Tool WebMCP Ready
Audit any MCP server for Claude in Chrome and WebMCP compatibility using a comprehensive 6-tier evaluation framework.
Quick Start
npx cbrowser webmcp-ready "https://your-server.com/mcp"
CLI Options
npx cbrowser webmcp-ready "https://your-server.com/mcp" \
--api-key YOUR_KEY # API key if server requires auth
--oauth-token TOKEN # OAuth token if server uses OAuth
--timeout 30000 # Request timeout (ms)
--html # Generate HTML report
--json # Output as JSON
MCP Tool
{
"tool": "webmcp_ready_audit",
"params": {
"url": "https://your-server.com/mcp",
"apiKey": "optional-api-key",
"oauthToken": "optional-oauth-token",
"timeout": 30000
}
}
6-Tier Evaluation Framework
Tier 1: Server Implementation (25%)
Core MCP protocol compliance and response handling.
Checks:
- Server responds to MCP requests
- Uses latest protocol version (2024-11-05)
- Returns proper JSON-RPC errors
- Response time under 5 seconds
Tier 2: Tool Discoverability (20%)
Tools exposed with proper schemas and descriptions.
Checks:
- Tools exposed via
tools/list - Descriptive tool descriptions (>10 chars)
- Input schemas defined for all tools
- Property descriptions in schemas
Tier 3: Instrumentation (15%)
Monitoring, health checks, and OAuth metadata.
Checks:
/healthendpoint available/infoendpoint with name and version- OAuth Protected Resource metadata (for Claude.ai)
Tier 4: Consistency (15%)
Session management, state persistence, and recovery.
Checks:
- Session management tools available
- Self-healing/recovery tools
- Consistent response structure
Tier 5: Agent Optimizations (15%)
Vision, natural language, and persona support.
Checks:
- Screenshot/vision capabilities
- Natural language input support
- Persona-based testing tools
- Accessibility testing tools
Tier 6: Documentation (10%)
AI-readable docs and machine-friendly metadata.
Checks:
/llms.txtor/webmcp.txtavailable- Documentation endpoint (
/docsor/README)
Grade Scale
| Grade | Score | Description |
|---|---|---|
| A | 90-100 | Excellent - Full WebMCP compatibility |
| B | 80-89 | Good - Minor gaps in compatibility |
| C | 70-79 | Fair - Some features missing |
| D | 60-69 | Poor - Significant compatibility issues |
| F | <60 | Failing - Major protocol/feature gaps |
Example Output
{
"url": "https://demo.cbrowser.ai/mcp",
"score": 92,
"grade": "A",
"summary": {
"totalChecks": 18,
"passedChecks": 17,
"criticalIssues": 0,
"highIssues": 0,
"serverResponded": true,
"protocolVersion": "2024-11-05",
"toolCount": 91
},
"tierScores": [
{ "tier": 1, "name": "Server Implementation", "score": 100, "weight": "25%" },
{ "tier": 2, "name": "Tool Discoverability", "score": 95, "weight": "20%" },
{ "tier": 3, "name": "Instrumentation", "score": 100, "weight": "15%" },
{ "tier": 4, "name": "Consistency", "score": 85, "weight": "15%" },
{ "tier": 5, "name": "Agent Optimizations", "score": 100, "weight": "15%" },
{ "tier": 6, "name": "Documentation", "score": 75, "weight": "10%" }
],
"topIssues": [],
"topRecommendations": [
"[QUICK WIN] Add /docs endpoint with usage examples"
],
"duration": 2341
}
Common Issues
| Issue | Tier | Severity | Remediation |
|---|---|---|---|
| Server not responding | 1 | Critical | Verify MCP endpoint is accessible |
| Old protocol version | 1 | Medium | Upgrade to 2024-11-05 |
| No tools exposed | 2 | Critical | Implement at least one MCP tool |
| Tools lack schemas | 2 | High | Define JSON Schema for inputs |
| No /health endpoint | 3 | Medium | Add GET /health for monitoring |
| No session management | 4 | Medium | Add session save/load tools |
| No /llms.txt | 6 | Medium | Create AI-readable docs |
Use Cases
1. Pre-deployment Check
# Verify your MCP server before going live
npx cbrowser webmcp-ready https://staging.yourserver.com/mcp --html
2. Claude.ai Custom Connector Validation
# Check OAuth and metadata requirements
npx cbrowser webmcp-ready https://yourserver.com/mcp \
--oauth-token YOUR_TOKEN
3. Competitive Analysis
# Compare your server against CBrowser's demo
npx cbrowser webmcp-ready https://demo.cbrowser.ai/mcp
Related Tools
- Tool-Agent-Ready-Audit - Audit websites for AI-friendliness
- Tool-AI-Benchmark - Compare AI-friendliness across sites
- Tool-Llms-Txt - Generate and validate llms.txt files
Version
- Added in: v18.15.0
- MCP Tool:
webmcp_ready_audit - Category: Audit