{ "schema_version": "alliance-calibration-snapshot.v1", "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", "benchmark_schema": "vignette.outcome_alliance_benchmark.v1", "generated_at": "2026-08-06T12:48:10.139793Z", "predictions": [ { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "goal", "value": 0.15, "confidence": 0.85, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "79cae1ab-6977-49e2-958a-4bf604699bb3", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "task", "value": 0.15, "confidence": 0.85, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "79cae1ab-6977-49e2-958a-4bf604699bb3", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "bond", "value": 0.3, "confidence": 0.3, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "79cae1ab-6977-49e2-958a-4bf604699bb3", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "goal", "value": 0.15, "confidence": 0.35, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "77f90cdd-09b2-4b37-af08-fbc7fb7b5a0c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "task", "value": 0.1, "confidence": 0.6, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "77f90cdd-09b2-4b37-af08-fbc7fb7b5a0c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "bond", "value": 0.2, "confidence": 0.3, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "77f90cdd-09b2-4b37-af08-fbc7fb7b5a0c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "goal", "value": 0.3, "confidence": 0.15, "evidence_turn_indices": [ 0 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "63634a6b-ccd7-48e5-8300-51762d2157fe", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "task", "value": 0.15, "confidence": 0.55, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "63634a6b-ccd7-48e5-8300-51762d2157fe", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "bond", "value": 0.3, "confidence": 0.2, "evidence_turn_indices": [ 0, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "63634a6b-ccd7-48e5-8300-51762d2157fe", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "goal", "value": 0.2, "confidence": 0.2, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "597d56b0-2d62-4a0e-828d-72e3c9bb41f5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "task", "value": 0.1, "confidence": 0.7, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "597d56b0-2d62-4a0e-828d-72e3c9bb41f5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "bond", "value": 0.1, "confidence": 0.75, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "597d56b0-2d62-4a0e-828d-72e3c9bb41f5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "goal", "value": 0.35, "confidence": 0.3, "evidence_turn_indices": [ 0, 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "7f2ddbb2-9265-4554-b566-d45789d061b5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "task", "value": 0.8, "confidence": 0.75, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "7f2ddbb2-9265-4554-b566-d45789d061b5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "bond", "value": 0.5, "confidence": 0.45, "evidence_turn_indices": [ 1, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "7f2ddbb2-9265-4554-b566-d45789d061b5", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "goal", "value": 0.15, "confidence": 0.2, "evidence_turn_indices": [ 0, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "9bce3592-6c88-4cdd-a62d-4f8161b30937", "attempt_count": 2, "prior_model_run_ids": [ "72683e07-94b7-453a-a0c4-faede6de9e06" ], "prior_error_codes": [ "agent_engineerror" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "task", "value": 0.1, "confidence": 0.55, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "9bce3592-6c88-4cdd-a62d-4f8161b30937", "attempt_count": 2, "prior_model_run_ids": [ "72683e07-94b7-453a-a0c4-faede6de9e06" ], "prior_error_codes": [ "agent_engineerror" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "bond", "value": 0.1, "confidence": 0.75, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "9bce3592-6c88-4cdd-a62d-4f8161b30937", "attempt_count": 2, "prior_model_run_ids": [ "72683e07-94b7-453a-a0c4-faede6de9e06" ], "prior_error_codes": [ "agent_engineerror" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "goal", "value": 0.3, "confidence": 0.35, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "aaec0b36-39c7-441f-a8ee-a00e9808a34f", "attempt_count": 2, "prior_model_run_ids": [ "c38909d9-7150-4cd7-9604-d924681feeed" ], "prior_error_codes": [ "agent_validation_missing_task_rationale" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "task", "value": 0.4, "confidence": 0.35, "evidence_turn_indices": [ 0, 1, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "aaec0b36-39c7-441f-a8ee-a00e9808a34f", "attempt_count": 2, "prior_model_run_ids": [ "c38909d9-7150-4cd7-9604-d924681feeed" ], "prior_error_codes": [ "agent_validation_missing_task_rationale" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "bond", "value": 0.7, "confidence": 0.45, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "aaec0b36-39c7-441f-a8ee-a00e9808a34f", "attempt_count": 2, "prior_model_run_ids": [ "c38909d9-7150-4cd7-9604-d924681feeed" ], "prior_error_codes": [ "agent_validation_missing_task_rationale" ], "prompt_bundle_hash": "44a5b06787cff6a3a31e734bc9a0b6fa4a3ae543635927adc324c8a3196a6933", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "goal", "value": 0.2, "confidence": 0.45, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "fd084a6a-aac9-4708-acd0-1930dc8c7237", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "task", "value": 0.4, "confidence": 0.4, "evidence_turn_indices": [ 0, 1, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "fd084a6a-aac9-4708-acd0-1930dc8c7237", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "bond", "value": 0.65, "confidence": 0.5, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.2.0", "model_run_id": "fd084a6a-aac9-4708-acd0-1930dc8c7237", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "8e5ad200cb35ca3c01cca12c12cecf8c6237c9ad1af25ddfd9b766ad8a538fd8", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" } ] }