feature-inventory

作者: microsoft

從程式碼庫中編錄現有功能(API端點、使用者流程、UI畫面、可觀察行為),並生成帶有REQ-XXX…的結構化功能規格。

npx skills add https://github.com/microsoft/github-copilot-modernization --skill feature-inventory

User Input

You MUST consider the user input before proceeding (if not empty).

The text in the task description is the feature description. Assume you always have it available. Do not ask the user to repeat it unless they provided an empty command.

Outline

Given that feature description, do this:

Output

All outputs are written under your task's Artifact path: (from task metadata).

spec.md                     ← main artifact (feature specification)
requirements-checklist.md   ← quality checklist (if applicable)

Workflow

  1. Load constitution document (if available) to understand principles and constraints.

  2. Load clarification.md if it exists ({{BASE_PATH}}/clarification.md or path from dependencyArtifacts). Treat every field that has an explicit value (i.e., resolution: user or resolution: default) as already resolved — do NOT re-ask the user about it and do NOT count it against the [NEEDS CLARIFICATION] budget below. Items listed under blocking_gaps should be treated as known risks but still do not count against the budget (they are surfaced separately by the coordinator).

  3. Read knowledge-graph.json (and, if needed, the query guidance in skills/building-java-knowledge-graph/SKILL.md) to understand the architecture and dependencies related to this feature. Also check for research files if available.

  4. Load skills/writing-feature-spec/templates/spec-template.md for required sections.

  5. Execute:

    1. Parse feature description from task If empty: ERROR "No feature description provided"
    2. Extract key concepts: actors, actions, data, constraints
    3. Assign unique requirement IDs (CRITICAL for traceability):
      • Format: REQ-XXX where XXX is a 3-digit sequence number (001, 002, ...)
      • Example: REQ-001, REQ-002, REQ-003
      • Each requirement MUST have unique ID across entire feature
    4. For unclear aspects:
      • Make informed guesses based on context and industry standards
      • Items already answered in clarification.md are NOT unclear — use the user's answer (or the recorded default) and do not mark them.
      • Only mark with [NEEDS CLARIFICATION: specific question] if:
        • The choice significantly impacts feature scope or user experience
        • Multiple reasonable interpretations exist with different implications
        • No reasonable default exists
      • LIMIT: Maximum 3 [NEEDS CLARIFICATION] markers total
      • Prioritize clarifications by impact: scope > security/privacy > user experience > technical details
    5. Fill User Scenarios & Testing section If no clear user flow: ERROR "Cannot determine user scenarios"
    6. Generate Functional Requirements Each requirement must be testable Use reasonable defaults for unspecified details (document assumptions in Assumptions section)
    7. Define Success Criteria Create measurable, technology-agnostic outcomes Include both quantitative metrics (time, performance, volume) and qualitative measures (user satisfaction, task completion) Each criterion must be verifiable without implementation details
    8. Identify Key Entities (if data involved)
    9. Return: SUCCESS (spec ready for planning)
  6. Write the specification to the assigned artifact path using the template structure, replacing placeholders with concrete details derived from the feature description while preserving section order and headings.

  7. Report completion with spec file path, checklist results, and readiness for the next phase.

Research-Driven Mode

When pre-built research files are available (provided by a design/explore phase), use this alternate workflow instead of reading source code directly:

⚠️ NO Direct Codebase Exploration

When research files exist, do NOT read project source files directly (source code, build files, config files). All codebase analysis is already in the research files.

Research File Inputs

Look for these files in the current working directory or as referenced in the task:

  • project-structure.md — functional domain list (REQUIRED)
  • tech-stack.md — technology inventory (REQUIRED)
  • data-model.md — key entities summary (optional)
  • architecture-summary.md — substitute for direct knowledge graph reading (optional)
  • Any additional *.md research files (optional)

Domain-Driven Generation

Use the functional domain list from project-structure.md to drive:

  • User Scenarios: For each domain, generate 1–3 user stories (P1/P2/P3, Given-When-Then)
  • Requirements: For each domain, generate 2–4 REQ-XXX items referencing constitution principles
  • Key Entities: Use entities from data-model.md directly (top 6–8)

Incremental Section Writing

Write the spec section by section, appending each to the assigned artifact before moving to the next:

  1. Header + Scope Baseline → create/write artifact
  2. User Scenarios & Testing → append
  3. Requirements (Functional per domain + Non-Functional) → append
  4. Success Criteria → append

Checklist Generation

After spec is complete, append a quality checklist section at the end of the artifact:

Requirements Quality Checklist

Requirement ID Coverage

  • All requirements use REQ-XXX format
  • IDs are unique and sequential

Testability

  • Every requirement is independently testable
  • Acceptance criteria are concrete (Given-When-Then)

Completeness

  • Scope Baseline section complete
  • User Scenarios prioritized (P1, P2, P3)
  • Functional requirements cover all in-scope items
  • Success criteria are measurable

Constitution Alignment

  • All constitution principles referenced
  • Migration mode constraints respected

General Guidelines

  • Focus on WHAT users need and WHY.
  • Avoid HOW to implement (no tech stack, APIs, code structure).
  • Written for business stakeholders, not developers.
  • DO NOT create any checklists that are embedded in the spec. That will be a separate command.

Section Requirements

  • Mandatory sections: Must be completed for every feature
  • Optional sections: Include only when relevant to the feature
  • When a section doesn't apply, remove it entirely (don't leave as "N/A")

For AI Generation

When creating this spec from a user prompt:

  1. Make informed guesses: Use context, industry standards, and common patterns to fill gaps
  2. Document assumptions: Record reasonable defaults in the Assumptions section
  3. Limit clarifications: Maximum 3 [NEEDS CLARIFICATION] markers - use only for critical decisions that:
    • Significantly impact feature scope or user experience
    • Have multiple reasonable interpretations with different implications
    • Lack any reasonable default
  4. Prioritize clarifications: scope > security/privacy > user experience > technical details
  5. Think like a tester: Every vague requirement should fail the "testable and unambiguous" checklist item
  6. Common areas needing clarification (only if no reasonable default exists):
    • Feature scope and boundaries (include/exclude specific use cases)
    • User types and permissions (if multiple conflicting interpretations possible)
    • Security/compliance requirements (when legally/financially significant)

Examples of reasonable defaults (don't ask about these):

  • Data retention: Industry-standard practices for the domain
  • Performance targets: Standard web/mobile app expectations unless specified
  • Error handling: User-friendly messages with appropriate fallbacks
  • Authentication method: Standard session-based or OAuth2 for web apps
  • Integration patterns: RESTful APIs unless specified otherwise

Success Criteria Guidelines

Success criteria must be:

  1. Measurable: Include specific metrics (time, percentage, count, rate)
  2. Technology-agnostic: No mention of frameworks, languages, databases, or tools
  3. User-focused: Describe outcomes from user/business perspective, not system internals
  4. Verifiable: Can be tested/validated without knowing implementation details

Good examples:

  • "Users can complete checkout in under 3 minutes"
  • "System supports 10,000 concurrent users"
  • "95% of searches return results in under 1 second"
  • "Task completion rate improves by 40%"

Bad examples (implementation-focused):

  • "API response time is under 200ms" (too technical, use "Users see results instantly")
  • "Database can handle 1000 TPS" (implementation detail, use user-facing metric)
  • "React components render efficiently" (framework-specific)
  • "Redis cache hit rate above 80%" (technology-specific)

Resources

Templates

  • templates/spec-template.md — Specification document template with all mandatory sections

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