MicrosoftKnown publishermicrosoft/physical-ai-toolchain/.github/skills/infrastructure/SKILL.md

infrastructure

Skill138 starsChanged 14 days agoMITView on GitHub

Deploy and manage Azure infrastructure for the Physical AI Toolchain including Terraform IaC, Kubernetes setup, GPU configuration, and network topology

12 other projects use a version of this.

Adopted by

Most recently taken up by

BermanRoman/promt-seo-geotheir own version, first seen Nov 8, 2026
xiaoyangtx996/Searching-skilltheir own version, first seen Nov 8, 2026
theopenco/llmgatewaytheir own version, first seen Nov 6, 2026
grafana/skillstheir own version, first seen Oct 27, 2026
edjchapman/claude-code-configtheir own version, first seen Oct 23, 2026
rcelebrone/DotAgentstheir own version, first seen Oct 14, 2026

Other versions12 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 12
theopenco/llmgateway★ 1.7ksame name
grafana/skills★ 302same name
rcelebrone/DotAgents★ 15same name
edjchapman/claude-code-config★ 8same name
Kur1sulab/blackbox★ 3same name
ieuanign/skills★ 2same name
cwklurks/ship-score★ 1same name
manastalukdar/ai-devstudio★ 1same name
mgoericke/claude-skills-devcontainer★ 1same name
xiaoyangtx996/Searching-skill★ 1same name
2 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository10 other files it gives its agents.
- .github/copilot-instructions.mdCopilot instructions
- azureml-k3s-compute-target-setupSkill
- dataviewerSkill
- e2e-infrastructureSkill
- e2e-testsSkill
- environment-deploymentSkill
- fleet-deploymentSkill
- fleet-intelligenceSkill
- osmo-lerobot-trainingSkill
- synthetic-dataSkill

The file

We split this skill into the rules it gives an agent

1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.

0
Common
Most of the 8 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Follow these steps in order for a complete deployment.
line 22

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

SKILL.md

Infrastructure Skill

Deploy and manage Azure cloud infrastructure for the Physical AI Toolchain — Terraform IaC, AKS cluster configuration, GPU node pools, and network topology.

Prerequisites

Tool	Requirement
Azure CLI	`az login` authenticated
Terraform	1.5+
kubectl	Matching cluster version
Helm	4.2+
shellcheck	For script validation

Deployment Workflow

Follow these steps in order for a complete deployment.

Step 1 — Initialize Azure subscription

source infrastructure/terraform/prerequisites/az-sub-init.sh

Exports ARM_SUBSCRIPTION_ID and validates Azure CLI authentication.

Step 2 — Configure Terraform variables

cd infrastructure/terraform
cp terraform.tfvars.example terraform.tfvars

Edit terraform.tfvars with environment-specific values. Example configurations are in infrastructure/examples/:

File	Scenario
`terraform.tfvars.dev`	Single spot GPU pool, public networking
`terraform.tfvars.prod`	Multiple GPU pools, full private networking, HA
`terraform.tfvars.hybrid`	Private data services, public AKS API server

Step 3 — Provision infrastructure

terraform init
terraform plan -var-file=terraform.tfvars
terraform apply -var-file=terraform.tfvars

Step 4 — Deploy VPN (private clusters only)

Required when should_enable_private_aks_cluster = true:

cd infrastructure/terraform/vpn
terraform init && terraform apply

Step 5 — Connect to cluster

az aks get-credentials --resource-group <rg> --name <aks>
kubectl cluster-info

Step 6 — Run setup scripts

cd infrastructure/setup
./01-deploy-robotics-charts.sh
./02-deploy-azureml-extension.sh
./03-deploy-osmo.sh --private-service-ip <unused-aks-subnet-ip>

Scripts must run in numeric order. Each supports --config-preview for dry-run output. On a new cluster, 03-deploy-osmo.sh needs --private-service-ip with a free address in the AKS subnet for the internal load balancer in front of OSMO; later runs reuse that address.

Network Mode Selection

Three network modes control connectivity and security:

Mode	`should_enable_private_endpoint`	`should_enable_private_aks_cluster`	VPN Required
Full Private	`true`	`true`	Yes
Hybrid	`true`	`false`	No
Full Public	`false`	`false`	No

Full Private is the default and recommended for production. Hybrid mode allows kubectl access without VPN while keeping data services private.

Common Operations

Plan changes

cd infrastructure/terraform
terraform plan -var-file=terraform.tfvars

Apply changes

terraform apply -var-file=terraform.tfvars

Destroy infrastructure

terraform destroy -var-file=terraform.tfvars

VPN setup

cd infrastructure/terraform/vpn
terraform init && terraform apply

DNS configuration

cd infrastructure/terraform/dns
terraform init && terraform apply

Validate setup scripts

shellcheck infrastructure/setup/01-deploy-robotics-charts.sh
infrastructure/setup/01-deploy-robotics-charts.sh --config-preview

Check Terraform formatting

terraform fmt -check -recursive infrastructure/terraform/

Directory Structure

infrastructure/
├── terraform/                         # Infrastructure as Code
│   ├── main.tf                        # Module composition
│   ├── variables.tf                   # Input variables
│   ├── outputs.tf                     # Output values
│   ├── versions.tf                    # Provider requirements
│   ├── terraform.tfvars.example       # Example configuration
│   ├── prerequisites/                 # Azure subscription setup
│   ├── modules/                       # Terraform modules
│   ├── vpn/                           # Standalone VPN deployment
│   ├── automation/                    # Standalone automation deployment
│   └── dns/                           # Standalone DNS deployment
├── setup/                             # Post-deploy cluster configuration
│   ├── 01-deploy-robotics-charts.sh   # GPU Operator, KAI Scheduler
│   ├── 02-deploy-azureml-extension.sh # AzureML K8s extension
│   ├── 03-deploy-osmo.sh             # OSMO control plane and backend
│   ├── defaults.conf                  # Central version and namespace config
│   └── lib/                           # Shared shell libraries
├── specifications/                    # Domain specification documents
└── examples/                          # Example tfvars configurations

GPU Configuration Reference

GPU	VM SKU	Driver Source	`gpu_driver`	MIG Strategy
A10	`Standard_NV36ads_A10_v5`	AKS-managed	`Install`	N/A
RTX PRO 6000	`Standard_NC144ds_xl_RTXPRO6000BSE_v6` (1 GPU, 96 GB)	AKS-managed GRID driver	`Install`	`single`
H100	`Standard_NC40ads_H100_v5`	GPU Operator	`None`	Disabled

Only RTX PRO 6000 pools created with gpu_driver = "None" need the nvidia.com/gpu.deploy.driver=false label, which hands them to the fallback GRID driver DaemonSet. Preview RTX sizes (128, 256, or 320 vCPUs) no longer deploy. Each NC144ds_xl node needs 144 vCPUs of RTX PRO 6000 quota, plus one more node's worth for an upgrade surge; park a pool with autoscaling off and node_count = 0 until quota exists.

Documentation

Guide	Description
Infrastructure README	Domain overview and quick start
Terraform README	Terraform configuration reference
Setup README	Setup script reference
Infrastructure Deployment	Full deployment walkthrough
GPU Configuration	Detailed GPU driver and operator reference

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free