vm-case-triage

Triage Azure VM and Storage support cases into structured troubleshooting directions and routing recommendations.

Updated Aug 13, 2026
One-click install
npx skills add https://github.com/crystalzhangbai/TS-Agent-Azure --skill vm-case-triage-crystalzhangbai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: vm-case-triage
Source: https://github.com/crystalzhangbai/TS-Agent-Azure/tree/main/bridge/skills/vm-case-triage
Command: npx skills add https://github.com/crystalzhangbai/TS-Agent-Azure --skill vm-case-triage-crystalzhangbai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve? When a new Azure VM or Storage support case lands, engineers often don't know where to start or whether the case even belongs to their team. This Skill turns a pasted case body or free-form symptom into a structured troubleshooting direction and settles ownership questions with concrete Support Area Path recommendations. ## Core Features & Use Cases - Six-section triage cockpit: Classifies the case (fault, performance, advisory, RCA, how-to), then produces interpretation, scope questions, ranked root-cause hypotheses with confidence levels, step-by-step troubleshooting, cited TSG/MS Learn references, and next-skill recommendations. - Scope and route decisions: Determines whether a case is in VM/Storage scope, resolves borderline ownership disputes (SQL-on-VM, Bastion, Azure Files vs managed disk), and recommends the correct Support Area Path plus owning team from a 53k+ path catalog. - Autopilot mode: Optionally chains up to three read-only diagnostic steps (Kusto queries, guest log analysis) automatically, with guardrails that never touch customer subscriptions or change case state. - Use Case: An engineer receives a case saying a production Linux VM rebooted at 03:15 UTC with no maintenance window. The Skill classifies it as a fault case, lists Service Healing and kernel panic hypotheses with confirmation steps, cites the reboot-attribution TSG, and recommends a platform Kusto query as the next move. ## Quick Start Paste the case body or describe the symptom and ask where to start troubleshooting this Azure VM case.

Frequently Asked Questions about vm-case-triage

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I triage a new Azure VM support case?▼

Paste the case body or describe the symptom, and the Skill classifies the case type, then emits a six-section cockpit: interpretation, scope questions, ranked root-cause hypotheses with confidence levels, troubleshooting steps, reference links, and recommended next skills.

How do I decide which team owns an Azure support case?▼

Use the Scope and Route stage, which applies a scope decision tree and boundary rules to determine if the case belongs to VM/Storage. For out-of-scope cases, it searches a 53k+ path SAP catalog to recommend the correct Support Area Path and owning team.

What is a Support Area Path and how do I find the right one?▼

A Support Area Path (SAP) is Microsoft's case routing taxonomy identifying the owning product team. The Skill searches a local JSON tree of over 53,000 paths by keyword, cross-references scope catalogs, and returns the full path with the owning team name.

Does the triage skill fetch case data from DFM automatically?▼

No, it never fetches from DFM or drives browser automation. All input comes from text the user pastes, such as a DFM Q&A block, advisory notification, or free-form symptom description; the user changes the SAP in DFM manually.

What are the autopilot safety limits for automated case diagnosis?▼

Autopilot chains at most three read-only diagnostic steps such as Kusto queries or log analysis. It never logs into customer subscriptions, restarts VMs, sends customer emails, or changes Support Area Paths; those actions stop the loop and require explicit user confirmation.

When should I not use case triage and go directly to another tool?▼

Skip triage when the next action is already known: draft FQR or RCA replies manually, use vm-kusto-query for a scoped platform query, or vm-log-analyzer for a specific log file. Triage is the entry point only when direction or ownership is unclear.