reviewing-aspire-architecture

作者: microsoft

觸發針對15個Aspire特定維度的深度架構審查。由深度審查、架構審查、模式審查或PR請求啟動…

npx skills add https://github.com/microsoft/aspire --skill reviewing-aspire-architecture

Aspire Architectural Review

Aspire-specific architectural review via the reviewing-aspire-architecture agent. Catches domain patterns that generic code-review cannot.

When to Use

Use only when at least one of these conditions is true:

  • The user explicitly asks for a deep architectural review, Aspire pattern review, or equivalent domain-focused review of an existing PR or diff.
  • A generic review has completed and found a concrete question that cannot be resolved without Aspire-specific architectural knowledge. The reviewer must state the relevant diff evidence, unresolved question, applicable Aspire contract, and why its normal review cannot determine correctness; invoke this skill with only that focused question. Do not escalate merely for additional confidence or a second opinion.

Changed file location is routing information after this skill is selected, not a reason to select it. A PR touching hosting core, Azure integrations, dashboard, CLI, components, resource types, the app model, or deployment behavior does not by itself qualify.

Outside the post-review escalation path, the user's primary requested action must be architectural review. Asking to understand, explain, compare, or design something does not qualify, even when the request references a PR or diff.

Don't use this skill for implementation, debugging, testing, routine code changes, post-change validation, design questions or brainstorming, explanatory investigation, generic PR review, CI failures, flaky tests, or API surface review. Use code-review, ci-test-failures, fix-flaky-test, or api-review where appropriate.

For ordinary review routing, use code-review for doc/config-only PRs. This does not prevent a direct invocation when the user explicitly requests a deep architectural or pattern review of such a diff.

Invocation Guard

  • A parent may load this skill once and launch one reviewing-aspire-architecture agent for a user review request and change-set revision.
  • If this skill context is already loaded, do not invoke the skill again.
  • If the current agent is reviewing-aspire-architecture, do not invoke this skill or launch another reviewing-aspire-architecture agent.
  • Continue follow-up work in the existing agent conversation. Do not launch a replacement agent for refinements or narrower follow-up questions.
  • Completing or modifying code does not automatically trigger another architectural review. Re-run only when the user explicitly requests it for a materially changed diff.

Relationship to code-review

code-review covers generic bugs, security, perf, concurrency, and error handling and is sufficient for ordinary PR reviews. This skill covers only checks requiring Aspire domain knowledge. Some topic areas (Security, Performance, Error Handling) exist in both, but the checks are disjoint: generic patterns belong to code-review, Aspire-specific patterns belong here.

When the user explicitly requests a deep architectural or pattern review, invoke this skill directly; a preliminary code-review pass is not required.

For an ordinary review, always complete code-review before considering an organic escalation to this skill. Do not front-load the architectural review or use it to decide what the generic reviewer should inspect. After the generic findings are complete, escalate when correctness depends on a named Aspire-domain contract that the generic review cannot resolve from its own rules and evidence.

If code-review can already state a concrete, high-confidence defect, report it directly rather than escalating it. If escalation is justified, preserve the completed generic findings, run one focused architectural analysis, and merge only its net-new high-confidence findings. Do not rerun code-review afterward.

Review Scope

For an explicit full architectural review, inspect the PR context and only the dimensions selected by the changed-file routing below. For an escalation from another review, inspect only the supplied question, relevant diff, surrounding code, and existing comments needed to avoid duplication. Do not expand a focused escalation into a full PR review or repository-wide inventory.

Folder → Dimension Routing

FolderDimensions
src/Aspire.Hosting/**Resource Model, API Design, Pattern Conformance, Containers
src/Aspire.Hosting.Azure*/**Azure Provisioning, Resource Model, API Design, Security
src/Aspire.Dashboard/**Dashboard UI/UX, Security, Performance
src/Aspire.Cli/**CLI Behavior, Error Handling, Platform Compatibility
src/Components/**Pattern Conformance, API Design, Build & Contributor Workflow
tests/**Test Quality + mirror dimensions of code under test
eng/**, .github/**Build & Contributor Workflow, Documentation & Naming

Full review rules in the reviewing-aspire-architecture agent.

來自 microsoft 的更多技能

oss-growth
microsoft
開源增長駭客角色
agent-framework-azure-ai-py
microsoft
使用Microsoft Agent Framework Python SDK(agent-framework-azure-ai)构建Azure AI Foundry代理。适用于使用AzureAIAgentsProvider创建持久化代理、使用托管工具(代码解释器、文件搜索、网络搜索)、集成MCP服务器、管理对话线程或实现流式响应。涵盖函数工具、结构化输出和多工具代理。
development
airunway-aks-setup
microsoft
在AKS上設定AI Runway——從裸叢集到執行模型。涵蓋叢集驗證、控制器安裝、GPU評估、供應商設定及首次部署。時機:「設定AI Runway」、「上線AKS叢集」、「安裝AI Runway」、「airunway設定」、「部署模型至AKS」、「在AKS上進行GPU推論」、「在AKS上設定KAITO」、「在AKS上執行LLM」、「在AKS上使用vLLM」、「在AKS上設定模型服務」、「AI Runway控制器」。
devops
appinsights-instrumentation
microsoft
使用Azure Application Insights檢測Web應用程式的指南。提供遙測模式、SDK設定與組態參考。適用時機:如何檢測應用程式、App Insights SDK、遙測模式、什麼是App Insights、Application Insights指南、檢測範例、APM最佳實踐。
devops
applicationinsights-web-ts
microsoft
使用Application Insights JavaScript SDK(@microsoft/applicationinsights-web)為瀏覽器/Web應用程式進行檢測。適用於真實使用者監控(RUM)——頁面檢視、點擊、AJAX/fetch依賴、例外、自訂事件,以及與後端OpenTelemetry追蹤關聯的瀏覽器端GenAI代理追蹤。涵蓋SDK載入器指令碼與npm設定、框架擴充(React、React Native、Angular)、點擊分析、遙測初始化器,以及從瀏覽器發出的代理/工具/模型span的OTel GenAI語意慣例。
devops
azure-ai-anomalydetector-java
microsoft
使用適用於 Java 的 Azure AI 異常偵測器 SDK 建置異常偵測應用程式。在實作單變量/多變量異常偵測、時間序列分析或 AI 驅動監控時使用。
development
azure-ai-language-conversations-py
microsoft
使用 azure-ai-language-conversations Python SDK 實作對話語言理解(CLU)。當使用 ConversationAnalysisClient 分析對話意圖與實體、建置 NLP 功能,或將語言理解整合至應用程式時使用。
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python。用於機器學習工作區、作業、模型、資料集、計算資源與管線。 觸發詞:「azure-ai-ml」、「MLClient」、「workspace」、「model registry」、「training jobs」、「datasets」。
development