Files
cdsl-cad/backend/live-evals/20260831T094745Z/report.json
T
2026-09-01 14:10:23 +08:00

795 lines
27 KiB
JSON

{
"status": "failed",
"author": {
"provider": "openai",
"model": "gpt-5.4-mini"
},
"reviewer": {
"provider": "openai",
"model": "gpt-5.5"
},
"author_capability": {
"schema_hash": "fb899a5de2f869b648df88a32d04a5eee92ae3d193ab23ee641892d6b59cd8e7",
"tool_count": 9,
"supported": true,
"failures": [],
"usage": {
"prompt_tokens": 41789,
"completion_tokens": 1156
},
"mode": "openai_compatible_tool_calling"
},
"reviewer_capability": {
"schema_hash": "fb899a5de2f869b648df88a32d04a5eee92ae3d193ab23ee641892d6b59cd8e7",
"tool_count": 9,
"supported": true,
"failures": [],
"usage": {
"prompt_tokens": 41789,
"completion_tokens": 541
},
"mode": "openai_compatible_tool_calling"
},
"results": [
{
"scenario": "rectangular_plate",
"repetition": 1,
"success": true,
"started_at": "2026-08-31T09:49:06.244495+00:00",
"finished_at": "2026-08-31T09:50:11.574252+00:00",
"terminal": {
"taskId": "cad_52c2f993c7e5",
"lifecycle": "completed",
"revisionId": "rev_001"
},
"projection": {
"schema_version": "3.0",
"task_id": "cad_52c2f993c7e5",
"phase": "COMPLETED",
"lifecycle": "completed",
"state_version": 9,
"active_revision": "rev_001",
"current_revision": "rev_001",
"published_revision": "rev_001",
"pending_action": null,
"active_candidate_id": "",
"repair_required": false,
"last_error": "",
"action_ledger_summary": [
{
"sequence": 1,
"at": "2026-08-31 09:49:13",
"draft_revision": 1,
"event": "requirements_draft_updated",
"item_count": 1
},
{
"sequence": 2,
"at": "2026-08-31 09:49:18",
"draft_revision": 1,
"event": "requirements_review_requested"
},
{
"sequence": 3,
"at": "2026-08-31 09:49:22",
"contract_hash": "42badf61b3707fe34319e2f2c145d0e64a781a656c8f37ad9552d96733095503",
"event": "requirements_contract_frozen",
"requirement_count": 1
},
{
"sequence": 4,
"at": "2026-08-31 09:49:31",
"action_id": "act_004",
"atomic_id": "extrude_add_blind",
"contract_hash": "fa9ffc9fd2141453fa18a389813dde2e2ef3fb4c7ef5d05e61d51d614309e95f",
"event": "proposed",
"expected_change": "Create a single solid rectangular plate with footprint 80 mm by 50 mm and thickness 8 mm.",
"intent": "Build the required base solid for the plate in one verifiable step.",
"requirement_ids": [
"req_001"
],
"working_head": "cad_52c2f993c7e5:root:v3"
},
{
"sequence": 5,
"at": "2026-08-31 09:49:39",
"action_id": "act_004",
"candidate_id": "candidate_4642dbd9452f51e3",
"event": "candidate_building",
"fragment_hash": "34b4567ab12fdc5a54b67ab39a791a6a0550daeb157fe6e40abd7f67726046bb"
},
{
"sequence": 6,
"at": "2026-08-31 09:49:40",
"action_id": "act_004",
"candidate_id": "candidate_4642dbd9452f51e3",
"claim_results": [
{
"claim_id": "claim_001",
"claim_kind": "solid_count_equals",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_002",
"claim_kind": "single_connected_body",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_003",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 80.0,
"axis": "x",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_004",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 50.0,
"axis": "y",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_005",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 8.0,
"axis": "z",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_006",
"claim_kind": "visual",
"deterministic": false,
"evidence": {
"reason": "requires independent visual review"
},
"status": "pending"
}
],
"event": "candidate_built"
},
{
"sequence": 7,
"at": "2026-08-31 09:49:49",
"action_id": "act_004",
"actual_atomic_id": "extrude_add_blind",
"candidate_id": "candidate_4642dbd9452f51e3",
"coverage": [
{
"claim_id": "claim_001",
"claim_kind": "solid_count_equals",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_002",
"claim_kind": "single_connected_body",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_003",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 80.0,
"axis": "x",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_004",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 50.0,
"axis": "y",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_005",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 8.0,
"axis": "z",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_006",
"claim_kind": "visual",
"deterministic": false,
"evidence": {
"reason": "requires independent visual review"
},
"status": "pending"
}
],
"event": "accepted",
"fragment_hash": "34b4567ab12fdc5a54b67ab39a791a6a0550daeb157fe6e40abd7f67726046bb",
"review_path": "revisions/rev_001/candidate-review.json",
"revision_id": "rev_001",
"selector_snapshot_id": "",
"working_head_before": "cad_52c2f993c7e5:root:v3"
},
{
"sequence": 8,
"at": "2026-08-31 09:50:00",
"claim_results": [
{
"claim_id": "claim_001",
"claim_kind": "solid_count_equals",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_002",
"claim_kind": "single_connected_body",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_003",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 80.0,
"axis": "x",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_004",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 50.0,
"axis": "y",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_005",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 8.0,
"axis": "z",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_006",
"claim_kind": "visual",
"deterministic": false,
"evidence": {
"reason": "requires independent visual review"
},
"status": "pending"
}
],
"event": "final_validation_requested",
"revision_id": "rev_001"
},
{
"sequence": 9,
"at": "2026-08-31 09:50:11",
"claim_results": [
{
"claim_id": "claim_001",
"claim_kind": "solid_count_equals",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_002",
"claim_kind": "single_connected_body",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_003",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 80.0,
"axis": "x",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_004",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 50.0,
"axis": "y",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_005",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 8.0,
"axis": "z",
"tolerance_mm": 0.0
},
"status": "pass"
},
{
"claim_id": "claim_006",
"claim_kind": "visual",
"deterministic": false,
"evidence": {
"reason": "requires independent visual review"
},
"status": "pending"
}
],
"event": "completed",
"final_review_path": "revisions/rev_001/final-review.json",
"revision_id": "rev_001"
}
],
"revisions": [
{
"revision_id": "rev_001",
"status": "success",
"visibility": "final",
"cdsl_path": "revisions/rev_001/model.cdsl.json",
"step_path": "revisions/rev_001/model.step",
"glb_path": "revisions/rev_001/model.glb",
"report_path": "revisions/rev_001/rebuild-report.json",
"candidate_review_path": "revisions/rev_001/candidate-review.json"
}
]
},
"usage": {
"calls": 8,
"prompt_tokens": 67140,
"completion_tokens": 2343,
"context_chars": 44387,
"records": [
{
"completion_tokens": 256,
"context_chars": 878,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5234,
"provider_id": "openai",
"raw_arguments_hash": "2e18aabb4b24b0d6a4d9f1e99aac7c625e266fb9de1b6cecc9ae0ff110b32f3a",
"tool": "submit_requirements_draft_batch",
"total_tokens": 5490
},
{
"completion_tokens": 84,
"context_chars": 1715,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5450,
"provider_id": "openai",
"raw_arguments_hash": "44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
"tool": "finalize_requirements_draft",
"total_tokens": 5534
},
{
"completion_tokens": 71,
"context_chars": 1195,
"model_id": "gpt-5.5",
"prompt_tokens": 5013,
"provider_id": "openai",
"raw_arguments_hash": "00996cae282886750e3169d668db3145d5aad3fa73c825b1b9d5d423613bdb52",
"role": "reviewer",
"tool": "review_requirements",
"total_tokens": 5084
},
{
"completion_tokens": 412,
"context_chars": 1301,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 4972,
"provider_id": "openai",
"raw_arguments_hash": "14880afd2112d4805a5e96df077239ec12139c82dbfce41f563fce53471cf537",
"tool": "propose_next_action",
"total_tokens": 5384
},
{
"completion_tokens": 366,
"context_chars": 2253,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5651,
"provider_id": "openai",
"raw_arguments_hash": "d52011a37183174b1d6e1937fdbede565bda754adb6894e35e85e2fe0883774a",
"tool": "submit_cdsl_fragment",
"total_tokens": 6017
},
{
"completion_tokens": 258,
"context_chars": 21618,
"model_id": "gpt-5.5",
"prompt_tokens": 19414,
"provider_id": "openai",
"raw_arguments_hash": "077328095f2b06d7dc28fa29ad1fcf5b876c88bc7f32b7298db29d24f048e722",
"role": "reviewer",
"tool": "review_candidate",
"total_tokens": 19672
},
{
"completion_tokens": 612,
"context_chars": 4419,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 6172,
"provider_id": "openai",
"raw_arguments_hash": "44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
"tool": "complete_task",
"total_tokens": 6784
},
{
"completion_tokens": 284,
"context_chars": 11008,
"model_id": "gpt-5.5",
"prompt_tokens": 15234,
"provider_id": "openai",
"raw_arguments_hash": "f67fcf857c12765312b6a70ca5ccac33ddd5a9c84d1075142a92f3c0eb572870",
"role": "reviewer",
"tool": "review_final",
"total_tokens": 15518
}
]
},
"checks": {
"expected_terminal_phase": true,
"token_budget": true,
"call_budget": true,
"author_turn_budget": true,
"reviewer_turn_budget": true,
"required_claims": true,
"required_operations": true,
"required_operation_alternative": true,
"immutable_artifacts": true,
"unique_revisions": true,
"raw_argument_audit": true,
"no_schema_or_decision_rejections": true
},
"rejection_codes": [],
"artifact_root": "/Users/lk/Projects/cdsl-cad/backend/live-evals/20260831T094745Z/artifacts/cad_52c2f993c7e5"
},
{
"scenario": "simple_flange",
"repetition": 1,
"success": false,
"started_at": "2026-08-31T09:50:11.576785+00:00",
"finished_at": "2026-08-31T09:52:12.813990+00:00",
"terminal": {
"taskId": "cad_e7a33b780edd",
"lifecycle": "failed",
"code": "FAILED_AUTHOR_FORMAT",
"message": "Author repeatedly failed the same canonical schema or state contract."
},
"projection": {
"schema_version": "3.0",
"task_id": "cad_e7a33b780edd",
"phase": "FAILED",
"lifecycle": "failed",
"state_version": 9,
"active_revision": "",
"current_revision": "",
"published_revision": "",
"pending_action": null,
"active_candidate_id": "",
"repair_required": true,
"last_error": "FAILED_AUTHOR_FORMAT",
"action_ledger_summary": [
{
"sequence": 10,
"at": "2026-08-31 09:50:25",
"draft_revision": 1,
"event": "requirements_draft_updated",
"item_count": 1
},
{
"sequence": 11,
"at": "2026-08-31 09:50:28",
"draft_revision": 1,
"event": "requirements_review_requested"
},
{
"sequence": 12,
"at": "2026-08-31 09:50:34",
"contract_hash": "5e2b5bbbc2e75d41288e49ae3eb844a50e370d781cb9ce70cac8122ce7576ccb",
"event": "requirements_contract_frozen",
"requirement_count": 1
},
{
"sequence": 13,
"at": "2026-08-31 09:50:45",
"action_id": "act_004",
"atomic_id": "extrude_add_blind",
"contract_hash": "fa9ffc9fd2141453fa18a389813dde2e2ef3fb4c7ef5d05e61d51d614309e95f",
"event": "proposed",
"expected_change": "Create the base flange solid as a centered 120 mm diameter, 12 mm thick round blank in millimetres.",
"intent": "Establish the primary connected body that will later receive the centered 40 mm through bore.",
"requirement_ids": [
"req_001"
],
"working_head": "cad_e7a33b780edd:root:v3"
},
{
"sequence": 14,
"at": "2026-08-31 09:50:53",
"action_id": "act_004",
"candidate_id": "candidate_b92d6c008fb7c644",
"event": "candidate_building",
"fragment_hash": "95bd1e739aed6d4a791a162b2ce05d84d0a52f4d18dc87c11dc7d6e8d7761078"
},
{
"sequence": 15,
"at": "2026-08-31 09:50:55",
"action_id": "act_004",
"candidate_id": "candidate_b92d6c008fb7c644",
"claim_results": [
{
"claim_id": "claim_001",
"claim_kind": "single_connected_body",
"deterministic": true,
"evidence": {
"solid_count": 1
},
"status": "pass"
},
{
"claim_id": "claim_002",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 120.0,
"axis": "x",
"tolerance_mm": 0.1
},
"status": "pass"
},
{
"claim_id": "claim_003",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 120.0,
"axis": "y",
"tolerance_mm": 0.1
},
"status": "pass"
},
{
"claim_id": "claim_004",
"claim_kind": "bbox_dimension_mm",
"deterministic": true,
"evidence": {
"actual_mm": 12.0,
"axis": "z",
"tolerance_mm": 0.1
},
"status": "pass"
},
{
"claim_id": "claim_005",
"claim_kind": "through_cylindrical_bore",
"deterministic": true,
"evidence": {
"reason": "the target bore has not been introduced at this checkpoint"
},
"status": "pending"
}
],
"event": "candidate_built"
},
{
"sequence": 16,
"at": "2026-08-31 09:51:16",
"candidate_id": "candidate_b92d6c008fb7c644",
"deterministic_fail": [],
"event": "candidate_rejected",
"review_path": "candidate-review.json"
},
{
"sequence": 17,
"at": "2026-08-31 09:51:22",
"action_id": "act_008",
"atomic_id": "extrude_cut_blind",
"contract_hash": "fe00ac225362469afe00cd4aee9b48259e8ceb74aa11d9a54ff038c6d12f0d50",
"event": "proposed",
"expected_change": "Cut a centered 40 mm diameter cylindrical void through the full 12 mm thickness of the flange, leaving the outer 120 mm round body intact.",
"intent": "Introduce the centered through bore required by the specification while preserving the single connected flange body.",
"requirement_ids": [
"req_001"
],
"working_head": "cad_e7a33b780edd:root:v7"
},
{
"sequence": 18,
"at": "2026-08-31 09:52:12",
"event": "failed_author_format",
"field_errors": [],
"tool": "submit_cdsl_fragment"
}
],
"revisions": []
},
"usage": {
"calls": 10,
"prompt_tokens": 70265,
"completion_tokens": 5557,
"context_chars": 44063,
"records": [
{
"completion_tokens": 745,
"context_chars": 934,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5242,
"provider_id": "openai",
"raw_arguments_hash": "bea9adcee320f20c42e965058dc3d445a4772c88cf97d8881b39c0703c1f640d",
"tool": "submit_requirements_draft_batch",
"total_tokens": 5987
},
{
"completion_tokens": 73,
"context_chars": 1877,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5499,
"provider_id": "openai",
"raw_arguments_hash": "44136fa355b3678a1146ad16f7e8649e94fb4fc21fe77e8310c060f61caaff8a",
"tool": "finalize_requirements_draft",
"total_tokens": 5572
},
{
"completion_tokens": 71,
"context_chars": 1331,
"model_id": "gpt-5.5",
"prompt_tokens": 5058,
"provider_id": "openai",
"raw_arguments_hash": "00996cae282886750e3169d668db3145d5aad3fa73c825b1b9d5d423613bdb52",
"role": "reviewer",
"tool": "review_requirements",
"total_tokens": 5129
},
{
"completion_tokens": 544,
"context_chars": 1329,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 4970,
"provider_id": "openai",
"raw_arguments_hash": "0c61ad302145316bcf37a93f2320442bf7f98bcfaf97f0007d02aa46f2bd4b38",
"tool": "propose_next_action",
"total_tokens": 5514
},
{
"completion_tokens": 393,
"context_chars": 2354,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5660,
"provider_id": "openai",
"raw_arguments_hash": "502aa2808018698f3ef18837ee58198e069831fe477e7ce8e02bb9b579b30696",
"tool": "submit_cdsl_fragment",
"total_tokens": 6053
},
{
"completion_tokens": 396,
"context_chars": 21651,
"model_id": "gpt-5.5",
"prompt_tokens": 19233,
"provider_id": "openai",
"raw_arguments_hash": "cb2915023d05ca8ee84de886fecf34aab773d93ec642f57e4c5ab0579411a283",
"role": "reviewer",
"tool": "review_candidate",
"total_tokens": 19629
},
{
"completion_tokens": 265,
"context_chars": 3019,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 5719,
"provider_id": "openai",
"raw_arguments_hash": "00b71ef952389c68fcfbbd7871941cfcb755e180486f7420f01272158acd42d0",
"tool": "propose_next_action",
"total_tokens": 5984
},
{
"completion_tokens": 1167,
"context_chars": 3614,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 6244,
"provider_id": "openai",
"raw_arguments_hash": "a61be86bfd477b481a340d77c8cca329acedd0ca32b69cf7df218e44435af80a",
"tool": "submit_cdsl_fragment",
"total_tokens": 7411
},
{
"completion_tokens": 1271,
"context_chars": 3977,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 6320,
"provider_id": "openai",
"raw_arguments_hash": "a61be86bfd477b481a340d77c8cca329acedd0ca32b69cf7df218e44435af80a",
"tool": "submit_cdsl_fragment",
"total_tokens": 7591
},
{
"completion_tokens": 632,
"context_chars": 3977,
"model_id": "gpt-5.4-mini",
"prompt_tokens": 6320,
"provider_id": "openai",
"raw_arguments_hash": "a61be86bfd477b481a340d77c8cca329acedd0ca32b69cf7df218e44435af80a",
"tool": "submit_cdsl_fragment",
"total_tokens": 6952
}
]
},
"checks": {
"expected_terminal_phase": false,
"token_budget": true,
"call_budget": true,
"author_turn_budget": true,
"reviewer_turn_budget": true,
"required_claims": false,
"required_operations": false,
"required_operation_alternative": false,
"immutable_artifacts": false,
"unique_revisions": true,
"raw_argument_audit": true,
"no_schema_or_decision_rejections": true
},
"rejection_codes": [],
"artifact_root": "/Users/lk/Projects/cdsl-cad/backend/live-evals/20260831T094745Z/artifacts/cad_e7a33b780edd"
}
],
"suite": "smoke",
"require_live": true,
"report_root": "/Users/lk/Projects/cdsl-cad/backend/live-evals/20260831T094745Z"
}