{ "schema_version": "alliance-calibration-snapshot.v1", "run_id": "g1-candidate-prompt-1.4.0-gatewayfix-2026-08-07", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "benchmark_schema": "vignette.outcome_alliance_benchmark.v1", "generated_at": "2026-08-07T07:13:16.861691Z", "predictions": [ { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "goal", "value": 0.15, "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": "1.4.0", "model_run_id": "a97bf9b3-8e84-4689-9b3b-50effb82f009", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "task", "value": 0.15, "confidence": 0.75, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "a97bf9b3-8e84-4689-9b3b-50effb82f009", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-001", "perspective": "independent_observer", "dimension": "bond", "value": 0.3, "confidence": 0.3, "evidence_turn_indices": [ 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "a97bf9b3-8e84-4689-9b3b-50effb82f009", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "0349809a6e585d3c178310e94ef79e11ed967ce3409f3961517cb0fc15b1d7dd" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "goal", "value": null, "confidence": null, "evidence_turn_indices": [], "status": "error", "error_code": "no_structured_output", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "6256340f-1b05-426b-906d-f219b56e15c6", "attempt_count": 3, "prior_model_run_ids": [ "4b2a83fb-c6ed-4ecd-8d4d-46db84afc9b5", "62c5777a-75d4-45ff-882a-c27510cddd66" ], "prior_error_codes": [ "no_structured_output", "no_structured_output" ], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "task", "value": null, "confidence": null, "evidence_turn_indices": [], "status": "error", "error_code": "no_structured_output", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "6256340f-1b05-426b-906d-f219b56e15c6", "attempt_count": 3, "prior_model_run_ids": [ "4b2a83fb-c6ed-4ecd-8d4d-46db84afc9b5", "62c5777a-75d4-45ff-882a-c27510cddd66" ], "prior_error_codes": [ "no_structured_output", "no_structured_output" ], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-002", "perspective": "independent_observer", "dimension": "bond", "value": null, "confidence": null, "evidence_turn_indices": [], "status": "error", "error_code": "no_structured_output", "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "6256340f-1b05-426b-906d-f219b56e15c6", "attempt_count": 3, "prior_model_run_ids": [ "4b2a83fb-c6ed-4ecd-8d4d-46db84afc9b5", "62c5777a-75d4-45ff-882a-c27510cddd66" ], "prior_error_codes": [ "no_structured_output", "no_structured_output" ], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6754fd04565d42b83214b23a8938a55a2275f8e28bcd94f0caa027a43ca13fab" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "goal", "value": 0.3, "confidence": 0.15, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "1d56ef71-9ea3-4a4d-afde-7aa6de152fae", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "task", "value": 0.2, "confidence": 0.55, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "1d56ef71-9ea3-4a4d-afde-7aa6de152fae", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-003", "perspective": "independent_observer", "dimension": "bond", "value": 0.3, "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": "1.4.0", "model_run_id": "1d56ef71-9ea3-4a4d-afde-7aa6de152fae", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "6012f42d024bc0a8a07c5ffbc647eef6424c3e653291ebfd51d067e4b81f6362" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "goal", "value": 0.3, "confidence": 0.25, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "e4fbce38-9dc5-4360-9a24-0be98a63e284", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "task", "value": 0.25, "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": "1.4.0", "model_run_id": "e4fbce38-9dc5-4360-9a24-0be98a63e284", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-005", "perspective": "client_agent_report", "dimension": "bond", "value": 0.1, "confidence": 0.8, "evidence_turn_indices": [ 0, 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "e4fbce38-9dc5-4360-9a24-0be98a63e284", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "49a0a720dd80edee39b128619fbe9a052139ef889acde66c8e189dc5d662d736" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "goal", "value": 0.6, "confidence": 0.45, "evidence_turn_indices": [ 0, 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "3e3c34b5-88a6-439a-a163-8934e79f0bc8", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "task", "value": 0.78, "confidence": 0.7, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "3e3c34b5-88a6-439a-a163-8934e79f0bc8", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-006", "perspective": "client_agent_report", "dimension": "bond", "value": 0.55, "confidence": 0.4, "evidence_turn_indices": [ 1, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "3e3c34b5-88a6-439a-a163-8934e79f0bc8", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "36a6237c838477ebd684f7d623af80272d241d81f1151b3df703821a38a4db27" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "goal", "value": 0.15, "confidence": 0.25, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "d88e59f0-9e4d-4283-81e3-4d605366dd5e", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "task", "value": 0.1, "confidence": 0.7, "evidence_turn_indices": [ 1, 2 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "d88e59f0-9e4d-4283-81e3-4d605366dd5e", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-007", "perspective": "client_agent_report", "dimension": "bond", "value": 0.05, "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": "1.4.0", "model_run_id": "d88e59f0-9e4d-4283-81e3-4d605366dd5e", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "236c869a1b7c94810f9a7a838ac6e2e212647105ff7f229051ff12ccf9c98b54" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "goal", "value": 0.25, "confidence": 0.5, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "5dfd31ce-41a4-496f-84c5-542dac8d092c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "task", "value": 0.3, "confidence": 0.45, "evidence_turn_indices": [ 0, 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "5dfd31ce-41a4-496f-84c5-542dac8d092c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "client_agent_report", "dimension": "bond", "value": 0.7, "confidence": 0.55, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "5dfd31ce-41a4-496f-84c5-542dac8d092c", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "b4891122ed4fb44392954ee4981fccaae99db38ba7757418641d676853ca985b", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "goal", "value": 0.2, "confidence": 0.5, "evidence_turn_indices": [ 2, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "4a0c4e61-28ec-4452-ac0b-c5a40573bc09", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "task", "value": 0.2, "confidence": 0.4, "evidence_turn_indices": [ 0, 3 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "4a0c4e61-28ec-4452-ac0b-c5a40573bc09", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" }, { "case_id": "oas-g0-008", "perspective": "independent_observer", "dimension": "bond", "value": 0.6, "confidence": 0.5, "evidence_turn_indices": [ 0, 1 ], "status": "ready", "error_code": null, "provider": "claude_cli", "model": "claude-opus-4-8", "prompt_bundle_version": "1.4.0", "model_run_id": "4a0c4e61-28ec-4452-ac0b-c5a40573bc09", "attempt_count": 1, "prior_model_run_ids": [], "prior_error_codes": [], "prompt_bundle_hash": "9e2b925c94949101b527beb5c39c46469655f81742b8758f0d36bc368029d2a4", "input_evidence_hash": "dab09723f26a5e58eb3b321abb8f4aec01d25063e6aaa198dffdcbd1ca7ad360" } ] }