网站审计工具

# Required
- Python 3.12+
- Node.js 22+
- VS Code with MCP support

# Optional (for Docker mode)
- Docker Desktop

# 1. Clone repository
git clone https://github.com/iberi22/WebAuditMCP.git
cd WebAuditMCP

# 2. Install Python dependencies
pip install -r requirements.txt

# 3. Install Node.js dependencies
cd node-tools
npm install
cd ..

# 4. Install Playwright browsers
playwright install chromium

# 5. Configure environment
cp .env.example .env
# Edit .env with your API keys (WAVE_API_KEY optional)

# Build and start container
docker-compose -f docker/docker-compose.yml up -d --build

# Check status
docker logs mcp-auditor

{
  "servers": {
    "webaudit": {
      "command": "python",
      "args": ["e:\\path\\to\\WebAuditMCP\\mcp\\server.py"],
      "env": {
        "CHROME_MCP_ENABLED": "true",
        "PYTHONUNBUFFERED": "1"
      }
    }
  }
}

{
  "servers": {
    "auditor-docker": {
      "url": "http://localhost:8000/mcp"
    }
  }
}

prompts/
├── complete-audit.md          # Full website audit (all categories)
├── security/
│   ├── headers-audit.md       # HTTP security headers
│   ├── vulnerabilities.md     # OWASP ZAP security scan
│   └── ssl-tls.md            # Certificate and HTTPS validation
├── accessibility/
│   ├── wcag-compliance.md     # WCAG 2.1 AA/AAA compliance
│   ├── screen-reader.md       # Assistive technology testing
│   └── keyboard-navigation.md # Keyboard accessibility
├── performance/
│   ├── core-web-vitals.md     # LCP, FID, CLS metrics
│   ├── lighthouse-full.md     # Complete Lighthouse audit
│   └── mobile-performance.md  # Mobile-specific performance
├── visual/
│   ├── responsive-design.md   # Multi-device layout testing
│   ├── cross-browser.md       # Browser compatibility
│   └── visual-regression.md   # Screenshot comparison
└── styles-layouts/
    ├── design-system.md       # Style consistency audit
    ├── typography.md          # Font and readability
    └── color-contrast.md      # WCAG color contrast ratios

Using WebAuditMCP, run lighthouse_fast on http://localhost:3000 with device desktop

Using WebAuditMCP, perform a comprehensive audit of https://example.com following the complete-audit.md prompt. Generate a single Markdown report with:
1. Executive summary with scores
2. Critical issues by priority
3. Actionable recommendations
4. Before/after comparison if baseline exists

Using WebAuditMCP, follow the localhost-audit.md prompt for http://localhost:3000/login

Audit https://website.com for WCAG 2.1 AA compliance using accessibility/wcag-compliance.md. Include:
- Automated axe-core scan
- Manual WAVE review
- Screen reader compatibility notes
- Remediation priority matrix

# Verify all dependencies and tool availability
health_check()

# Mobile audit with all categories
audit_lighthouse(
    url="https://example.com",
    device="mobile"
)

# Desktop audit
audit_lighthouse(
    url="https://example.com",
    device="desktop"
)

# Desktop accessibility audit
scan_axe(
    url="https://example.com",
    device="desktop"
)

# Check HTTP security headers
security_headers(url="https://example.com")

# Test multiple viewports
responsive_audit(
    url="https://example.com",
    viewports=["375x667", "768x1024", "1920x1080"]
)

# Quick baseline security scan (5 minutes)
zap_baseline(
    url="https://example.com",
    minutes=5
)

# Merge multiple audit results with budget enforcement
report_merge(
    items=[lighthouse_result, axe_result, security_result],
    budgets={
        "accessibility": 95,
        "performance": 90,
        "security": 85
    }
)

{
  "status": "ok",
  "url": "https://example.com",
  "timestamp": "2025-10-29T03:30:00Z",
  "scores": {
    "overall": 85,
    "performance": 90,
    "accessibility": 95,
    "security": 70
  },
  "issues": [...],
  "recommendations": [...],
  "artifacts": {
    "screenshots": ["path/to/screenshot.png"],
    "reports": ["path/to/report.html"]
  }
}

export AUDIT_LANGUAGE=es  # Spanish
export AUDIT_LANGUAGE=pt  # Portuguese
export AUDIT_LANGUAGE=zh  # Chinese
export AUDIT_LANGUAGE=en  # English (default)

{
  "servers": {
    "auditor-docker": {
      "url": "http://localhost:8000/mcp"
    }
  }
}

# 1. Clonar repositorio
git clone https://github.com/tuusuario/WebAuditMCP.git
cd WebAuditMCP

# 2. Instalar dependencias
pip install -r requirements.txt
cd node-tools && npm install && cd ..

# 3. Configurar VS Code
# Agregar servidor MCP en .vscode/mcp.json

# 4. Ejecutar servidor
python mcp/server.py

Usando WebAuditMCP, audita https://ejemplo.com siguiendo el prompt prompts/complete-audit.md.
Genera un informe Markdown único con resumen ejecutivo, issues críticos y recomendaciones.

# 1. Clonar repositório
git clone https://github.com/seuusuario/WebAuditMCP.git
cd WebAuditMCP

# 2. Instalar dependências
pip install -r requirements.txt
cd node-tools && npm install && cd ..

# 3. Configurar VS Code
# Adicionar servidor MCP em .vscode/mcp.json

# 4. Executar servidor
python mcp/server.py

# 1. 克隆仓库
git clone https://github.com/yourusername/WebAuditMCP.git
cd WebAuditMCP

# 2. 安装依赖
pip install -r requirements.txt
cd node-tools && npm install && cd ..

# 3. 配置 VS Code
# 在 .vscode/mcp.json 中添加 MCP 服务器

# 4. 运行服务器
python mcp/server.py