// THE PROBLEM
Stateless agents.
Every session starts from zero. Users re-explain context. Token costs spiral. The agent never gets smarter at its job.
Security veto.
Private data access plus untrusted content plus external API calls plus persistent memory equals the exact attack surface security teams are now blocking. No security signoff equals no production deployment.
Vendor lock-in.
The wrong agent runtime, memory layer, or model provider today is the migration tax tomorrow. Architecture decisions compound.
// THE STACK
Every action crosses four security perimeters before it touches your network. Inside the runtime, work splits across nine specialist roles, one per body part. Same architecture, Azure-native controls.
Private Azure Network / VNet
OpenShell / NemoClaw Sandbox
Agent Runtime
L0
Brain
Azure AI Foundry, Nemotron local
L1
Hands
MCP Servers, Azure Functions, Logic Apps
L2
Heart
OpenClaw or JARVIS Core
L3
Session
Cosmos DB, Postgres, Redis Cache
L4
Badge
Managed Identity, Entra ID, Key Vault
L5
Mouth
Teams, Web App, API Management
L6
Library
AI Search, pgvector, Blob, Hindsight
L7
Manager
AKS, Service Bus, Container Apps
L8
Receipt
App Insights, Monitor, Langfuse, Sentinel
// SPECIALIST ROLES
| Layer | Body Part | What It Does | Why It Matters |
|---|---|---|---|
| L0 | Brain | Reasoning and inference. Routes by data sensitivity. | The decision-maker. Smart enough to know when to ask a smaller, local model instead of a frontier one. |
| L1 | Hands | Tool use, MCP servers, external APIs, functions. | Where the agent does things. Books meetings, files forms, sends emails, runs code. |
| L2 | Heart | Agent loop. ReAct cycles, planning, tool selection. | The pulse. Reasons, picks a tool, observes the result, repeats until the task completes. |
| L3 | Session | Short-term context, working memory. | The current conversation. State that persists during the session and gets archived at session end. |
| L4 | Badge | Agent identity, RBAC, scopes, capability policies. | The agent ID card. Defines what it can touch, which secrets it can use, which client tenant it serves. |
| L5 | Mouth | User-facing surfaces. Trust-boundary aware sessions. | How the agent talks to humans and other systems. Knows the difference between main session, DM, and group chat. |
| L6 | Library | Long-term memory with entity resolution. | The accumulated knowledge that turns a stateless agent into a worker who learns the job over time. |
| L7 | Manager | Multi-agent orchestration, queues, scheduled jobs. | The conductor. Coordinates specialists across parallel workflows so nothing falls through the cracks. |
| L8 | Receipt | Per-agent observability and audit trail. | Every prompt, response, tool call, and decision logged. Audit-ready, replayable, exportable to your SIEM. |
// COMPONENT CHOICES
Each layer has Azure-native defaults and portable alternates. Every layer is swappable.
Agent Runtime layer
- Azure option: Azure AI Foundry agents
- Open option: OpenClaw
- TAG AI default: JARVIS Core on Azure Container Apps or AKS
Specialist agents work as a coordinated team while identity, routing, and deployment stay aligned to Azure.
Sandbox and Policy layer
- Azure option: Defender for Cloud plus Azure Policy
- Network option: Private Link, NSG, Azure Firewall
- TAG AI default: NemoClaw pattern plus OpenShell
Compromised prompts cannot escape the sandbox. Egress to unapproved domains gets blocked before it reaches the public internet.
Memory Engine layer
- Azure option: AI Search, Cosmos DB, Azure Database for PostgreSQL
- Open option: Hindsight or pgvector
- TAG AI default: Hindsight plus AI Search plus Postgres
Long-term memory stays inside your tenant, with retrieval logs and data residency controls your security team already understands.
Observability layer
- Azure option: Monitor, Application Insights, Log Analytics
- Security option: Microsoft Sentinel
- TAG AI default: Azure telemetry plus Langfuse
Every prompt, response, and tool call is traceable, replayable, and exportable into the same operations console your team already watches.
Model and Infrastructure layer
- Frontier option: Azure AI Foundry and Azure OpenAI
- Local and regulated: Nemotron, Ollama, vLLM on AKS
- TAG AI default: Hybrid sensitivity routing
Sensitive data stays in your network. Best-in-class model performance is used only where policy allows it.
// HOW IT FLOWS
For your security review, here is exactly what crosses each boundary, every Azure service touched, every log generated.
| 01 | User sends message | Mouth, L5 | Auth event in Entra ID |
| 02 | Identity and scope validated | Badge, L4 | RBAC decision logged |
| 03 | Sandbox boundary enforced | OpenShell | Network policy decisions, capability check |
| 04 | Long-term context retrieved | Library, L6 | AI Search and database query trace logged |
| 05 | Sensitivity classified | Brain, L0 | Model routing decision logged |
| 06 | Agent loop reasons and acts | Heart plus Hands | ReAct trace, function calls, and Logic Apps runs captured |
| 07 | Response generated | Receipt, L8 | Prompt, response, token cost, and App Insights event logged |
| 08 | Memory updated for next session | Library, L6 | Memory write event logged |
// DEFENSIBILITY
Compliance
Every action crosses four trust boundaries. Every decision lands in Sentinel. Audit responses go from days to seconds.
No lock-in
Every layer is swappable. New frontier model? Update the Brain. Better memory framework? Swap the Library. The body metaphor is the abstraction that lets each part evolve.
Production tested
We run this stack on our own business: E-Rate consulting, real estate operations, sales pipelines. We deploy what we depend on.
Microsoft aligned
If you are already an Azure shop, every layer plugs into existing identity, billing, support, and compliance contracts. No second cloud invoice. No second security review.
Architecture reviewed. Stack validated. Ship it.
If you are evaluating AI agent deployments and your security team has questions you cannot answer yet, that is the conversation we are built for. Operator grade architecture. Production ready in weeks, not quarters.
Book an architecture review