borghei/Claude-Skills / engineering/azure-cloud-architect
azure-cloud-architect
VerifiedDesign, review, and validate Azure cloud architectures. Use when choosing Azure compute, storage, networking, or identity services, or applying the Azure Well-Architected Framework to a workload.
Azure Cloud Architect
End-to-end Azure-specific architecture: service selection, Well-Architected Framework assessment, identity and networking patterns, cost optimization, and operational defaults. Provider-specific complement to our generic senior-cloud-architect skill — that one covers cross-cloud patterns; this one knows AKS pricing tiers, when to pick Cosmos over SQL DB, and how Front Door differs from Application Gateway.
Core Capabilities
- Compute selection — decision tree across AKS, App Service, Container Apps, Functions, VMs/VMSS, Batch, Static Web Apps, and API gateways.
- Data store selection — relational, NoSQL/document, cache, blob, time-series, search, vector, and warehouse decision trees.
- Networking design — VNets, Private/Service Endpoints, gateways (App Gateway, Front Door, Firewall, NAT, VPN/ER), hub-and-spoke and Private Link patterns.
- Identity design — Entra ID, Managed Identity (system/user-assigned), Workload Identity, Service Principals, and least-privilege RBAC scoping.
- WAF assessment — score workloads across Reliability, Security, Cost, Operational Excellence, and Performance Efficiency pillars.
- Cost optimization — right-sizing, reservations, autoscaling, spot, storage tiering, and egress-reduction levers with anti-pattern detection.
When to Use
| Situation | Skill applies |
|---|---|
| Designing an Azure architecture from scratch | Yes — start with the compute decision tree |
| Reviewing an existing Azure architecture | Yes — run WAF assessment via scripts/azure_waf_scorer.py |
| Validating a Bicep template or a YAML workload spec | Yes — scripts/azure_architecture_validator.py (ARM JSON and Terraform are not checked) |
| Estimating Azure cost for a workload | Yes — scripts/azure_cost_estimator.py |
| Picking compute, data store, networking, or identity | Yes — see the decision-trees reference |
| Going to production without WAF review | Don't — run the WAF scorer first |
Clarify First
Before designing or assessing, settle the inputs below. When one is missing or vague and would change the result, ask about it; when it wouldn't, use a sensible default and say which one you used.
- Task: design from scratch, review an existing architecture, validate IaC, or estimate cost (selects
azure_architecture_validator.pyvsazure_cost_estimator.pyvsazure_waf_scorer.py). - Workload spec: the YAML workload config, or the Bicep files (the input the scripts parse). ARM JSON templates and Terraform are not checked; for Terraform, describe the workload in the YAML spec instead.
- Priority pillar: reliability, security, cost, operational excellence, or performance (weights the WAF assessment and which recommendations lead).
Ask about the two or three inputs that matter most for this request, not the whole list. If the user wants a result straight away, produce it and state your assumptions at the top.
Quick Start Workflow
All three scripts read one YAML workload spec. Its fields are listed in
references/workload-spec.md; start from assets/sample_workload.yaml.
# 1. Score the five Well-Architected pillars (10 checks each) and list failed checks by severity
python scripts/azure_waf_scorer.py --workload-config assets/sample_workload.yaml # 89/100
python scripts/azure_waf_scorer.py --workload-config assets/sample_workload_flawed.yaml # 6/100
# 2. Look for anti-patterns in the spec and in Bicep (ARM JSON and Terraform are not read)
python scripts/azure_architecture_validator.py --workload-config assets/sample_workload.yaml # exit 0
python scripts/azure_architecture_validator.py --bicep assets/sample_flawed.bicep # exit 1, 4 critical
# 3. Rough monthly cost with optimisation hints
python scripts/azure_cost_estimator.py --workload-config assets/sample_workload.yaml # about $1,755/mo
Tools overview. Output flags differ by tool: the scorer and validator take --format markdown|json,
the estimator --format human|json; all take --output FILE. The validator's --severity hides
findings below a level.
Exit codes. Validator: 0 no critical finding, 1 at least one critical finding, 2 unreadable
spec, a spec that is not a YAML mapping, or a Bicep path that matches nothing. Scorer and estimator:
0 report written, 2 the same input errors (the estimator also when nothing in the spec can be
priced). The WAF score itself does not set the exit code; read it.
Done when:
- Every critical WAF check passes or has a written exception with an owner.
- The validator exits 0 on the spec and on every Bicep file you deploy.
- The cost estimate was checked against the Azure Pricing Calculator before anyone quotes it.
- Each
truein the spec is backed by evidence (a dashboard, a runbook, a restore-test date).
Anti-Patterns
Scoring the spec instead of the system
Mistake: Marking backup_tested: true or mfa_enforced: true because the team believes it, then reporting a 90/100 WAF score.
Why it happens: The spec is a checklist, and ticking boxes is faster than checking them.
Instead: Set a field to true only with evidence. The scorer never calls Azure, so its score is exactly as honest as the YAML.
Quoting the estimator as a budget
Mistake: Putting the estimator's total into a budget or a customer quote. Why it happens: It prints a precise-looking dollar figure. Instead: Use it to compare designs and spot the big lines. Its table holds approximate early-2026 list prices without reservations; price the final design in the Azure Pricing Calculator.
Public PaaS endpoints "for now"
Mistake: Shipping storage, SQL or Cosmos with public network access and shared keys, planning to lock it down later.
Why it happens: Private Endpoints and Managed Identity take extra setup on day one.
Instead: Start private with Managed Identity. The validator flags publicNetworkAccess: 'Enabled', allowBlobPublicAccess: true and allowSharedKeyAccess: true in Bicep.
Files
scripts/azure_waf_scorer.py,scripts/azure_architecture_validator.py,scripts/azure_cost_estimator.py— the three CLIs above.assets/sample_workload.yaml(sound tier-1),assets/sample_workload_flawed.yaml(gaps everywhere),assets/sample_flawed.bicep(five Bicep findings).references/workload-spec.md— every spec field and which tool reads it; other references listed below.
Scope & Limitations
In Scope:
- Azure service selection through decision trees for compute, data stores, networking and identity
- Well-Architected Framework scoring with
azure_waf_scorer.py: five pillars, ten checks each, read from a YAML workload spec - Anti-pattern checks with
azure_architecture_validator.pyon a YAML workload spec or Bicep files: public endpoints, shared keys instead of Managed Identity, single-AZ tier-1, wide RBAC scopes, missing diagnostics and locks, and credentials in parameters - Rough monthly cost estimates with
azure_cost_estimator.pyfrom a YAML workload spec, with prioritized optimization suggestions
Out of Scope:
- Cross-cloud patterns and the AWS and GCP equivalents (see
engineering/senior-cloud-architect/,engineering/aws-solution-architect/andengineering/gcp-cloud-architect/) - Writing or reviewing Terraform modules (see
engineering/terraform-patterns/) - Security posture audits of an exported cloud inventory and IAM (see
engineering/cloud-security/) - Kubernetes operator design for AKS workloads (see
engineering/kubernetes-operator/) - Compliance evidence and control mapping (see
ra-qm-team/soc2-compliance-expert/andra-qm-team/information-security-manager-iso27001/)
Important Caveats:
- The scripts never call Azure. They read files the user provides, and the WAF score reflects what the YAML spec declares, not what is deployed.
- The validator scans Bicep text with six pattern rules (public network access, anonymous blob access, shared keys, TLS below 1.2, HTTP allowed, short inline passwords); its rules do not match ARM JSON syntax, so ARM templates pass without findings. It does not compile templates or resolve modules and parameters, and it does not read Terraform.
- Cost figures come from a built-in table of approximate list prices from early 2026, with no reservation or savings-plan discounts. They are for relative comparison, not a quote; check the Azure Pricing Calculator before budgeting.
- Service tiers, SLAs, limits and prices in the references change often and should be confirmed against current Azure documentation.
Reference Documentation
Load the reference that matches the task — keep this file lean and pull detail on demand:
- references/workload-spec.md — the YAML fields the three scripts read. Read before writing a spec.
- references/azure-decision-trees.md — full compute, data store, networking, and identity decision trees with building-block, gateway, and RBAC tables. Read when selecting services or designing topology.
- references/azure-workflows-and-antipatterns.md — the five WAF pillars, cost levers and anti-patterns, the design/review/migrate workflows, Azure-specific anti-patterns, and tooling outputs. Read when running a design or review.
- references/azure-services-reference.md — per-service depth: tiers, SLAs, limits, when to upgrade. Read when choosing a specific SKU or tier.
- references/azure-well-architected.md — 5-pillar WAF assessment with the 10-question checklist per pillar, common findings, and remediations. Read during a WAF review.
- references/azure-cost-optimization.md — full cost lever catalog, anti-patterns, and detection heuristics. Read when optimizing spend.
Related skills
engineering/senior-cloud-architect— generic multi-cloud architecture patternsengineering/aws-solution-architect— AWS counterpartengineering/gcp-cloud-architect— GCP counterpartengineering/kubernetes-operator— for AKS operator-pattern workloadsra-qm-team/information-security-manager-iso27001— for compliance-mapped controls (Azure has built-in Defender / Compliance Manager)ra-qm-team/soc2-compliance-expert— Azure-specific SOC 2 evidence collection
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |