frontmatter-description

作者: microsoft

检查并优化VS Code文档中的MetaDescription前置元数据字段。用于审计、添加或改进页面描述以提升SEO及…

npx skills add https://github.com/microsoft/vscode-docs --skill frontmatter-description

Frontmatter Description

Check and optimize MetaDescription frontmatter fields in markdown documentation files. Produces actionable fixes for descriptions that violate the rules below.

When to Use

  • Adding or editing a documentation page and need to write a MetaDescription.
  • Auditing existing pages for SEO or discoverability improvements.
  • Asked to check, review, or optimize frontmatter descriptions.
  • Editing content articles and want to ensure the MetaDescription reflects the updated content.

Rules

Every MetaDescription value must satisfy all of the following:

RuleDetail
LengthMaximum 160 characters.
ToneAction-oriented, value-focused, factual, and impersonal.
VoiceNo "you can", "users can", or "this page explains".
SentencesComplete sentences, not fragments or labels.
No colonsDo not include : in the value — it breaks YAML parsing.
UniquenessEach description must be unique across the docs.
Reusable variablesPreserve existing {% data variables.<group>.<name> %} references. Do not replace them with rendered product names.

Descriptions must also reflect the page's primary persona, reader intent, and article purpose. Establish that framing by following the content-framing guidance before suggesting or applying a description. If the existing page and writer's request support multiple materially different interpretations, ask the writer to confirm the framing before editing.

Procedure

1. Identify target files

Determine which files to check:

  • If given specific files, use those.
  • If asked to audit a folder (e.g., docs/copilot/), find all .md files in that folder recursively.
  • Skip files that have no YAML frontmatter (e.g., README.md files without --- delimiters).

2. Extract and validate

For each file, read enough of the page to identify its primary persona, reader intent, and article purpose. Then read the YAML frontmatter and check the MetaDescription field:

  1. Missing — Flag if MetaDescription is absent.
  2. Length — Flag if the value exceeds 160 characters. Report the current length.
  3. Voice — Flag if it contains "you can", "users can", "this page explains", or similar reader-addressing phrases.
  4. Tone — Flag if the description is a fragment, a label, or passive rather than action-oriented.
  5. Context — Flag if the description does not mention the relevant feature or tool.
  6. Content alignment — Flag if the description does not represent the page's purpose or the outcome sought by its primary persona.
  7. Forbidden words — Flag any occurrence of "teaching", "enable", "disable", or condescending terms.
  8. Colons — Flag any : character in the value.
  9. Versions — Flag any version numbers (e.g., "v1.90", "VS Code 1.90").
  10. Formatting — Flag Jinja2 variables ({{...}}), HTML tags, or Markdown formatting. Do not flag supported reusable data variables ({% data variables.<group>.<name> %}).
  11. Reusable variables — When a description contains a reusable data variable, verify that its full path exists in data/variables and resolves to the intended text. Preserve valid variables exactly as written.

3. Report findings

Present results as a table:

| File | Issue | Current value | Suggested fix |
|------|-------|---------------|---------------|
  • Group by file path.
  • For each issue, provide the current MetaDescription and a rewritten suggestion that passes all rules.
  • If a description passes all checks, omit it from the table (or mark it as passing if the user asked for a full report).

4. Apply fixes (when asked)

  • Edit the MetaDescription value in the YAML frontmatter.
  • Do not change any other frontmatter fields.
  • Do not alter the page body content.
  • Preserve existing reusable data variables. Never replace a variable with its rendered product name.
  • Verify the fix passes all rules and reflects the confirmed page framing before applying.

Examples

Good descriptions:

  • Get AI-powered inline suggestions from GitHub Copilot in VS Code, including ghost text completions and next edit suggestions.
  • Configure and manage extensions in Visual Studio Code to customize your development environment.
  • Debug Python applications in Visual Studio Code with breakpoints, variable inspection, and integrated terminal output.
  • Build a web app with an AI agent in {% data variables.product.prodname_vscode_shortname %}, then review and validate the result.

Bad descriptions and why:

DescriptionProblem
This page explains how to use Copilot.Uses "this page explains".
You can debug your code with VS Code.Uses "you can".
CopilotFragment, not a sentence.
Learn how to enable the new feature in VS Code 1.90.Uses "enable" and includes a version number.
Set up debugging: configure launch.json for Python.Contains a colon.

来自 microsoft 的更多技能

oss-growth
microsoft
OSS增长黑客角色
agent-framework-azure-ai-py
microsoft
使用Microsoft Agent Framework Python SDK(agent-framework-azure-ai)构建Azure AI Foundry代理。在创建使用AzureAIAgentsProvider的持久化代理、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时使用。涵盖函数工具、结构化输出和多工具代理。
development
airunway-aks-setup
microsoft
在AKS上设置AI Runway——从裸集群到运行模型。涵盖集群验证、控制器安装、GPU评估、提供商设置和首次部署。适用场景:“设置AI Runway”、“接入AKS集群”、“安装AI Runway”、“airunway设置”、“将模型部署到AKS”、“在AKS上进行GPU推理”、“在AKS上配置KAITO”、“在AKS上运行LLM”、“在AKS上使用vLLM”、“在AKS上设置模型服务”、“AI Runway控制器”。
devops
appinsights-instrumentation
microsoft
使用Azure Application Insights对Web应用进行插桩的指南。提供遥测模式、SDK设置和配置参考。适用场景:如何对应用进行插桩、App Insights SDK、遥测模式、什么是App Insights、Application Insights指南、插桩示例、APM最佳实践。
devops
applicationinsights-web-ts
microsoft
使用Application Insights JavaScript SDK(@microsoft/applicationinsights-web)为浏览器/Web应用添加检测。用于真实用户监控(RUM)——页面视图、点击、AJAX/fetch依赖项、异常、自定义事件,以及与后端OpenTelemetry追踪关联的浏览器端GenAI代理追踪。涵盖SDK加载器脚本和npm设置、框架扩展(React、React Native、Angular)、点击分析、遥测初始化器,以及从浏览器发出的代理/工具/模型跨度所遵循的OTel GenAI语义约定。
devops
azure-ai-anomalydetector-java
microsoft
使用适用于 Java 的 Azure AI 异常检测器 SDK 构建异常检测应用程序。在实现单变量/多变量异常检测、时间序列分析或 AI 驱动的监控时使用。
development
azure-ai-language-conversations-py
microsoft
使用azure-ai-language-conversations Python SDK实现对话语言理解(CLU)。当使用ConversationAnalysisClient分析对话意图和实体、构建NLP功能或将语言理解集成到应用程序中时使用。
development
azure-ai-ml-py
microsoft
Azure Machine Learning SDK v2 for Python。用于机器学习工作区、作业、模型、数据集、计算资源和管道。 触发词:“azure-ai-ml”、“MLClient”、“工作区”、“模型注册表”、“训练作业”、“数据集”。
development