
|
|
"evidence_precision_delta": -0.2727272727272727
|
|
}
|
|
}
|