{ "created_at": "2026-11-12T07:59:12Z", "result": { "engine": "model_orchestration", "limits": "Reviews supplied routing and outcome records. It does not call any model or verify a quality score. The run is Ed25519-signed; pin the issuer key out of band.", "module": "cain_a75_engines/modelorchestration/engine.py", "profile": "cross_lab_routing", "run": { "label": "Recomputes multi-model routing: model capabilities, context, provider, residency, cost, fallbacks, cross-lab review for high-risk tasks, outcome coverage and claim honesty. It does not call any model.", "public_key_b64": "xtu45xaxR222DBRx5eSF+5SHGv0v5HigytXM0IbN1Us=", "records": [ { "checks": [ { "detail": "model 'model-a-fast'", "id": "T1-summarize.model_exists", "ok": true }, { "detail": "needs ['summarize'], model has ['code', 'summarize']", "id": "T1-summarize.capabilities", "ok": true }, { "detail": "2000 <= 32000", "id": "T1-summarize.context_fits", "ok": true }, { "detail": "provider 'lab-a' in ['lab-a', 'lab-b', 'lab-c']", "id": "T1-summarize.provider_allowed", "ok": true }, { "detail": "data residency None", "id": "T1-summarize.residency", "ok": true }, { "detail": "1.0 <= 20.0", "id": "T1-summarize.cost_ceiling", "ok": true }, { "detail": "chain ['model-c-cheap']", "id": "T1-summarize.fallback_exists", "ok": true }, { "detail": "primary 'model-a-fast', chain ['model-c-cheap']", "id": "T1-summarize.fallback_distinct", "ok": true }, { "detail": "not flagged high-risk", "id": "T1-summarize.review_not_required", "ok": true }, { "detail": "model 'model-b-deep'", "id": "T2-fix.model_exists", "ok": true }, { "detail": "needs ['code', 'review'], model has ['code', 'review', 'reason']", "id": "T2-fix.capabilities", "ok": true }, { "detail": "8000 <= 128000", "id": "T2-fix.context_fits", "ok": true }, { "detail": "provider 'lab-b' in ['lab-a', 'lab-b', 'lab-c']", "id": "T2-fix.provider_allowed", "ok": true }, { "detail": "data residency 'eu'", "id": "T2-fix.residency", "ok": true }, { "detail": "16.0 <= 20.0", "id": "T2-fix.cost_ceiling", "ok": true }, { "detail": "chain ['model-a-fast']", "id": "T2-fix.fallback_exists", "ok": true }, { "detail": "primary 'model-b-deep', chain ['model-a-fast']", "id": "T2-fix.fallback_distinct", "ok": true }, { "detail": "reviewer 'model-a-fast'", "id": "T2-fix.review_present", "ok": true }, { "detail": "reviewer provider 'lab-a' vs actor 'lab-b'", "id": "T2-fix.review_cross_lab", "ok": true }, { "detail": "every route has an outcome", "id": "outcomes.cover_routes", "ok": true }, { "detail": "1.0 >= 0.9", "id": "outcomes.success_rate", "ok": true }, { "detail": "17.0 <= 30.0", "id": "cost.run_budget", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.all_routes_valid", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.success_meets_gate", "ok": true }, { "detail": "claimed True, computed True", "id": "claims.within_budget", "ok": true } ], "created_at": "2026-11-12T07:59:12Z", "engine": "a75-013/1.0.0", "extra": { "computed": { "all_routes_valid": true, "success_meets_gate": true, "within_budget": true }, "routes": { "T1-summarize": { "cost": 1.0, "model": "model-a-fast", "provider": "lab-a" }, "T2-fix": { "cost": 16.0, "model": "model-b-deep", "provider": "lab-b" } }, "success_rate": 1.0, "total_cost": 17.0 }, "failed": [], "label": "Recomputes multi-model routing: model capabilities, context, provider, residency, cost, fallbacks, cross-lab review for high-risk tasks, outcome coverage and claim honesty. It does not call any model.", "prev_record_hash": "de8087a6b491a6e99233d89705e8112831a8e1490d6037bbf252ca9a6bc0b6fb", "public_key_b64": "xtu45xaxR222DBRx5eSF+5SHGv0v5HigytXM0IbN1Us=", "record_hash": "2e0411a70c828025e8ef80faa165456dd881a04da70b0b5725f12a67b1125ef8", "schema": "cain42.a75.orchestration.record.v1", "signature_b64": "zfvCS+Wj0TKlPUfwrYjmaqhBsCQd21hslL1/ego+TTWJJioLebrY/iqLYYbhQSLABICl4hvNEH7FH93Li0/rBA==", "sku": "CAIN-A75-013", "subject": { "claims": { "all_routes_valid": true, "success_meets_gate": true, "within_budget": true }, "models": [ { "capabilities": [ "code", "summarize" ], "context_window": 32000, "cost_per_1k": 0.5, "id": "model-a-fast", "provider": "lab-a", "residency": [ "us" ], "version": "1.2" }, { "capabilities": [ "code", "review", "reason" ], "context_window": 128000, "cost_per_1k": 2.0, "id": "model-b-deep", "provider": "lab-b", "residency": [ "us", "eu" ], "version": "2.0" }, { "capabilities": [ "summarize" ], "context_window": 16000, "cost_per_1k": 0.2, "id": "model-c-cheap", "provider": "lab-c", "residency": [ "us" ], "version": "0.9" } ], "outcomes": [ { "cost": 1.0, "latency_ms": 900, "quality_score": 0.91, "success": true, "task_id": "T1-summarize" }, { "cost": 16.0, "latency_ms": 2400, "quality_score": 0.95, "success": true, "task_id": "T2-fix" } ], "policy": { "allowed_providers": [ "lab-a", "lab-b", "lab-c" ], "max_cost_per_task": 20.0, "min_success_rate": 0.9, "run_budget": 30.0 }, "routes": [ { "fallback_chain": [ "model-c-cheap" ], "high_risk": false, "model": "model-a-fast", "required_capabilities": [ "summarize" ], "task_id": "T1-summarize", "tokens": 2000 }, { "data_residency": "eu", "fallback_chain": [ "model-a-fast" ], "high_risk": true, "model": "model-b-deep", "required_capabilities": [ "code", "review" ], "reviewer": { "model": "model-a-fast" }, "task_id": "T2-fix", "tokens": 8000 } ], "schema": "cain42.a75.orchestration.record.v1" }, "subject_sha256": "a99de43ae82914f0c2a50f3021cb06f9701cd4b7def6732cb096c558a792f4dc", "verdict": "SOUND" } ], "run_hash": "882cc456706bba088148ae016729fd03213ea8bb3bee35ba0f91ae6d586e4722", "schema": "cain42.a75.orchestration.run.v1", "sku": "CAIN-A75-013", "verdict": "VERIFIED", "verdicts": [ "SOUND" ] }, "scenarios": [ { "name": "cross_lab_routing", "run_hash": "882cc456706bba088148ae016729fd03213ea8bb3bee35ba0f91ae6d586e4722", "success_rate": 1.0, "total_cost": 17.0, "verdict": "VERIFIED" }, { "expected": "SOUND", "name": "honest_cross_lab_routing", "verdict": "SOUND" }, { "expected": "UNSOUND", "name": "unknown_model", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "capability_missing", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "context_too_small", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "provider_not_allowed", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "residency_violation", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "task_cost_over_ceiling", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "fallback_points_at_primary", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "fallback_unknown", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "high_risk_no_review", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "review_same_lab", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "duplicate_task", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "duplicate_model", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "missing_outcome", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "success_below_gate", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "run_budget_exceeded", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "claim_overstates", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "unknown_claim", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "wrong_schema", "verdict": "UNSOUND" }, { "expected": "UNSOUND", "name": "garbage", "verdict": "UNSOUND" } ], "summary": "VERIFIED: capabilities, context, provider/residency, cost, fallbacks and cross-lab review recomputed; 20/20 route-failure cases as expected" }, "schema": "cain42.agentic75.findings.v1", "sku": "CAIN-A75-013" }