# Spor 1 — Edge-Set Classification Review (Step 4)

**Date:** 2026-07-03
**Scope:** every manifest entry flagged `reviewFlag:true` by the mechanical classifier (106 of 389)
**Mechanism:** main-context executor content review, two tiers — (1) title + section headings + intro per file, (2) full-text sampling where tier 1 was inconclusive (6 files). The plan prescribed read-only subagent fan-out; the executor's Agent tool is disabled during plan execution (trekexecute Hard Rule 10), so adjudication ran in the main context against the same criteria. Recorded as `resolvedBy: executor-content-review-2026-07-03`.

## Criteria (Port-1 closed type set)

- **reference** — dominant substance is machine-verifiable Microsoft product/service claims (verifiable against learn.microsoft.com); will carry a `**Source:**` URL.
- **template** — reusable fill-in scaffold; must never carry an MS source.
- **methodology** — reusable process/framework guidance; must never carry an MS source.
- **regulatory** — law/standard text or obligations derived directly from it.

Judged by content substance, not filename. No file was deferred — every flagged file had a defensible type.

## Outcome

Resolved: **106/106** (0 deferred). Distribution within the edge set: {"template":8,"methodology":20,"reference":69,"regulatory":9}.
Corpus-wide (389): reference 352 · methodology 20 · regulatory 9 · template 8.

Notable calls:
- `ms-ai-advisor/development/agent-framework.md` → **reference** (pure MS API fact base; the path heuristic tripped on 'framework').
- `responsible-ai/ai-ethics-in-public-sector.md` → **reference** (substance = MS RAI standard + Foundry capabilities) while `norwegian-public-sector-governance/public-sector-ai-ethics-framework.md` → **methodology** (substance = Norwegian governance landscape). Same theme, different dominant substance.
- `norwegian-public-sector-governance/norwegian-nlp-benchmarks.md` → **reference** with zero MS citations — expected to land in the honest `unsourced/deferred` bucket in Step 6.
- The 67 zero-MS engineering/infrastructure files (multi-modal, hybrid-edge, api-management, agent-orchestration, performance-scalability) are Azure product-fact files written without citations — all **reference**; Step 6 will defer them unless a defensible MS authority exists.

## Per-file adjudication

| File (under skills/) | Type | Classifier signal | Rationale |
|---|---|---|---|
| `ms-ai-advisor/references/architecture/adr-template.md` | template | ms-citation+path-conflict | MADR fill-in scaffold; MS citations are illustrative |
| `ms-ai-advisor/references/architecture/ai-utredning-template.md` | template | ms-citation+path-conflict | utredningsmal (fill-in document skeleton) grounded in utredningsinstruksen |
| `ms-ai-advisor/references/architecture/alternativanalyse-methodology.md` | methodology | path-heuristic | weighted MCA scoring process for architecture comparison |
| `ms-ai-advisor/references/architecture/capacity-feasibility-benchmarks.md` | methodology | path-heuristic | feasibility assessment framework (gap matrices, timeline validation) |
| `ms-ai-advisor/references/architecture/decision-trees.md` | methodology | path-heuristic | platform decision frameworks; embedded MS facts are supporting |
| `ms-ai-advisor/references/architecture/diagram-prompt-templates.md` | template | path-heuristic | prompt templates for Imagen 3 diagram generation |
| `ms-ai-advisor/references/architecture/poc-template.md` | template | ms-citation+path-conflict | POC plan/rubric/checklist scaffolds |
| `ms-ai-advisor/references/architecture/public-sector-checklist.md` | template | ms-citation+path-conflict | comprehensive fill-in checklist for public-sector adoption |
| `ms-ai-advisor/references/architecture/rag-maturity-model.md` | methodology | path-heuristic | maturity model + level-selection decision framework |
| `ms-ai-advisor/references/architecture/recommended-mcp-servers.md` | methodology | path-heuristic | advisory tool-selection guidance with matrix |
| `ms-ai-advisor/references/architecture/source-traceability-assumption-register.md` | methodology | path-heuristic | traceability/assumption-register process framework |
| `ms-ai-advisor/references/development/agent-framework.md` | reference | ms-citation+path-conflict | MS Agent Framework API facts, verified against MS Learn |
| `ms-ai-advisor/references/prompt-engineering/regulatory-and-compliance-prompting.md` | methodology | ms-citation+path-conflict | reusable compliance-aware prompting patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-compliance-and-audit-trails.md` | reference | path-heuristic | Azure agent audit/compliance product patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-cost-optimization-strategies.md` | reference | path-heuristic | Foundry/Azure cost-optimization product facts |
| `ms-ai-engineering/references/agent-orchestration/agent-ecosystem-and-plugin-marketplace.md` | reference | path-heuristic | agent ecosystem/marketplace platform patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-evaluation-testing-frameworks.md` | reference | ms-citation+path-conflict | Azure AI Evaluation SDK facts (GA/Preview status) |
| `ms-ai-engineering/references/agent-orchestration/agent-feedback-and-learning-loops.md` | reference | path-heuristic | Foundry continuous-evaluation product facts |
| `ms-ai-engineering/references/agent-orchestration/agent-latency-optimization.md` | reference | path-heuristic | Azure agent latency/performance product patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-monitoring-observability.md` | reference | path-heuristic | Azure tracing/observability product patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-routing-and-specialization.md` | reference | path-heuristic | Semantic Kernel/Azure routing product patterns |
| `ms-ai-engineering/references/agent-orchestration/agent-security-threat-modeling.md` | reference | path-heuristic | Azure agent security product patterns |
| `ms-ai-engineering/references/agent-orchestration/copilot-agent-integration-patterns.md` | reference | path-heuristic | Copilot Studio integration product facts |
| `ms-ai-engineering/references/agent-orchestration/declarative-vs-imperative-agent-design.md` | reference | path-heuristic | MS agent-platform design tradeoff facts |
| `ms-ai-engineering/references/agent-orchestration/multi-tenant-agent-isolation.md` | reference | path-heuristic | Azure OpenAI isolation model facts |
| `ms-ai-engineering/references/api-management/circuit-breaker-ai-resilience.md` | reference | path-heuristic | APIM circuit-breaker configuration facts |
| `ms-ai-engineering/references/api-management/load-balancing-openai-instances.md` | reference | path-heuristic | APIM backend-pool/load-balancing facts |
| `ms-ai-engineering/references/api-management/semantic-caching-apim.md` | reference | path-heuristic | APIM semantic-caching policy facts |
| `ms-ai-engineering/references/api-management/token-rate-limiting-policies.md` | reference | path-heuristic | APIM token-rate-limit policy facts |
| `ms-ai-engineering/references/data-engineering/data-quality-ai-frameworks.md` | reference | ms-citation+path-conflict | MS data-quality tooling (Purview/Fabric) facts |
| `ms-ai-engineering/references/mlops-genaiops/model-evaluation-frameworks.md` | reference | ms-citation+path-conflict | Azure ML/Foundry evaluation tooling facts |
| `ms-ai-engineering/references/multi-modal/accessibility-multimodal-ai.md` | reference | path-heuristic | Azure AI accessibility product facts (WCAG tooling) |
| `ms-ai-engineering/references/multi-modal/audio-video-transcription-workflow.md` | reference | path-heuristic | Azure Speech batch-transcription product facts |
| `ms-ai-engineering/references/multi-modal/azure-video-indexer-patterns.md` | reference | path-heuristic | Azure Video Indexer product facts |
| `ms-ai-engineering/references/multi-modal/cv-llm-integration.md` | reference | path-heuristic | Azure vision+LLM integration product facts |
| `ms-ai-engineering/references/multi-modal/dalle-image-generation.md` | reference | path-heuristic | Azure OpenAI DALL-E product facts |
| `ms-ai-engineering/references/multi-modal/document-vision-processing.md` | reference | path-heuristic | Document Intelligence product facts |
| `ms-ai-engineering/references/multi-modal/gpt4o-vision-architecture.md` | reference | path-heuristic | GPT-4o vision capability/token-model facts |
| `ms-ai-engineering/references/multi-modal/image-classification-understanding.md` | reference | path-heuristic | Azure vision model selection/training facts |
| `ms-ai-engineering/references/multi-modal/multimodal-content-safety.md` | reference | path-heuristic | Azure AI Content Safety product facts |
| `ms-ai-engineering/references/multi-modal/multimodal-evaluation-metrics.md` | reference | path-heuristic | multimodal evaluation metric/tooling facts |
| `ms-ai-engineering/references/multi-modal/multimodal-prompt-engineering.md` | reference | path-heuristic | multimodal prompting capability facts (GPT-4o) |
| `ms-ai-engineering/references/multi-modal/multimodal-rag-architecture.md` | reference | path-heuristic | Azure multimodal embedding/RAG product facts |
| `ms-ai-engineering/references/multi-modal/ocr-pipeline-architecture.md` | reference | path-heuristic | Azure OCR engine/pipeline product facts |
| `ms-ai-engineering/references/multi-modal/real-time-audio-api.md` | reference | path-heuristic | Azure OpenAI Realtime API product facts |
| `ms-ai-engineering/references/multi-modal/speech-to-ai-pipelines.md` | reference | path-heuristic | Azure Speech recognition pipeline facts |
| `ms-ai-engineering/references/multi-modal/text-to-speech-citizen.md` | reference | path-heuristic | Azure neural TTS product facts |
| `ms-ai-engineering/references/multi-modal/video-analysis-patterns.md` | reference | path-heuristic | Azure video analysis product facts |
| `ms-ai-engineering/references/multi-modal/whisper-speech-recognition.md` | reference | path-heuristic | Whisper on Azure product facts |
| `ms-ai-engineering/references/rag-architecture/rag-evaluation-frameworks.md` | reference | ms-citation+path-conflict | RAG evaluation tooling/SDK facts |
| `ms-ai-governance/references/monitoring-observability/application-insights-llm-monitoring.md` | reference | path-heuristic | Application Insights LLM telemetry facts |
| `ms-ai-governance/references/monitoring-observability/azure-monitor-setup-ai-workloads.md` | reference | path-heuristic | Azure Monitor configuration facts (Portal/PS/CLI) |
| `ms-ai-governance/references/norwegian-public-sector-governance/anskaffelser-ai-procurement-framework.md` | methodology | ms-citation+path-conflict | procurement process framework (lovgrunnlag + DFØ + SSA) |
| `ms-ai-governance/references/norwegian-public-sector-governance/dpia-norwegian-methodology-ai.md` | methodology | ms-citation+path-conflict | Datatilsynet DPIA process methodology for AI |
| `ms-ai-governance/references/norwegian-public-sector-governance/gevinstrealisering-dfo-methodology.md` | methodology | path-heuristic | DFØ 5-step benefit-realization process |
| `ms-ai-governance/references/norwegian-public-sector-governance/norwegian-nlp-benchmarks.md` | reference | path-heuristic | factual benchmark/model-comparison content; no MS authority → expected unsourced/deferred in Step 6 |
| `ms-ai-governance/references/norwegian-public-sector-governance/public-sector-ai-ethics-framework.md` | methodology | ms-citation+path-conflict | Norwegian ethics governance landscape/framework guidance |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-ai-threat-library.md` | methodology | path-heuristic | reusable AI threat catalog for ROS assessment (OWASP-anchored) |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-dpia-security-integration.md` | methodology | path-heuristic | assessment sequencing/integration process guide |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-maestro-multiagent.md` | methodology | path-heuristic | MAESTRO 7-layer assessment framework + checklist |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-methodology-ns5814-iso31000.md` | regulatory | path-heuristic | NS 5814/ISO 31000/ISO 23894/AI Act Art.9 standard-text mapping |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-report-templates.md` | template | path-heuristic | Quick/Full ROS report fill-in scaffolds |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-scoring-rubrics-7x5.md` | methodology | path-heuristic | 7x5 scoring rubric framework for ROS |
| `ms-ai-governance/references/norwegian-public-sector-governance/ros-sector-checklists.md` | template | path-heuristic | sector fill-in checklists (helse/transport/finans/...) |
| `ms-ai-governance/references/norwegian-public-sector-governance/samfunnsokonomisk-analyse-nnv.md` | methodology | path-heuristic | NNV/samfunnsøkonomisk analysis process (DFØ) |
| `ms-ai-governance/references/norwegian-public-sector-governance/statistical-ethics-ssa-methodology.md` | methodology | ms-citation+path-conflict | SSB statistical-ethics application methodology |
| `ms-ai-governance/references/norwegian-public-sector-governance/utredningsinstruksen-ai-methodology.md` | methodology | ms-citation+path-conflict | utredningsinstruksen applied as AI assessment process |
| `ms-ai-governance/references/responsible-ai/ai-act-annex-iii-checklist.md` | regulatory | ms-citation+path-conflict | Annex III high-risk categories/checkpoints (obligations from the Act) |
| `ms-ai-governance/references/responsible-ai/ai-act-classification-methodology.md` | regulatory | path-heuristic | classification per the Act’s own rules (Annex III, provider/deployer roles) |
| `ms-ai-governance/references/responsible-ai/ai-act-compliance-guide.md` | regulatory | ms-citation+path-conflict | the Act’s risk classes + obligations are the substance; MS tooling is a section |
| `ms-ai-governance/references/responsible-ai/ai-act-conformity-assessment.md` | regulatory | path-heuristic | Art. 43/47 + Annex IV conformity obligations |
| `ms-ai-governance/references/responsible-ai/ai-act-deployer-obligations.md` | regulatory | path-heuristic | Art. 26/27 deployer obligations |
| `ms-ai-governance/references/responsible-ai/ai-act-fria-template.md` | template | path-heuristic | FRIA fill-in scaffold (7 sections) grounded in Art. 27 |
| `ms-ai-governance/references/responsible-ai/ai-act-microsoft-tools-mapping.md` | reference | path-heuristic | MS tool capabilities mapped to AI Act articles (Purview/Compliance Manager facts) |
| `ms-ai-governance/references/responsible-ai/ai-act-provider-obligations.md` | regulatory | path-heuristic | Art. 9–15 provider obligations |
| `ms-ai-governance/references/responsible-ai/ai-act-transparency-notices.md` | regulatory | path-heuristic | Art. 13/50 transparency obligations (maler secondary) |
| `ms-ai-governance/references/responsible-ai/ai-ethics-in-public-sector.md` | reference | ms-citation+path-conflict | MS RAI standard + Foundry capability facts are the substance |
| `ms-ai-governance/references/responsible-ai/ai-governance-structure-framework.md` | methodology | ms-citation+path-conflict | organizational governance framework guidance |
| `ms-ai-governance/references/responsible-ai/ai-impact-assessment-framework.md` | methodology | ms-citation+path-conflict | impact assessment process framework |
| `ms-ai-governance/references/responsible-ai/gdpr-compliance-ai-systems.md` | regulatory | ms-citation+path-conflict | GDPR obligations applied to AI systems |
| `ms-ai-governance/references/responsible-ai/model-explainability-interpretability.md` | reference | ms-citation+path-conflict | Azure ML RAI dashboard/InterpretML product facts dominate |
| `ms-ai-governance/references/responsible-ai/responsible-ai-framework-overview.md` | reference | ms-citation+path-conflict | Microsoft RAI framework facts, MS-verifiable |
| `ms-ai-infrastructure/references/hybrid-edge/azure-arc-ai-management.md` | reference | path-heuristic | Azure Arc product facts |
| `ms-ai-infrastructure/references/hybrid-edge/azure-confidential-computing-ai.md` | reference | path-heuristic | Azure Confidential Computing product facts |
| `ms-ai-infrastructure/references/hybrid-edge/azure-iot-hub-ai-pipeline.md` | reference | path-heuristic | Azure IoT Hub product facts |
| `ms-ai-infrastructure/references/hybrid-edge/azure-local-ai-workloads.md` | reference | path-heuristic | Azure Local product facts |
| `ms-ai-infrastructure/references/hybrid-edge/data-sovereignty-norway-public-sector.md` | reference | path-heuristic | Azure data residency/EUDB/sovereign cloud facts |
| `ms-ai-infrastructure/references/hybrid-edge/edge-ai-inferencing-patterns.md` | reference | path-heuristic | edge inference product patterns (IoT Edge, quantization) |
| `ms-ai-infrastructure/references/hybrid-edge/edge-to-cloud-data-synchronization.md` | reference | path-heuristic | edge-cloud sync product patterns |
| `ms-ai-infrastructure/references/hybrid-edge/hybrid-rag-architecture.md` | reference | path-heuristic | hybrid RAG architecture product patterns |
| `ms-ai-infrastructure/references/hybrid-edge/iot-operations-ai-integration.md` | reference | path-heuristic | Azure IoT Operations product facts |
| `ms-ai-infrastructure/references/hybrid-edge/kubernetes-edge-aks-edge.md` | reference | path-heuristic | AKS Edge Essentials product facts |
| `ms-ai-infrastructure/references/hybrid-edge/network-constrained-ai-deployment.md` | reference | path-heuristic | constrained-network deployment product patterns |
| `ms-ai-infrastructure/references/hybrid-edge/offline-first-ai-applications.md` | reference | path-heuristic | offline-first AI application product patterns |
| `ms-ai-infrastructure/references/hybrid-edge/on-premises-slm-phi-deployment.md` | reference | path-heuristic | Phi-3/Phi-4 on-prem deployment facts |
| `ms-ai-infrastructure/references/hybrid-edge/onnx-runtime-edge-deployment.md` | reference | path-heuristic | ONNX Runtime product facts |
| `ms-ai-infrastructure/references/hybrid-edge/regulatory-compliance-edge-ai.md` | reference | path-heuristic | MS compliance tooling for edge (Arc/Purview/Defender/Compliance Manager) is the substance |
| `ms-ai-infrastructure/references/hybrid-edge/sovereign-cloud-norway.md` | reference | path-heuristic | MS sovereign cloud capability facts |
| `ms-ai-infrastructure/references/hybrid-edge/windows-ai-apc-capabilities.md` | reference | path-heuristic | Windows ML/NPU/Copilot+ PC facts |
| `ms-ai-security/references/ai-security-engineering/ai-security-scoring-framework.md` | reference | ms-citation+path-conflict | documents Microsoft’s AI Risk Assessment Framework (MS-verifiable) |
| `ms-ai-security/references/ai-security-engineering/secure-model-deployment-hardening.md` | reference | path-heuristic | Azure ML/ACR/Defender/Key Vault hardening facts |
| `ms-ai-security/references/performance-scalability/auto-scaling-ai-infrastructure.md` | reference | path-heuristic | Azure Container Apps scaling facts |
| `ms-ai-security/references/performance-scalability/cdn-edge-caching-ai.md` | reference | path-heuristic | Azure Front Door/CDN product facts |
| `ms-ai-security/references/performance-scalability/latency-optimization-azure-openai.md` | reference | path-heuristic | Azure OpenAI latency optimization facts |
| `ms-ai-security/references/performance-scalability/performance-benchmarking-frameworks.md` | reference | ms-citation+path-conflict | azure-openai-benchmark/Azure Load Testing/Foundry eval facts |
| `ms-ai-security/references/performance-scalability/streaming-response-patterns.md` | reference | path-heuristic | SSE/streaming implementation facts for Azure OpenAI |