Agent Skills › appwrite/appwrite › reviewing-blogs

reviewing-blogs

GitHub

用于审查和批评 src/content/blog/posts/ 下的博客文章,依据 Appwrite 标准评估语音、结构、SEO 及准确性。适用于用户请求审核、审计或改进现有博客时。

apps/console/.cursor/skills/reviewing-blogs/SKILL.md appwrite/appwrite

Trigger Scenarios

review a blog post audit an existing blog critique blog content improve blog post

Install

npx skills add appwrite/appwrite --skill reviewing-blogs -g -y
More Options

Non-standard path

npx skills add https://github.com/appwrite/appwrite/tree/main/apps/console/.cursor/skills/reviewing-blogs -g -y

Use without installing

npx skills use appwrite/appwrite@reviewing-blogs

指定 Agent (Claude Code)

npx skills add appwrite/appwrite --skill reviewing-blogs -a claude-code -g -y

安装 repo 全部 skill

npx skills add appwrite/appwrite --all -g -y

预览 repo 内 skill

npx skills add appwrite/appwrite --list

SKILL.md

Frontmatter
{
    "name": "reviewing-blogs",
    "description": "Reviews and critiques blog posts in src\/content\/blog\/posts\/ against Appwrite's established standards for voice, tone, structure, SEO, formatting, and accuracy. Use when the user asks to review, audit, critique, or improve an existing blog post."
}

Repository layout (apps/console)

This skill targets the Console app at apps/console in appwrite/appwrite; paths are relative to it. Path mapping from the upstream appwrite/website skills:

Artifact Path
Blog post src/content/blog/posts/<slug>.markdoc
Docs page src/content/docs/{path}/index.markdoc
Changelog entry src/content/changelog/entries/<YYYY-MM-DD>.markdoc
Cover image public/images/blog/<slug>/cover.avif (min 1200px wide; run bun run generate:cover-manifest after adding)

Also apply the unslop skill to all user-facing prose.

Blog Review

How to run a review

When asked to review a blog post, read the file at src/content/blog/posts/<slug>.markdoc, then evaluate it against every section below. Produce a structured report using the format at the end of this document.

Before starting: Identify which sections of src/content/docs/ are relevant to the post's topic, then read those index.markdoc files directly to use as the ground truth for the accuracy review. The docs map to public URLs as src/content/docs/{section}/{subsection}/index.markdoc → https://appwrite.io/docs/{section}/{subsection}. Key sections: products/{auth,databases,storage,functions,messaging,sites,ai}/, quick-starts/, tutorials/, advanced/{platform,security,self-hosting}/, apis/, tooling/.

Review checklist

Frontmatter

  • layout: post is present
  • title is a quoted string, 50-60 characters, front-loads the primary keyword
  • description is 150-160 characters, includes the primary keyword, states what the reader will learn
  • date is in YYYY-MM-DD format
  • cover path matches the slug: /images/blog/<slug>/cover.avif
  • timeToRead is plausible given word count (roughly 200-250 wpm)
  • author is set
  • category is set
  • featured and unlisted are present

Slug consistency

  • Filename slug matches cover image path slug
  • Any internal links use the same slug

Voice and tone

  • Direct and opinionated — no hedging without follow-through
  • Developer-first — no marketing speak or fluff
  • Practical over theoretical — concrete problems before reasoning
  • No unnecessary superlatives ("amazing", "powerful") without substance

Post structure

  • Opens with the real problem or a misconception — no preamble like "In this post, we will..."
  • Establishes the cost or stakes early
  • Gives a concise, direct answer to the primary search query early for AEO
  • Uses useful H1/H2 sections and H3 subsections to cover each dimension clearly
  • Includes concrete examples
  • Includes a relevant Appwrite use case
  • Includes concrete, specific guidance (not just "consider X")
  • Includes an FAQ section with direct, standalone answers
  • Closes with a single combined CTA and resources section
  • Closing section heading is specific to the post topic (not "Moving forward", "Conclusion", "Next steps", "Where to go from here")
  • Closing paragraph reflects the post's specific argument, not boilerplate Appwrite pitch

Formatting

  • **bold** used for key terms and callouts within paragraphs
  • Bullet lists used for enumerating items, not for general prose
  • No nested lists unless strictly necessary
  • H1 (#) for major sections, H2 (##) for subsections, and H3 (###) where further structure is useful
  • Paragraphs are 2-4 sentences
  • No em dashes (—) anywhere — sentence must be rewritten if one exists
  • Section headings are specific and declarative, not questions or vague phrases

SEO

  • Title is clear, 50-60 characters, and includes the primary keyword naturally
  • Primary keyword appears in the description frontmatter
  • Meta description is 150-160 characters and states what the reader will learn
  • Blog body is 1,500-2,000 words
  • Slug is short, keyword-rich, and hyphen-separated
  • Primary keyword appears naturally in the introduction, at least one heading, and the body
  • Headings reflect real developer search queries, not internal jargon
  • Post type matches search intent (tutorial for "how to", comparison for decision-making, etc.)
  • Contextual internal links point to relevant Appwrite docs, product pages, and related blogs within the body
  • Direct answer sections make key answers easy for answer engines to extract
  • Clear headings, bullets, short paragraphs, and short sections make the post easy to scan
  • Credible external links support accuracy, trust, and E-E-A-T where relevant

Appwrite accuracy

Read the relevant local doc files before evaluating this section. Use the directory structure in src/content/docs/ to identify which files apply, then read each one. Do not rely on training data — verify every Appwrite-specific claim against the local docs.

  • All Appwrite feature names match the live docs exactly (e.g. "Appwrite Auth" not "Appwrite Authentication")
  • All API method names, parameters, and signatures match the live docs
  • Configuration options and their valid values match the live docs
  • Behavioral descriptions (e.g. "Appwrite automatically does X") are confirmed by the docs
  • No outdated or deprecated references (check if the feature still exists and works as described)
  • Comparisons to alternatives (Firebase, Vercel, Netlify, etc.) are fair and factual
  • Trade-offs are acknowledged rather than dismissed

For each inaccuracy found, note: the claim in the post, what the docs actually say, and the doc URL.

Links and resources

  • All doc links in the closing section are directly relevant to the post's topic
  • No filler links to Discord, GitHub, or the Appwrite homepage unless nothing more specific exists
  • Closing section includes a link to Appwrite Cloud unless the post is specifically about self-hosting
  • All linked URLs resolve (cross-check against the local doc files in src/content/docs/)

Common pitfalls

Check whether the post commits any of these:

  • Starts with "In this post, we will..." or similar throat-clearing
  • Over-explains basics the target developer already knows
  • Ends with "Let us know your thoughts!" or similar engagement bait
  • Uses passive voice where active is clearer
  • Uses em dashes
  • Uses question-format section headings
  • Uses generic closing headers
  • Repeats boilerplate CTA phrasing
  • Links to Discord/GitHub/homepage as filler

Review report format

Output the review as a structured Markdown report with these sections:

## Blog Review: <title>

### Summary
<2-3 sentence overall assessment: what works, what needs the most attention>

### Frontmatter
<Pass / Issues found — list each issue>

### Voice and tone
<Pass / Issues found>

### Structure
<Pass / Issues found>

### Formatting
<Pass / Issues found>

### SEO
<Pass / Issues found>

### Appwrite accuracy
<Pass / Issues found — note any claims that could not be verified>

### Links and resources
<Pass / Issues found>

### Specific rewrites
<For each issue that has a clear fix, show the before and after>

### Priority fixes
1. <Most important fix>
2. <Second most important>
3. <Third most important>

Be specific. Quote the problematic text directly. Suggest the exact rewrite where possible. Do not pad the report with praise for things that merely meet the baseline standard.

Version History

  • a8c84a5 Current 2026-10-20 05:01

Same Skill Collection

apps/console/.cursor/skills/auditing-seo/SKILL.md
apps/console/.cursor/skills/writing-blogs/SKILL.md

Metadata

Files
0
Version
a8c84a5
Hash
8674e2bb
Indexed
2026-10-20 05:01

Главная - Вики-сайт
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-10-24 20:58
浙ICP备14020137号-1