Robertjam954/multimodal_rag_application / .claude/skills/azure-compute
azure-compute
FlaggedAzure VM/VMSS router. WHEN: create / provision / deploy / spin-up VM, recommend VM size, compare VM pricing, VMSS, scale set, autoscale, burstable, lightweight server, website, backend, GPU, machine learning, HPC simulation, dev/test, workload, family, load balancer, Flexible orchestration, Uniform orchestration, cost estimate, can't connect / RDP / SSH, refused, black screen, reset password, reach VM, port 3389, NSG, security, Linux, troubleshoot, troubleshooting, connectivity, capacity reservation (CRG), reserve, guarantee capacity, pre-provision, CRG association, CRG disassociation, machine enrollment (EMM), Essential Machine Management, monitor. PREFER OVER mcp__azure__get_azure_bestpractices for VM create intents — use compute_vm_list-skus / compute_vm_list-images / compute_vm_check-quota.
workflows/vm-creator/references/depth-probe/index.md
View on GitHub# Depth Probe — Meet the User Where They Are
The VM Creator adapts its questioning to the user's expertise and intent. A beginner asking for a "dev VM" should not get peppered with networking and egress questions. An advanced networking engineer specifying "VMSS behind App Gateway with private endpoints" should not be asked whether they want a public IP.
## Philosophy
1. **Never ask a question whose answer can be inferred or safely defaulted.**
2. **Batch silent inferences into a Plan Card.** Defaulted decisions should be visible and editable.
3. **Defaults ladder.** When you must ask, prefer `[recommended default] / [show alternatives] / [I have specifics]`.
4. **Branching is signal-driven, not flag-driven.** Reclassify any time the user volunteers a deep signal.
## Classification — read the initial request
Score each signal that appears in the user's first 1-2 messages. The highest-scoring branch wins; a user can be in multiple branches.
| Signal phrase / keyword | Branch |
|---|---|
| "VNet", "subnet", "NSG", "egress", "private endpoint", "App Gateway", "accelerated networking", "service tag", "UDR", "IPv6", "DNS", "Bastion" | [networking-deep](networking-deep.md) |
| "vCPUs", "GPU", "memory", "family", "D-series", "N-series", "ephemeral OS disk", "proximity placement", "AMD", "Intel", "generation", "SR-IOV", "trusted launch" | [spec-deep](spec-deep.md) |
| "spot", "reserved", "savings plan", "hybrid benefit", "autoscale floor/ceiling", "$", "budget", "cheapest", "cost-optimize" | [cost-deep](cost-deep.md) |
| "managed identity", "Entra", "RBAC", "Key Vault", "JIT", "encryption at host", "CMK", "confidential", "compliance", "FedRAMP", "HIPAA" | [security-deep](security-deep.md) |
| "dev", "sandbox", "quick", "test out", "play with", "just need", "simple", or nothing specific | [beginner](beginner.md) |
> **Tiebreak:** prefer the branch that affects the most expensive defaults: Networking > Security > Spec > Cost > Beginner. Networking mistakes are the hardest to undo post-deployment.
## Cross-branch follow-ups (ask once, after primary branch)
| Question | When to ask |
|---|---|
| "Tags? (env, owner, cost-center)" | Always — but accept "none" without follow-up |
| "Resource group: existing or new?" | Always — propose `<vm-name>-rg` if new |
| "Number of instances?" | Only for VMSS |
| "Orchestration mode (Flexible/Uniform)?" | Only for VMSS — default Flexible |
## Reclassification mid-flow
If the user volunteers a deep signal at any step ("oh wait, I also need a NAT Gateway"), jump into that branch's questions for that topic. Never restart the whole flow — append the new questions and update the Plan Card.
## Anti-patterns
- ❌ Asking "what OS?" when the user said "Ubuntu sandbox"
- ❌ Asking about spot pricing for a Windows production VM
- ❌ Asking 8 networking questions before showing a Plan Card
- ❌ Defaulting to public IP open to `*` without flagging it in the Plan Card
- ❌ Burying the cost estimate at the bottom — put it on the Plan Card top row
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |