vignette/docs/ops/evidence/alliance-calibration-comparison-prompt-1.1.0-retry2-2026-08-06.json
Yun Chan 16e791e044 G0~G8 성과·동맹 측정 OS 작업 일괄 고정
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.
2026-08-08 01:30:53 +09:00

297 lines
8.2 KiB
JSON

{
"schema_version": "alliance-calibration-comparison.v1",
"generated_at": "2026-08-06 12:41:01.831699+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 12:41:01.830634+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-1.1.0-retry2-2026-08-06",
"provider": "claude_cli",
"model": "claude-opus-4-8",
"prompt_bundle_version": "1.1.0",
"generated_at": "2026-08-06 12:40:42.603600+00:00"
},
"expected_count": 9,
"ready_count": 8,
"missing_or_failed_count": 1,
"error_count": 1,
"recovered_after_retry_count": 1,
"direction_accuracy": 0.7777777777777778,
"evidence_precision": 0.7,
"evidence_recall": 0.7777777777777778,
"dimension_means": {
"bond": 0.2925,
"goal": 0.15000000000000002,
"task": 0.465
},
"rows": [
{
"case_id": "oas-g0-001",
"perspective": "independent_observer",
"dimension": "goal",
"expected_direction": "low",
"status": "ready",
"error_code": null,
"direction_match": true,
"value": 0.1,
"confidence": 0.75,
"evidence_precision": 1.0,
"evidence_recall": 1.0
},
{
"case_id": "oas-g0-002",
"perspective": "independent_observer",
"dimension": "task",
"expected_direction": "low",
"status": "error",
"error_code": "no_structured_output",
"direction_match": false,
"value": null,
"confidence": null,
"evidence_precision": 0.0,
"evidence_recall": 0.0
},
{
"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": 0.25,
"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": 0.12,
"confidence": 0.65,
"evidence_precision": 0.3333333333333333,
"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": 0.78,
"confidence": 0.7,
"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": 0.85,
"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": false,
"value": 0.65,
"confidence": 0.55,
"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": 0.5,
"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": 0.15,
"confidence": 0.4,
"evidence_precision": 0.6666666666666666,
"evidence_recall": 1.0
}
]
},
"drift": {
"comparable_predictions": 8,
"mean_absolute_score_delta": 0.07500000000000001,
"dimension_mean_delta": {
"bond": -0.057500000000000016,
"goal": -0.05,
"task": -0.03500000000000002
},
"direction_accuracy_delta": -0.2222222222222222,
"evidence_recall_delta": -0.2222222222222222
}
}