{ "timestamp_utc": "2026-10-08T21:14:48.025782+00:00", "dataset_version": "synthetic-regression-v1", "metric_definition": "requirement coverage, valid dependency DAG, zero outbound handoff; no delivery estimate accuracy claim", "real_inference": true, "fictional_data_only": true, "case_count": 8, "completed": 8, "elapsed_seconds": 93.52, "environment": { "gpu": "NVIDIA GeForce RTX 5060 Ti, 595.84, 16311 MiB", "python": "3.12.3", "model_runtime": "Existing Qwen 3.5 9B GGUF / MiniLM L12-H384 / Whisper small.en / Kokoro af_heart; only applicable capabilities used", "providers_enabled": { "llm": true, "embedding": false, "rerank": false, "asr": false, "tts": false, "ocr": false } }, "metrics": { "requirement_coverage": { "passed": 8, "total": 8 }, "dependency_ids_valid": { "passed": 8, "total": 8 }, "zero_outbound_handoff": { "passed": 8, "total": 8 } }, "cases": [ { "id": "plan-01", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define supported file formats and upload mechanism for fictional invoices", "days": 2, "depends_on": [], "requirement_ids": [ "R1" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Design and implement data correction interface and logic", "days": 3, "depends_on": [ "T1" ], "requirement_ids": [ "R2" ], "start_day": 2, "end_day": 5 }, { "id": "T3", "title": "Define JSON export schema constraints and validation rules", "days": 1, "depends_on": [ "T2" ], "requirement_ids": [ "R3" ], "start_day": 5, "end_day": 6 }, { "id": "T4", "title": "Implement JSON export functionality", "days": 2, "depends_on": [ "T3" ], "requirement_ids": [ "R3" ], "start_day": 6, "end_day": 8 }, { "id": "T5", "title": "Design and implement mandatory human review workflow", "days": 3, "depends_on": [ "T4" ], "requirement_ids": [ "R4" ], "start_day": 8, "end_day": 11 }, { "id": "T6", "title": "Integrate review step as a gate before final JSON export", "days": 2, "depends_on": [ "T5", "T4" ], "requirement_ids": [ "R4" ], "start_day": 11, "end_day": 13 }, { "id": "T7", "title": "End-to-end verification of upload, correction, review, and export flow", "days": 2, "depends_on": [ "T6" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 13, "end_day": 15 } ], "completed": true, "elapsed_seconds": 11.381 }, { "id": "plan-02", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define report schema and validation rules for fictional inspection reports", "days": 2, "depends_on": [], "requirement_ids": [ "R1", "R3", "R4" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Implement local workspace file upload mechanism", "days": 3, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 2, "end_day": 5 }, { "id": "T3", "title": "Design and implement content correction interface and logic", "days": 4, "depends_on": [ "T1", "T2" ], "requirement_ids": [ "R2" ], "start_day": 5, "end_day": 9 }, { "id": "T4", "title": "Implement JSON export logic with human review gate enforcement", "days": 3, "depends_on": [ "T1", "T3" ], "requirement_ids": [ "R3", "R4" ], "start_day": 9, "end_day": 12 }, { "id": "T5", "title": "Develop human review workflow UI and state management", "days": 3, "depends_on": [ "T1", "T4" ], "requirement_ids": [ "R4" ], "start_day": 12, "end_day": 15 }, { "id": "T6", "title": "Integrate human review step into export pipeline", "days": 2, "depends_on": [ "T3", "T4", "T5" ], "requirement_ids": [ "R4" ], "start_day": 15, "end_day": 17 }, { "id": "T7", "title": "End-to-end testing of upload, correction, review, and export flows", "days": 3, "depends_on": [ "T2", "T3", "T4", "T5", "T6" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 17, "end_day": 20 } ], "completed": true, "elapsed_seconds": 11.413 }, { "id": "plan-03", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define Data Schema and Validation Rules", "days": 3, "depends_on": [], "requirement_ids": [ "R1", "R2" ], "start_day": 0, "end_day": 3 }, { "id": "T2", "title": "Implement File Upload Interface", "days": 2, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 3, "end_day": 5 }, { "id": "T3", "title": "Implement Data Correction Functionality", "days": 4, "depends_on": [ "T1", "T2" ], "requirement_ids": [ "R2" ], "start_day": 5, "end_day": 9 }, { "id": "T4", "title": "Implement Human Review Workflow", "days": 3, "depends_on": [ "T3" ], "requirement_ids": [ "R4" ], "start_day": 9, "end_day": 12 }, { "id": "T5", "title": "Implement JSON Export Feature", "days": 2, "depends_on": [ "T4" ], "requirement_ids": [ "R3", "R4" ], "start_day": 12, "end_day": 14 }, { "id": "T6", "title": "End-to-End Integration and Verification", "days": 2, "depends_on": [ "T5" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 14, "end_day": 16 } ], "completed": true, "elapsed_seconds": 9.499 }, { "id": "plan-04", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define supported file formats and upload mechanism", "days": 3, "depends_on": [], "requirement_ids": [ "R1" ], "start_day": 0, "end_day": 3 }, { "id": "T2", "title": "Implement local workspace file upload functionality", "days": 5, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 3, "end_day": 8 }, { "id": "T3", "title": "Design data correction interface and logic", "days": 4, "depends_on": [ "T2" ], "requirement_ids": [ "R2" ], "start_day": 8, "end_day": 12 }, { "id": "T4", "title": "Implement data correction functionality", "days": 5, "depends_on": [ "T3" ], "requirement_ids": [ "R2" ], "start_day": 12, "end_day": 17 }, { "id": "T5", "title": "Define 'human review' criteria and workflow state machine", "days": 3, "depends_on": [ "T4" ], "requirement_ids": [ "R4" ], "start_day": 17, "end_day": 20 }, { "id": "T6", "title": "Implement mandatory human review step logic", "days": 4, "depends_on": [ "T5", "T4" ], "requirement_ids": [ "R4" ], "start_day": 20, "end_day": 24 }, { "id": "T7", "title": "Implement JSON export functionality with review gate", "days": 4, "depends_on": [ "T6", "T4" ], "requirement_ids": [ "R3", "R4" ], "start_day": 24, "end_day": 28 }, { "id": "T8", "title": "End-to-end integration testing and verification", "days": 5, "depends_on": [ "T7" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 28, "end_day": 33 } ], "completed": true, "elapsed_seconds": 11.617 }, { "id": "plan-05", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Clarify File Format Support and JSON Schema", "days": 2, "depends_on": [], "requirement_ids": [ "R1", "R3", "R4" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Define User Roles and Review Workflow", "days": 2, "depends_on": [], "requirement_ids": [ "R2", "R4" ], "start_day": 0, "end_day": 2 }, { "id": "T3", "title": "Specify Local Workspace Technical Constraints", "days": 1, "depends_on": [], "requirement_ids": [ "R5", "R6", "R7" ], "start_day": 0, "end_day": 1 }, { "id": "T4", "title": "Finalize Acceptance Criteria and Compliance Standards", "days": 2, "depends_on": [], "requirement_ids": [ "R6", "R7" ], "start_day": 0, "end_day": 2 }, { "id": "T5", "title": "Design Database Schema for Notes and Reviews", "days": 3, "depends_on": [ "T1", "T2", "T3" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 2, "end_day": 5 }, { "id": "T6", "title": "Implement File Upload Service", "days": 4, "depends_on": [ "T1", "T5" ], "requirement_ids": [ "R1" ], "start_day": 5, "end_day": 9 }, { "id": "T7", "title": "Implement Content Correction Interface", "days": 4, "depends_on": [ "T2", "T5" ], "requirement_ids": [ "R2" ], "start_day": 5, "end_day": 9 }, { "id": "T8", "title": "Implement Human Review Workflow", "days": 5, "depends_on": [ "T2", "T5" ], "requirement_ids": [ "R4" ], "start_day": 5, "end_day": 10 }, { "id": "T10", "title": "Configure Local Workspace Deployment", "days": 3, "depends_on": [ "T3", "T5" ], "requirement_ids": [ "R5" ], "start_day": 5, "end_day": 8 }, { "id": "T9", "title": "Implement JSON Export Functionality", "days": 3, "depends_on": [ "T1", "T8" ], "requirement_ids": [ "R3", "R4" ], "start_day": 10, "end_day": 13 }, { "id": "T11", "title": "System Integration and End-to-End Testing", "days": 5, "depends_on": [ "T6", "T7", "T8", "T9", "T10" ], "requirement_ids": [ "R1", "R2", "R3", "R4", "R5" ], "start_day": 13, "end_day": 18 }, { "id": "T12", "title": "Project Finalization and Handover", "days": 2, "depends_on": [ "T11", "T4" ], "requirement_ids": [ "R6", "R7" ], "start_day": 18, "end_day": 20 } ], "completed": true, "elapsed_seconds": 17.808 }, { "id": "plan-06", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define supported file formats and upload interface requirements", "days": 2, "depends_on": [], "requirement_ids": [ "R1" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Implement local workspace file upload functionality", "days": 3, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 2, "end_day": 5 }, { "id": "T3", "title": "Design and implement content correction UI and logic", "days": 4, "depends_on": [ "T2" ], "requirement_ids": [ "R2" ], "start_day": 5, "end_day": 9 }, { "id": "T4", "title": "Define JSON export schema and structure", "days": 2, "depends_on": [ "T3" ], "requirement_ids": [ "R3" ], "start_day": 9, "end_day": 11 }, { "id": "T5", "title": "Implement mandatory human review workflow and enforcement logic", "days": 3, "depends_on": [ "T4" ], "requirement_ids": [ "R4" ], "start_day": 11, "end_day": 14 }, { "id": "T6", "title": "Implement JSON export functionality gated by review status", "days": 2, "depends_on": [ "T5" ], "requirement_ids": [ "R3", "R4" ], "start_day": 14, "end_day": 16 }, { "id": "T7", "title": "End-to-end integration testing and verification", "days": 3, "depends_on": [ "T6" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 16, "end_day": 19 } ], "completed": true, "elapsed_seconds": 10.54 }, { "id": "plan-07", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define supported file formats and upload mechanism", "days": 2, "depends_on": [], "requirement_ids": [ "R1" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Implement local workspace file upload functionality", "days": 3, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 2, "end_day": 5 }, { "id": "T3", "title": "Design and implement data correction UI/mechanisms", "days": 4, "depends_on": [ "T2" ], "requirement_ids": [ "R2" ], "start_day": 5, "end_day": 9 }, { "id": "T4", "title": "Implement JSON export logic with validation rules", "days": 3, "depends_on": [ "T3" ], "requirement_ids": [ "R3" ], "start_day": 9, "end_day": 12 }, { "id": "T5", "title": "Implement mandatory human review workflow and state management", "days": 3, "depends_on": [ "T3", "T4" ], "requirement_ids": [ "R4" ], "start_day": 12, "end_day": 15 }, { "id": "T6", "title": "Integrate review gate to block JSON export until review is complete", "days": 2, "depends_on": [ "T5", "T4" ], "requirement_ids": [ "R4" ], "start_day": 15, "end_day": 17 }, { "id": "T7", "title": "End-to-end integration testing and verification", "days": 3, "depends_on": [ "T2", "T3", "T4", "T5", "T6" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 17, "end_day": 20 } ], "completed": true, "elapsed_seconds": 11.148 }, { "id": "plan-08", "measurements": { "requirement_coverage": true, "dependency_ids_valid": true, "zero_outbound_handoff": true }, "tasks": [ { "id": "T1", "title": "Define Document Schema and Validation Rules", "days": 2, "depends_on": [], "requirement_ids": [ "R1", "R3" ], "start_day": 0, "end_day": 2 }, { "id": "T2", "title": "Implement Local Workspace Upload Functionality", "days": 3, "depends_on": [ "T1" ], "requirement_ids": [ "R1" ], "start_day": 2, "end_day": 5 }, { "id": "T3", "title": "Design and Implement Content Correction Interface", "days": 4, "depends_on": [ "T1" ], "requirement_ids": [ "R2" ], "start_day": 2, "end_day": 6 }, { "id": "T4", "title": "Implement Human Review Workflow and State Management", "days": 3, "depends_on": [ "T2", "T3" ], "requirement_ids": [ "R4" ], "start_day": 6, "end_day": 9 }, { "id": "T5", "title": "Implement JSON Export Logic with Review Gate", "days": 3, "depends_on": [ "T1", "T4" ], "requirement_ids": [ "R3", "R4" ], "start_day": 9, "end_day": 12 }, { "id": "T6", "title": "End-to-End Integration and Verification", "days": 2, "depends_on": [ "T2", "T3", "T4", "T5" ], "requirement_ids": [ "R1", "R2", "R3", "R4" ], "start_day": 12, "end_day": 14 } ], "completed": true, "elapsed_seconds": 10.049 } ], "limits": "Small authored synthetic regression set. Literal comparisons are proxies, not semantic entailment. Synthetic speech is not representative of real-call noise, accents or diarization.", "latency": { "median_case_seconds": 11.264, "p95_case_seconds": 17.808, "includes": "Full HTTP case, including upload/ASR/OCR where applicable; first requests may include lazy loading." } }