blog-writer

作者: microsoft

規劃、撰寫和審查 VS Code 部落格文章。當使用者要求「寫一篇部落格文章」、「起草一篇部落格」、「為……新增一篇部落格文章」、「審查……

npx skills add https://github.com/microsoft/vscode-docs --skill blog-writer

Write or Review a Blog Post

Help author a new VS Code blog post or review an existing one in the blogs/ folder. The skill has two modes:

  • Write mode works in two phases: it first researches the topic and proposes an outline, then drafts the post only after you approve.
  • Review mode evaluates an existing draft for clarity, conciseness, and style, and reports concrete suggestions without editing unless asked.

All work must follow the blog-writing style guide, which in turn builds on the docs-writing style guide.

When to Use

Use this skill whenever the request maps to any of these:

  • The user says "write a blog post", "draft a blog", "add a blog post for…", or "announce this on the blog".
  • The user pastes a GitHub issue or PR link and asks you to turn it into a blog post.
  • The user asks you to "review this blog post", check a draft, or improve an existing post in blogs/.
  • Announcing or explaining a new or changed VS Code or GitHub Copilot feature in a narrative, conversational format.

For new posts, run the plan-first workflow rather than jumping straight to a draft, even when the topic seems small.

Do not use this skill for reference documentation, release notes, or API docs. Those are handled by other skills (doc-writer, release-note-writer) or direct edits. If the user wants reference docs rather than a narrative post, switch to doc-writer.

Guardrails

  • Blogs only. Limit changes to the blogs/ folder. Do not update reference docs, release notes, or API docs as part of this skill.
  • Public functionality only. Announce functionality only when it is publicly available in the named channel by the post's publication date. Do not announce internal dogfood or source-only functionality.
  • Folder structure. A new post MUST live in blogs/<year>/<month>/<day>/ with each date level as its own subfolder. Store images and the social image alongside the post.
  • Frontmatter is required. Every post needs the YAML frontmatter fields from the style guide (Order, TOCTitle, PageTitle, MetaDescription, MetaSocialImage, Date, Author). Order must be unique and higher than the most recent post, so check existing posts before assigning it.
  • Images and screenshots are human work. When a hero image or screenshot is needed, insert a TODO comment for a human to capture and add it later. Do not fabricate image references or invent file names beyond the placeholder.
  • Style compliance. Follow the blog-writing style guide: engaging, conversational tone, active voice, first-person plural ("we"), contractions, and a closing call to action ending with Happy coding! 💙.

Phase 1 — Research & Propose an Outline (no edits)

Do not modify any files in this phase.

  1. Understand the topic. Read the feature description, issue, or PR provided. If it is ambiguous or lacks detail, ask clarifying questions before continuing.

  2. Establish the content framing. Identify the primary persona, reader intent, and narrative purpose by following the blog content-framing guidance. Infer them from the topic, source material, existing coverage, and writer's request. If any part of the framing remains ambiguous and different interpretations would change the angle, ask the writer to confirm before proposing the outline.

  3. Check the source if needed. To understand the feature, inspect the source code in the microsoft/vscode and microsoft/vscode-copilot-chat repos. Use the gh CLI for all GitHub interactions (issues, PRs, code). See user memory gh-cli-powershell.md for PowerShell-specific gh patterns.

    Area being written aboutPrimary source repo
    Core editor, workbench, debug, terminal, tasks, settings, commands, keybindingsmicrosoft/vscode
    Copilot Chat, inline chat, agent mode, chat tools, chat participants, MCP in chatmicrosoft/vscode-copilot-chat
  4. Verify public availability. Confirm that each announced capability is available to users in the named public channel by the publication date. Public Preview, Experimental, and Insiders functionality qualifies when labeled clearly. Source-code existence or manual configurability is not enough. Exclude planned, hidden, internal dogfood, and source-only functionality.

  5. Determine placement and metadata. Decide the target folder (blogs/<year>/<month>/<day>/) and the next available Order value by checking recent posts. Confirm the author name and social media profile link with the user if unknown.

  6. Present the outline. Summarize:

    • The primary persona, reader intent, and narrative purpose.
    • The proposed title, target folder path, and draft frontmatter values.
    • A section-by-section outline (H2 headings) with a one-line description of each.
    • Any TODO image placeholders that will be needed (hero/social image, inline screenshots).
    • Open questions or assumptions.
  7. Stop and wait for approval. Do not proceed to Phase 2 until the user explicitly approves the outline (or adjusts it).

Phase 2 — Draft the Post (after approval)

Once the user approves the outline:

  1. Create the post at blogs/<year>/<month>/<day>/<slug>.md with complete YAML frontmatter.
  2. Write the body following the approved content framing and the content structure from the style guide:
    • An H1 title, followed by a byline <Month> <day>, <year> by [<author>](<social link>) and a blank line.
    • A brief introduction that summarizes the main points.
    • H2 sections as outlined, in an engaging, conversational tone using active voice, "we", and contractions.
    • Images via relative paths with descriptive alt text. Use absolute URLs for links to documentation articles.
    • A closing call to action, ending with Happy coding! 💙.
  3. Add TODO comments where images or screenshots need to be captured by a human.
  4. Verify that the opening, examples, terminology, key takeaways, and call to action serve the approved primary persona and reader intent.
  5. Summarize the post you created and call out any remaining TODOs for the user.

Review Mode

When the user asks you to review an existing blog post instead of writing one:

  1. Read the draft and check it against the blog-writing style guide:
    • Content framing (a clear primary persona, reader intent, and narrative purpose reflected consistently throughout the post).
    • Frontmatter completeness and correctness (all required fields, unique Order, MetaDescription under 160 chars, TOCTitle under 30 chars).
    • Folder structure (blogs/<year>/<month>/<day>/).
    • Content structure (H1 title, byline, intro, H2 sections, alt text on images, absolute URLs for doc links, closing call to action ending with Happy coding! 💙).
    • Writing style (engaging and conversational, active voice, first-person plural, contractions, clear language for a developer audience).
  2. Verify technical claims against the source where practical (setting names, command IDs, version availability), using the gh CLI.
  3. Verify that announced functionality is public in the named channel by the post's publication date. Flag content that presents planned, hidden, internal dogfood, or source-only functionality as available.
  4. Report concrete, practical suggestions grouped by theme. Do not edit the file unless the user asks you to apply the changes.

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