open/ms-ai-architect
1
0
Fork 0
Code Issues Releases Activity

feat(ms-ai-architect): R7.5 bølge 3 — payload-21..30 dømt+ingestet (206 claims, 48 flagg; ledger 212→222) [skip-docs]

Browse source
This commit is contained in:
Kjell Tore Guttormsen 2026-08-01 20:48:15 +02:00
commit 353b976425
1 changed files with 687 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

688
scripts/kb-eval/data/judge-pass-manifest.json
View file

@ -4,7 +4,7 @@
"derived_from": "scripts/kb-update/data/full-pass-worklist.json (243 due, all never-verified)",
"cadence": "R7–R10 (5 økter × ~49, 8–10 samtidige)",
"batch": "R7.1",
"count": 212,
"count": 222,
"generated": "2026-07-18"
},
"files": [
@ -12374,6 +12374,65 @@
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/azure-cost-management-ai.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 18,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/azure-cost-management-ai.md#1",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/cost-management-billing/costs/overview-cost-management",
"evidence_quote": "",
"reason": "Oversiktssiden beskriver Cost Management som en FinOps-verktøysuite tilgjengelig for alle med tilgang til billing account, subscription, resource group eller management group, men ingen Learn-side oppgir en GA-status og det finnes heller ingen preview-merking, så påstanden kan verken bekreftes eller avkreftes.",
"file": "skills/ms-ai-security/references/cost-optimization/azure-cost-management-ai.md",
"line": 4,
"claim": "Azure Cost Management (kostnadsovervåking, budsjettering og optimalisering for Azure-ressurser) har status GA (generelt tilgjengelig).",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/azure-cost-management-ai.md#5",
"judge_verdict": "not_grounded",
"rule": "R7",
"evidence_url": "https://learn.microsoft.com/azure/cost-management-billing/costs/quick-acm-cost-analysis",
"evidence_quote": "Forecasting employs a time series linear regression model, which adjusts to factors such as reserved instance purchases that temporarily affect forecasted costs.",
"reason": "Den kanoniske forecast-siden oppgir en time series linear regression-modell med tilbakeblikk på 28 eller 90 dager og prognose opptil ett år framover; ingen 36-timers forecast-algoritme er dokumentert — de 36 timene gjelder anomaly detection, så tallet er feilkoblet til forecast-alerts.",
"file": "skills/ms-ai-security/references/cost-optimization/azure-cost-management-ai.md",
"line": 38,
"claim": "Forecast alerts i Azure Cost Management bygger på en 36-timers forecast-algoritme.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/azure-cost-management-ai.md#11",
"judge_verdict": "not_grounded",
"rule": "R7",
"evidence_url": "https://learn.microsoft.com/azure/cost-management-billing/costs/tutorial-acm-create-budgets",
"evidence_quote": "If you want to filter the new costs so that budgets are evaluated against first party Azure consumption charges only, add the following filters to your budget:",
"reason": "Kilden foreskriver inkluderende filtre (Publisher Type: Azure og Charge Type: Usage) for å holde kjøp utenfor budsjettevalueringen, og Budgets-API-ets BudgetOperatorType har bare verdien «In» ifølge create-or-update-referansen, så filteruttrykket «ChargeType != Purchase» finnes ikke slik påstanden oppgir det.",
"file": "skills/ms-ai-security/references/cost-optimization/azure-cost-management-ai.md",
"line": 128,
"claim": "Budsjetter i Azure Cost Management kan filtrere bort kjøp (reservations/savings plans) med filteret `ChargeType != Purchase`.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/azure-cost-management-ai.md#17",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/understand/analyze-unexpected-charges",
"evidence_quote": "If you get an error message indicating that you reached the limit of five alerts per subscription, consider editing an existing anomaly alert rule.",
"reason": "Delen om at budsjetter og varsler er gratis stemmer, men den lastbærende delen «ubegrenset antall ... alerts» motsies: anomaly-alerts er begrenset til fem per abonnement, og et budsjett kan ha maks fem terskler og fem e-postmottakere ifølge tutorial-acm-create-budgets.",
"file": "skills/ms-ai-security/references/cost-optimization/azure-cost-management-ai.md",
"line": 218,
"claim": "Budgets og alerts i Azure Cost Management er gratis, med ubegrenset antall budsjetter og alerts.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/batch-processing-cost-reduction.md",
"batch": "R7.1",
@ -12397,6 +12456,207 @@
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 15,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/budget-forecasting-ai-projects.md#1",
"judge_verdict": "not_grounded",
"rule": "R2",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/costs/quick-acm-cost-analysis",
"evidence_quote": "Forecast costs are available in both smart and custom views when you select either an area or column stacked chart type. In either case, the forecast is calculated the same way based on your historical usage patterns for up to a year in the future.",
"reason": "Den kanoniske forecasting-seksjonen i Cost Analysis beskriver ÉN innebygd prognosemetode med horisont opptil ett år, og hverken firedelingen eller de navngitte metodene «AutoML-basert (3-24 måneder)» og «Hybrid approach (6-36 måneder)» finnes på siden; AutoML sin forecast_horizon er dessuten dokumentert i enheter av tidsseriens frekvens, ikke i måneder.",
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"line": 39,
"claim": "Forecasting i Azure Cost Management deles i fire metoder: Native Cost Analysis Forecast (1-12 måneder) | AutoML-basert forecasting (3-24 måneder) | Manual projection (variabel horisont) | Hybrid approach (6-36 måneder).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/budget-forecasting-ai-projects.md#2",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/foundry/concepts/manage-costs",
"evidence_quote": "",
"reason": "Den siterte Foundry-kostnadssiden og FinOps-rammeverkets forecasting-side omtaler kostnadsdrivere som compute, storage, networking og dataoverføring som en åpen eksempelliste, men ingen learn.microsoft.com-side fastsetter en femdelt budsjettdimensjon-taksonomi for AI-kostnader — påstanden kan verken bekreftes eller motbevises.",
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"line": 52,
"claim": "AI-kostnader segmenteres langs fem budsjettdimensjoner: Compute (training GPU-timer, inference TPM/RPM, PTU-hosting) | Storage (treningsdata, modellartefakter, feature stores, logging) | Networking (dataoverføring, API-kall, kryss-regional replikering) | Licensing (modell-API-er med token-kostnad, fine-tuning, commitment tiers) | Operational (Monitoring, Log Analytics, Application Insights).",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/budget-forecasting-ai-projects.md#7",
"judge_verdict": "not_grounded",
"rule": "R1",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/costs/quick-acm-cost-analysis",
"evidence_quote": "For accurate long-term forecasting, it's essential to have sufficient historical data. New subscriptions or contracts with limited historical data might result in less accurate forecasts. At least 90 days of historical data are recommended for a more precise annual forecast.",
"reason": "Siden oppgir ingen hard minimumsgrense på 10 dager; de operative tallene er en lookback-periode på 28 dager for prognoser opptil 28 dager og en anbefaling om minst 90 dagers historikk — den nedre grensen på 10 dager underdriver kravet med langt mer enn det dobbelte og villeder leseren.",
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"line": 259,
"claim": "Forecast-baseline i Azure Cost Management krever minimum 10 dager med historikk.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/budget-forecasting-ai-projects.md#8",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/costs/tutorial-acm-create-budgets",
"evidence_quote": "Budgets are supported for the following types of Azure account types and scopes:\n\n- Azure role-based access control (Azure RBAC) scopes\n\n - Management group\n - Subscription\n - Resource group",
"reason": "Kilden lister management group i tillegg til subscription og resource group — samt EA-scopene billing account, department og enrollment account og MCA-scopene billing profile, invoice section og customer — så påstanden om at budsjetter «kun» kan settes på subscription eller resource group er for snever.",
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"line": 260,
"claim": "Budsjetter i Azure Cost Management kan kun settes på subscription- eller resource group-scope.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/budget-forecasting-ai-projects.md#13",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/provisioned-throughput-billing",
"evidence_quote": "With Reservations, you commit to payment for a fixed number of PTUs over a one-month or one-year term, and in return, you receive a discounted effective $/PTU/hr rate.",
"reason": "Gjeldende dokumentasjon oppgir bindingstid på én måned eller ett år for PTU-reservasjoner (og én måned for commitment-perioden), ikke 1-3 år, og timebasert PTU-bruk er dessuten mulig helt uten binding.",
"file": "skills/ms-ai-security/references/cost-optimization/budget-forecasting-ai-projects.md",
"line": 386,
"claim": "Azure commitment tiers (Provisioned Throughput Units) krever langsiktig binding på 1-3 år.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/cost-allocation-chargeback.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 19,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/cost-allocation-chargeback.md#1",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/cloud-adoption-framework/scenarios/ai/platform/governance",
"evidence_quote": "",
"reason": "Verken den siterte CAF-siden eller allocate-costs-siden oppgir noen GA- eller preview-status for cost allocation; ingen Learn-side sier generally available, så påstanden kan hverken bekreftes eller avkreftes.",
"file": "skills/ms-ai-security/references/cost-optimization/cost-allocation-chargeback.md",
"line": 4,
"claim": "Cost allocation og chargeback-funksjonaliteten som beskrives i filen har status GA (generelt tilgjengelig).",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/cost-allocation-chargeback.md#7",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/costs/enable-tag-inheritance",
"evidence_quote": "Inherited tags are applied to child resource usage records and not the resources themselves.",
"reason": "Første del om Azure Policy og tagging-strategier holder, men den siterte siden avviser eksplisitt den andre lastbærende delen: arvede tags legges på child resource usage records, ikke på ressursene selv — og oversiktssiden legger til at de ikke er tilgjengelige for andre tjenester som Azure Policy (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/cost-allocation-chargeback.md",
"line": 51,
"claim": "Azure Policy kan håndheve tagging-strategier, og tag inheritance propagerer tags fra subscription/resource group ned til child resources.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/cost-allocation-chargeback.md#15",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/cloud-computing/finops/toolkit/power-bi/governance",
"evidence_quote": "The Summary page provides a summary of subscriptions, resource types, resources, and regions across your environment.",
"reason": "Chargeback-siden (subscription/resource group/resource-nivå) og Commitments-siden i Cost summary stemmer, men Governance-rapportens Summary-side oppsummerer subscriptions, resource types, resources og regions — ikke tagging compliance, som ligger på en egen Policy compliance-side; én lastbærende del er dermed feil (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/cost-allocation-chargeback.md",
"line": 226,
"claim": "FinOps Toolkits Power BI-rapporter omfatter: Cost Summary → Commitments (amortized cost for reservations og savings plans) | Rate Optimization → Chargeback (tabell på subscription/resource group/resource-nivå) | Governance → Summary (tagging compliance).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/cost-allocation-chargeback.md#16",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/cost-management-billing/costs/overview-cost-management",
"evidence_quote": "",
"reason": "Ingen Learn-side oppgir at Cost Management er inkludert uten ekstra kostnad for EA-, MCA- og Pay-As-You-Go-kunder, eller at cost allocation og chargeback er uten lisenskostnad; oversiktssiden sier bare hvem tjenesten er tilgjengelig for, så prispåstanden kan ikke verifiseres.",
"file": "skills/ms-ai-security/references/cost-optimization/cost-allocation-chargeback.md",
"line": 293,
"claim": "Azure Cost Management er inkludert uten ekstra kostnad for alle EA-, MCA- og Pay-As-You-Go-kunder, og det er ingen lisenskostnad for cost allocation og chargeback-funksjonalitet.",
"disposition": "unsourced"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 24,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#8",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/search/semantic-how-to-enable-disable",
"evidence_quote": "",
"reason": "Learn bekrefter at semantisk rangering har en gratis månedskvote og deretter forbruksbasert pris, men ingen Learn-side oppgir tallet 1 000 — det står kun på den JS-genererte prissiden.",
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"line": 81,
"claim": "Azure AI Search semantic ranker: de første 1 000 forespørslene per måned er gratis, deretter faktureres per 1 000 forespørsler.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#13",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/microsoft-copilot-studio/requirements-messages-management",
"evidence_quote": "| Agent flow actions *per 100 actions*^3^ | 13 Copilot Credits | No charge |",
"reason": "Klassisk svar (1 credit) og generativt svar (2 credits) stemmer, men agent flow-satsen er 13 Copilot Credits per 100 handlinger og ikke 1 credit per handling — én lastbærende del er feil.",
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"line": 95,
"claim": "Copilot Credit-forbruk i Copilot Studio: standard melding (ikke-generativ AI) 1 credit | generativt AI-svar (GenAnswers, orchestration) 2 credits | agent flow action 1 credit.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#17",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/ai-services/content-moderator/content-moderator",
"evidence_quote": "| F0 (Free) | 5K free transactions/month | 5 RPS |",
"reason": "Kilden oppgir 5 000 gratis transaksjoner per måned for F0, ikke per 20 dager.",
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"line": 119,
"claim": "Azure AI Content Safety gratisnivå F0 gir 5 000 transaksjoner per 20 dager.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#19",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/ai-services/document-intelligence/service-limits?view=doc-intel-4.0.0",
"evidence_quote": "",
"reason": "Tjenestegrensesiden for Document Intelligence viser til prissiden for F0-månedskvoten og oppgir ikke 500 sider selv; ingen Learn-dokumentasjonsside oppgir verdien.",
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"line": 128,
"claim": "Azure AI Document Intelligence gratisnivå F0 gir 500 sider per måned.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/deterministic-cost-calculation-model.md#21",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/azure-monitor/logs/cost-logs",
"evidence_quote": "Billed per GB. Includes 31 days of analytics retention at no extra cost.",
"reason": "Kilden sier at Analytics Logs inkluderer 31 dagers oppbevaring uten ekstra kostnad, ikke 0–90 dager.",
"file": "skills/ms-ai-security/references/cost-optimization/deterministic-cost-calculation-model.md",
"line": 136,
"claim": "Azure Monitor-datalagring (retention) er inkludert uten ekstra kostnad i 0–90 dager; lagring utover 90 dager faktureres per GB per måned.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/gpt5-gpt41-pricing-models.md",
"batch": "R7.1",
@ -12492,6 +12752,432 @@
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/inference-endpoint-cost-optimization.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 19,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/inference-endpoint-cost-optimization.md#6",
"judge_verdict": "not_grounded",
"rule": "R2",
"evidence_url": "https://learn.microsoft.com/en-us/azure/machine-learning/how-to-autoscale-endpoints?view=azureml-api-2",
"evidence_quote": "To implement autoscale for an online endpoint, you enable autoscale settings and then define a default autoscale profile that specifies the minimum, maximum, and default scale set capacity.",
"reason": "Den kanoniske siden for autoskalering av online endpoints oppgir minimum, maximum og default samt scale-out- og scale-in-regler med cooldown, men ingen «idle time before scale-down» — den parameteren hører til AmlCompute-klynger og finnes ikke i endepunkt-autoskaleringen.",
"file": "skills/ms-ai-security/references/cost-optimization/inference-endpoint-cost-optimization.md",
"line": 49,
"claim": "Autoscaling-konfigurasjonen for online endpoints består av parameterne: Minimum instances | Maximum instances | Default instances | Scale-out threshold | Scale-in threshold | Cooldown period | Idle time before scale-down.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/inference-endpoint-cost-optimization.md#11",
"judge_verdict": "not_grounded",
"rule": "R7",
"evidence_url": "https://learn.microsoft.com/en-us/azure/machine-learning/monitor-azure-machine-learning-reference?view=azureml-api-2",
"evidence_quote": "| **CPU Memory Utilization Percentage**Percentage of memory utilization on an instance. Utilization is reported at one minute intervals. | `CpuMemoryUtilizationPercentage` | No | Percent | Minimum, Maximum, Average | `instanceId` | PT1M | No |",
"reason": "CpuUtilizationPercentage, GpuUtilizationPercentage, RequestLatency og RequestsPerMinute finnes med de oppgitte scopene, men ingen metrikk heter MemoryUtilizationPercentage — minnemetrikkene heter CpuMemoryUtilizationPercentage og GpuMemoryUtilizationPercentage, og det er selve navnet påstanden bærer.",
"file": "skills/ms-ai-security/references/cost-optimization/inference-endpoint-cost-optimization.md",
"line": 73,
"claim": "Metrikker tilgjengelige for autoscaling av online endpoints: CpuUtilizationPercentage (scope deployment) | RequestLatency (scope endpoint) | RequestsPerMinute (scope endpoint) | GpuUtilizationPercentage (scope deployment med GPU) | MemoryUtilizationPercentage (scope deployment).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/inference-endpoint-cost-optimization.md#16",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/machine-learning/how-to-manage-optimize-cost?view=azureml-api-2",
"evidence_quote": "",
"reason": "Serverless-halvdelen er dokumentert andre steder (regionstøtte må verifiseres per modell), men ingen Learn-side oppgir at Norway East eller Norway West kan velges som compute-region for managed compute — AzureML sin regionliste ligger på den JS-genererte products-by-region-siden, så påstanden lar seg verken bekrefte eller avkrefte.",
"file": "skills/ms-ai-security/references/cost-optimization/inference-endpoint-cost-optimization.md",
"line": 376,
"claim": "Managed compute gir full kontroll over compute-region, slik at Norway East eller Norway West kan velges for datalagring i Norge, mens serverless har begrenset region-valg og krever verifisering av at modellene støtter norske regioner.",
"disposition": "unsourced"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 21,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#10",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/cloud-adoption-framework/scenarios/ai/plan",
"evidence_quote": "",
"reason": "Den siterte CAF-siden nevner ikke PTU-terskler i det hele tatt, og den kanoniske PTU-siden gir bare kvalitative kriterier (forutsigbar trafikk, latenssensitivitet, produksjonsvolum) uten noe tokenvolum per måned, så tallet 150 millioner kan verken bekreftes eller motbevises på Microsoft Learn.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 70,
"claim": "Ved over 150 millioner tokens per måned bør Provisioned Throughput Units (PTU) vurderes for Azure OpenAI ved stabil workload.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#12",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/azure/ai-services/translator/service-limits",
"evidence_quote": "The hourly quota should be consumed evenly throughout the hour. For example, at the F0 tier limit of 2 million characters per hour, characters should be consumed no faster than roughly 33,300 characters per minute.",
"reason": "Translator-delen er feil: gratisnivået F0 er 2 millioner tegn per time og siden sier uttrykkelig at det belastes per tegn og ikke per forespørsel, så «20 transaksjoner per minutt» er ikke gjeldende grense — og siden denne lastbærende delen er motsagt, faller hele claimet selv om 5 000 transaksjoner per måned for tekstanalyse stemmer med F-nivåets prediksjonskvote.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 75,
"claim": "Flere Azure AI Services har free tier: 5000 transaksjoner per måned for Text Analytics og 20 transaksjoner per minutt for Translator.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#15",
"judge_verdict": "not_grounded",
"rule": "R2",
"evidence_url": "https://learn.microsoft.com/en-us/ai-builder/credit-management",
"evidence_quote": "Some Microsoft products like Power Apps per app plan, Power Apps per user plan, and Power Automate Premium (previously Power Automate per user with attended RPA) plan include AI Builder capacity.",
"reason": "Jeg sjekket den kanoniske opplistingen «Number of AI Builder credits by license», og Microsoft 365 E3 finnes ikke i den — seeded credits kommer kun fra Power Apps- og Power Automate-planer, Dynamics 365 F&O og Cloud for Sustainability — så den lastbærende delen om at E3 inkluderer seeded AI Builder-credits er uriktig, selv om Copilot Chat-delen og add-on-delen stemmer.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 358,
"claim": "Microsoft 365 E3 inkluderer M365 Copilot Chat (web) og seeded AI Builder-credits (til november 2026), og M365 Copilot må kjøpes som add-on.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#16",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/microsoft-365/copilot/microsoft-365-copilot-licensing",
"evidence_quote": "",
"reason": "300-brukergrensen for Business Premium er godt dokumentert, men ingen Microsoft Learn-side sier at Business Premium har de samme AI-kapabilitetene som E3 — lisenssiden skiller tvert imot mellom tilbudene Microsoft 365 Copilot Business og Microsoft 365 Copilot uten å angi kapabilitetsforskjeller, og E3/E5/E7-sammenligningen omtaler ikke Business Premium — så likhetspåstanden kan verken bekreftes eller motbevises.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 360,
"claim": "Microsoft 365 Business Premium har samme AI-kapabiliteter som E3, men er begrenset til under 300 brukere.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#17",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/microsoft-365-copilot/extensibility/cost-considerations",
"evidence_quote": "Users with a Microsoft 365 Copilot license don't incur additional charges.",
"reason": "Delen om at extensibility er metered for add-on-lisensen er motsagt: lisenstabellen sier at Microsoft 365 Copilot add-on gir «No extra charges for accessing or using extensibility features (Copilot connectors, agents, plugins)», og forbruksbasert Copilot Credit-fakturering gjelder kun brukere uten add-on som treffer delte tenant-data, så én lastbærende del faller selv om Word/Excel/Teams-, work chat- og Pages/Notebooks-delene stemmer.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 361,
"claim": "Microsoft 365 Copilot add-on gir full Copilot i Word | Excel | Teams m.fl., work-grounded chat og Copilot Pages/Notebooks, mens extensibility (connectors og custom agents med work data) er metered.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#18",
"judge_verdict": "not_grounded",
"rule": "R3",
"evidence_url": "https://learn.microsoft.com/en-us/microsoft-copilot-studio/billing-licensing",
"evidence_quote": "Starting on September 1, 2025, the common currency for agents changed from messages to Copilot Credits.",
"reason": "Hele claimet er formulert i den erstattede enheten meldinger, og gjeldende dokumentasjon måler standalone-kapasitet i Copilot Credits samlet på tenant-nivå med overage håndtert via reallokering, kjøp av mer kapasitet eller pay-as-you-go og håndheving ved 125 % — det finnes ingen fakturering i blokker på 10 000 meldinger, så rammeverket claimet beskriver eksisterer ikke lenger.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 362,
"claim": "Copilot Studio standalone inkluderer 25 000 meldinger per måned per tenant, og overage faktureres i blokker på 10 000 meldinger.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#20",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/search/search-limits-quotas-capacity",
"evidence_quote": "Dedicated: Fixed pricing measured by Search Units (SUs). Service tier options include: Basic, Standard (S1-S3, including S3 HD), Storage Optimized (L1-L2), and a Free tier with limited search service capabilities.",
"reason": "Fastprisdelen stemmer for Dedicated-modellen som S1 tilhører, men delen om ekstra kostnad for lagring over 100 GB er uriktig: per-GB-lagringsbetaling finnes bare i Serverless-forhåndsvisningen, mens S1 har en partisjonsgrense på 160 GB (25 GB for tjenester opprettet før 3. april 2024) som utvides ved å legge til partisjoner, altså søkeenheter, og 100 GB er ingen S1-terskel.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 365,
"claim": "Azure AI Search S1 faktureres som fast månedsavgift, med ekstra kostnad for lagring over 100 GB.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/licensing-compliance-cost-avoidance.md#21",
"judge_verdict": "not_grounded",
"rule": "R1",
"evidence_url": "https://learn.microsoft.com/en-us/azure/search/search-limits-quotas-capacity",
"evidence_quote": "Each index supports up to the following number of documents: 24 billion on Basic, S1, S2, and S3",
"reason": "Dette er en øvre grense, og taket har flyttet seg dramatisk: gjeldende side oppgir 24 milliarder dokumenter per indeks for S1, ikke 1 million, og siden S3 har nøyaktig samme dokumentgrense som S1 er også kontrasten «S3 for større volumer» feil på dokumentantall.",
"file": "skills/ms-ai-security/references/cost-optimization/licensing-compliance-cost-avoidance.md",
"line": 388,
"claim": "Azure AI Search S1-tier dekker opptil 1 million dokumenter, mot S3-tier for større volumer.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 23,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#1",
"judge_verdict": "not_grounded",
"rule": "R1",
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
"evidence_quote": "`gpt-5.6-sol` (2026-07-09)",
"reason": "Påstanden angir generasjonens ytterpunkt (utvidet til GPT-5.5), men katalogen lister nå GPT-5.6-serien (gpt-5.6-sol/-terra/-luna, 2026-07-09) som nyeste GPT-5-generasjon; det oppgitte taket er passert, og modelleksemplene er mer enn én generasjon bak.",
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"line": 27,
"claim": "GPT-5-generasjonen er utvidet til GPT-5.1 til GPT-5.5, slik at referansens modelleksempler (GPT-5 / GPT-4.1 / GPT-4o-mini) er én generasjon bak.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#5",
"judge_verdict": "not_grounded",
"rule": "R2",
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
"evidence_quote": "Image generation models include `gpt-image-1`, `gpt-image-1-mini`, `gpt-image-1.5`, and `gpt-image-2`.",
"reason": "Den kanoniske opplistingen, som eksplisitt utelater utfasede og pensjonerte modeller, nevner ikke DALL-E i det hele tatt — bildegenerering er nå gpt-image-familien. Embeddings og Whisper står fortsatt, men DALL-E-leddet er feil.",
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"line": 52,
"claim": "Modellklassen spesialiserte modeller i Azure AI består av Embeddings | Whisper | DALL-E.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#12",
"judge_verdict": "not_grounded",
"rule": "R7",
"evidence_url": "https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/reasoning",
"evidence_quote": " reasoning={\n \"effort\": \"medium\",",
"reason": "Kallet client.responses.create stemmer, men i Responses API-eksempelet settes nivået via objektet reasoning med feltet effort; reasoning_effort er parameteren i client.chat.completions.create. Den lastbærende strengen en leser ville kopiere inn i kode er dermed feil for det API-et påstanden oppgir.",
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"line": 172,
"claim": "GPT-5 kalles via Responses API (client.responses.create) med parameteren reasoning_effort for å styre resonneringsnivå.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#17",
"judge_verdict": "not_grounded",
"rule": "R4",
"evidence_url": "https://learn.microsoft.com/microsoft-copilot-studio/prompt-model-availability",
"evidence_quote": "| GPT-4o | July 2025 | GPT-4.1 |",
"reason": "I Copilot Studio står GPT-4o kun i tabellen over pensjonerte modeller (pensjonert juli 2025, erstattet av GPT-4.1) og brukes videre bare i US government-regioner; gjeldende liste er GPT-4.1 mini (standard), GPT-4.1 og GPT-5-serien. GPT-4o er fortsatt gjeldende i Azure OpenAI (jf. claim 3), men det er en annen produktflate.",
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"line": 328,
"claim": "Copilot Studio bruker Azure OpenAI-modeller (GPT-4o eller GPT-4.1-serien).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/model-selection-price-performance.md#20",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/microsoft-copilot-studio/billing-licensing",
"evidence_quote": "Copilot Credits are the common currency across Copilot Studio capabilities. You can get them through pay-as-you-go meters, prepurchase plans, and Copilot Credit prepaid pack subscriptions.",
"reason": "Begge lastbærende ledd er motsagt: brukerlisensen for Copilot Studio er gratis og kapasiteten kjøpes på tenant-nivå (prepaid pakke eller pay-as-you-go), og inferens betales som Copilot Credits per forbruk — den er ikke inkludert i en lisens per bruker per måned.",
"file": "skills/ms-ai-security/references/cost-optimization/model-selection-price-performance.md",
"line": 393,
"claim": "Copilot Studio lisensieres per bruker per måned, og inferenskostnader er inkludert i lisensen.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 27,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#13",
"judge_verdict": "not_grounded",
"rule": "R3",
"evidence_url": "https://learn.microsoft.com/en-us/azure/architecture/ai-ml/guide/azure-openai-gateway-multi-backend",
"evidence_quote": "They can be classified in four key topologies:",
"reason": "Kilden opererer nå med fire topologier, ikke tre: den splitter samme-region-tilfellet i «Multiple instances in a single region and single subscription» og «Multiple instances in a single region across multiple subscriptions», så påstandens tredelte inndeling er erstattet.",
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"line": 65,
"claim": "Gateway-topologiene for flere Azure OpenAI-deployments er: Single Instance + Multiple Deployments | Multiple Instances (Same Region) | Multiple Instances (Multi-Region).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#15",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure",
"evidence_quote": "- 1,047,576 - 300,000 (standard deployments) - 128,000 (provisioned managed and batch deployments)",
"reason": "Kjernen holder — model-router-siden sier «The effective context window is limited by the smallest underlying model» — men den lastbærende tilleggspåstanden om at GPT-4.1-serien er 128k er feil: kilden oppgir 1 047 576 for gpt-4.1/-mini/-nano, med 300 000 for standard-deployments og 128 000 kun for provisioned managed og batch.",
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"line": 100,
"claim": "Model Routers context window er begrenset til den minste underliggende modellen — 128k for GPT-4.1-serien.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#20",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/api-management/backends",
"evidence_quote": "",
"reason": "Den kanoniske siden for backend circuit-breaker-regler dokumenterer funksjonen uten preview-merking, men slår heller ingen steder fast at den er GA, og både REST-referansen («current-preview») og Bicep/ARM-eksemplene bruker preview-API-versjoner; verken den siterte siden eller søk gir en side som bekrefter eller motsier preview-statusen.",
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"line": 376,
"claim": "Circuit breaker-policyen i Azure API Management (backend circuit-breaker rules) er i preview.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#24",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/foundry/openai/how-to/model-router",
"evidence_quote": "Don't deploy the underlying chat models separately. Model router works independently of your other deployed models.",
"reason": "Den lastbærende første delen er motsagt — kunden deployer ikke de underliggende modellene i det hele tatt, så noe krav om «samme data zone» for dem finnes ikke; kilden snur også unntaket: Claude-modellene er nettopp de eneste som MÅ deployes selv, til samme Foundry-konto med matchende SKU.",
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"line": 438,
"claim": "Underliggende modeller i Model Router må deployes i samme data zone, med unntak av Claude-modellene som krever separate deployments.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/multi-model-strategy-costs.md#25",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/foundry/openai/concepts/provisioned-throughput-billing",
"evidence_quote": "In exchange for a 1-month or 1-year term commitment, you receive a discounted effective $/PTU/hr rate compared to hourly billing.",
"reason": "Kilden oppgir 1-måneds eller 1-års bindingstid for Azure Reservations på provisioned throughput; 3-årsalternativet påstanden nevner finnes ikke, og siden gjentar «you commit to payment for a fixed number of PTUs over a one-month or one-year term».",
"file": "skills/ms-ai-security/references/cost-optimization/multi-model-strategy-costs.md",
"line": 460,
"claim": "Azure Reservations for provisioned throughput kan kjøpes som 1-års eller 3-års avtale.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 20,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#4",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/logs/data-retention-configure",
"evidence_quote": "Basic plan tables have a fixed period of 30 days for queries while Auxiliary plan tables can be queried for the total retention period.",
"reason": "Redusert ingestion-pris og egen query-kostnad stemmer, men kilden oppgir 30 dagers fast spørreperiode for Basic — ikke 8 dager; den lastbærende tidsangivelsen er motsagt (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 48,
"claim": "Basic Logs har redusert ingestion-pris, egen query-kostnad og begrenset query-tid på 8 dager.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#5",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/logs/data-retention-configure",
"evidence_quote": "Basic plan tables have a fixed period of 30 days for queries while Auxiliary plan tables can be queried for the total retention period.",
"reason": "Laveste ingestion-pris stemmer, men kilden sier at Auxiliary-tabeller kan spørres i hele total retention-perioden (og table plan-tabellen gir full KQL på én tabell), så påstanden om at de kun kan spørres via search jobs er motsagt (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 49,
"claim": "Auxiliary Logs har lavest ingestion-pris og kan kun spørres via search jobs.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#11",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/app/opentelemetry-sampling",
"evidence_quote": "The Application Insights OpenTelemetry distros include a default sampler. The specific sampler and its rate depend on the language and distro version.",
"reason": "De to strategiene (fixed-rate med ratio 0–1 og rate-limited i traces/sek) stemmer, men kilden sier nå at distroene inkluderer en default sampler — påstandens lastbærende del om at sampling ikke er på som default er motsagt (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 64,
"claim": "Azure Monitor OpenTelemetry-distroen har ikke sampling på som default og støtter to strategier: fixed-rate (ratio 0–1, f.eks. 0.1 = ~10 %) | rate-limited (traces/sek, f.eks. 5.0 = fem traces/sek).",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#13",
"judge_verdict": "not_grounded",
"rule": "",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/app/opentelemetry-configuration",
"evidence_quote": "Starting with Java agent version 3.4.0, rate-limited sampling is available and is now the default.",
"reason": "Kilden oppgir Java-agent 3.4.0 som versjonen der rate-limited sampling ble tilgjengelig og default, mens påstanden krever 3.7.5 eller nyere — versjonsverdien avviker fra kilden.",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 114,
"claim": "Rate-limited sampling med requestsPerSecond konfigureres i Java-agent versjon 3.7.5 eller nyere.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#16",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/logs/data-retention-configure",
"evidence_quote": "Basic plan tables have a fixed period of 30 days for queries while Auxiliary plan tables can be queried for the total retention period.",
"reason": "Analytics 30–730 dager stemmer, men kilden gir Basic fast 30-dagers spørreperiode (ikke 8 dager) og lar Auxiliary spørres i hele total retention-perioden, ikke kun long-term; to lastbærende deler er motsagt (R8).",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 186,
"claim": "Retention per table plan: Analytics 30–730 dager | Basic 8 dager interactive pluss long-term | Auxiliary kun long-term.",
"disposition": "outdated"
},
{
"id": "ms-ai-security/cost-optimization/observability-cost-reduction.md#18",
"judge_verdict": "not_grounded",
"rule": "R8",
"evidence_url": "https://learn.microsoft.com/en-us/azure/azure-monitor/app/managed-workspaces",
"evidence_quote": "Classic Application Insights resources have been retired and don't ingest new telemetry.",
"reason": "Kostnadsdelene stemmer (classic kan ikke bruke commitment tiers, workspace-based kan bruke commitment tiers og Basic Logs), men kilden sier at classic er retired og ikke lenger tar imot telemetri — en annen livssyklusstatus enn påstandens deprecated, og statusen er lastbærende (R8); den siterte URL-en gir 404, så jeg brukte gjeldende offisielle side.",
"file": "skills/ms-ai-security/references/cost-optimization/observability-cost-reduction.md",
"line": 220,
"claim": "Classic Application Insights er deprecated og støtter kun pay-as-you-go, mens workspace-based Application Insights lagrer data i Log Analytics workspace og kan bruke commitment tiers og Basic Logs.",
"disposition": "outdated"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/prompt-engineering-cost-reduction.md",
"batch": "R7.5",
"judged_at": "2026-08-01",
"per_file_verdict": "flagged",
"claim_count": 20,
"verified": null,
"verified_by": null,
"flags": [
{
"id": "ms-ai-security/cost-optimization/prompt-engineering-cost-reduction.md#13",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/foundry-classic/how-to/prompt-flow-tools/llm-tool",
"evidence_quote": "",
"reason": "Verken den siterte kostnadssiden eller den kanoniske LLM-verktøysiden for Prompt Flow lister hvilke modellfamilier verktøyet støtter, så ingen Microsoft Learn-side bekrefter eller motsier at GPT-4.1- og GPT-5-seriene støttes.",
"file": "skills/ms-ai-security/references/cost-optimization/prompt-engineering-cost-reduction.md",
"line": 250,
"claim": "Prompt Flow støtter gjeldende modeller, blant annet GPT-4.1-serien | GPT-5-serien.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/prompt-engineering-cost-reduction.md#15",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/manage-costs",
"evidence_quote": "",
"reason": "Den siterte kostnadssiden nevner ikke Copilot Studio, og søk i Copilot Studio-dokumentasjonen ga ingen side som sier noe om prompt caching er eksponert for brukeren.",
"file": "skills/ms-ai-security/references/cost-optimization/prompt-engineering-cost-reduction.md",
"line": 267,
"claim": "I Copilot Studio er prompt caching ikke eksponert til brukeren, selv om tjenesten bruker underliggende Azure OpenAI.",
"disposition": "unsourced"
},
{
"id": "ms-ai-security/cost-optimization/prompt-engineering-cost-reduction.md#17",
"judge_verdict": "source_silent",
"rule": "",
"evidence_url": "https://learn.microsoft.com/azure/foundry/concepts/manage-costs",
"evidence_quote": "",
"reason": "Ingen offisiell Microsoft Learn-side gir en slik regionanbefaling for norske virksomheter; treffene var community-svar på learn.microsoft.com/answers, som ikke er førstepartsdokumentasjon.",
"file": "skills/ms-ai-security/references/cost-optimization/prompt-engineering-cost-reduction.md",
"line": 292,
"claim": "Norske organisasjoner med residenskrav anbefales å bruke regionene Norway East | West Europe.",
"disposition": "unsourced"
}
]
},
{
"file": "skills/ms-ai-security/references/cost-optimization/token-counting-optimization.md",
"batch": "R7.1",