SEO Analyzer Skill

基于 Google 官方文档的 SEO 自动检测工具。自动分析网址的技术 SEO、内容元数据、性能体验和链接结构，输出符合 Google 最佳实践的检测报告。

使用方法

自动模式 (推荐)

只需告诉 agent 检查某个 URL，skill 会自动完成所有工作：

python skills/seo-analyzer/scripts/seo_analyzer.py --auto <url>

Skill 会自动:

检查 agent-browser 是否已安装
如果已安装，调用 agent-browser 获取页面内容
执行完整的 SEO 检查
生成详细的检测报告

手动模式

如果 agent-browser 未安装，可以使用手动模式：

# 方式1: 从文件分析
python skills/seo-analyzer/scripts/seo_analyzer.py page.html <url>

# 方式2: 从 stdin 读取 (需要先手动获取 HTML)
agent-browser get source > page.html
python seo_analyzer.py page.html <url>

对于 Agent 的调用方式

当用户请求分析某个 URL 时：

# 使用自动模式
python skills/seo-analyzer/scripts/seo_analyzer.py --auto https://example.com

Agent 内部会：

调用 check_agent_browser_available() 检查 agent-browser
如果可用，调用 fetch_page_with_agent_browser(url) 获取页面
如果不可用，提示用户安装 agent-browser
调用 analyze_page() 执行 SEO 分析
输出格式化报告

Alternative - save to file first:

# Step 1: Open page
agent-browser open <url> --timeout 30000

# Step 2: Save HTML to file
agent-browser get source > /tmp/page.html

# Step 3: Run analysis
python skills/seo-analyzer/scripts/seo_analyzer.py /tmp/page.html <url>

Command-line Options

# Analyze HTML file
python skills/seo-analyzer/scripts/seo_analyzer.py page.html https://example.com

# Read from stdin (pipe from agent-browser)
agent-browser get source | python skills/seo_analyzer.py - https://example.com

# Output in JSON format (for programmatic use)
python seo_analyzer.py page.html https://example.com --json

Workflow

Crawl the URL - Use agent-browser to load the page and capture HTML
Extract page content - Use agent-browser get source to get HTML
Run SEO checks - Pass HTML to seo_analyzer.py for evaluation
Generate report - Output findings with severity levels and recommendations

For Agents: How to Analyze a URL

When asked to analyze a website's SEO, follow these steps:

# 1. Open the URL with agent-browser
agent-browser open <target_url> --timeout 30000

# 2. Get the HTML source from the browser
agent-browser get source > /tmp/target_page.html

# 3. Run the SEO analyzer
python skills/seo-analyzer/scripts/seo_analyzer.py /tmp/target_page.html <target_url>

Or as a single command:

agent-browser open <target_url> --timeout 30000 && agent-browser get source | python skills/seo-analyzer/scripts/seo_analyzer.py - <target_url>

agent-browser open --timeout 30000


### Step 2: Extract the page HTML
After the page loads, get the HTML content:

agent-browser get source

Or use get text to extract visible content:

agent-browser get text "body"


### Step 3: Save HTML to file and run analysis
Write the HTML to a file, then run the analyzer:
```bash
# Save HTML content to a file
echo "<html>...</html>" > /tmp/page.html

# Run the SEO analyzer
python skills/seo-analyzer/scripts/seo_analyzer.py /tmp/page.html <url>

Integrated workflow (recommended)

Use stdin to pass HTML directly:

agent-browser get source | python skills/seo-analyzer/scripts/seo_analyzer.py - <url>

Alternative: Analyze existing HTML file

If you already have HTML content saved:

python skills/seo-analyzer/scripts/seo_analyzer.py /path/to/page.html https://example.com/page

Check Categories

1. Technical Requirements (搜索要素/技术要求)

HTTP status code (must be 200)
Googlebot accessibility
Indexable content presence
HTTPS usage
robots.txt blocking detection
noindex directive check

2. Title Tag (标题链接)

Presence of <title> element
Uniqueness across pages
Descriptive, non-generic text
Avoid keyword stuffing
Proper length (not truncated)
Brand inclusion if appropriate

3. Meta Description (片段)

Presence of <meta name="description">
Unique per page
Descriptive, summarizes content
Not keyword-stuffed
Appropriate length

4. Headings Structure (SEO新手指南)

H1 presence and uniqueness
Logical heading hierarchy (H1→H2→H3)
Heading text is descriptive
Content is organized with headings

5. Link Accessibility (链接)

Links use <a href="..."> format
Anchor text is descriptive
No empty links
No excessive link clustering
External links use appropriate rel attributes (nofollow, sponsored, ugc)

6. Image Optimization (图像)

Images have alt attributes
Alt text is descriptive
Images are near relevant content

7. Structured Data (结构化数据)

JSON-LD, Microdata, or RDFa presence
Valid schema.org types
Required properties for supported types
Content is visible to users
Not blocked by robots.txt or noindex

8. URL Structure (网址结构)

Descriptive URLs (not random IDs)
Human-readable
Logical directory structure
Uses hyphens, not underscores

9. Canonical URL (规范化)

Canonical tag presence
Points to preferred URL
HTTPS preferred over HTTP
Consistent with actual URL

10. Content Quality (创建实用可靠以用户为中心的内容)

Unique, ori

...

seo-analyzer

SKILL.md